Sensory profiles and oxidative stability of linseed oil microencapsulated
with pea, soy, and whey proteins in high-fat food models
Sari A. Hakanen a, Annelie Damerau a, Dorota Ogrodowska b, Annalisa Seubert a,
Waldemar Brandt c, Oskar Laaksonen a, Małgorzata Tanska b, Kaisa M. Linderborg a,*
a Food Sciences, Department of Life Technologies, University of Turku, 20014 Turun Yliopisto, Turku, Finland
b Department of Food Plant Chemistry and Processing, Faculty of Food Science, University of Warmia and Mazury, 10-718, Olsztyn, Poland
c Department of Dairy Science and Quality Management, Faculty of Food Science, University of Warmia and Mazury, 10-719, Olsztyn, Poland
A R T I C L E I N F O
Keywords:
Dark chocolate
Functional foods
HS-SPME-GC-MS
Microencapsulation
Omega-3 fatty acids
Shortbread cookies
Flaxseed
A B S T R A C T
To hinder oxidation and mask flavor, linseed oil was encapsulated using maltodextrin with pea, soy, or whey
protein, and incorporated in dark chocolate and shortbread cookie models. While the spray dried emulsions with
whey protein and maltodextrin were most stable, this behaviour did not translate to food models. In volatiles, the
fortification with pea protein capsules had closer resemblance to the control than other spray-dried samples.
Storage time increased the volatiles of particularly soy and whey samples. In flavor, the pea sample, and in
appearance and texture, the soy sample resembled closest the control chocolate. Addition of the spray dried
emulsion increased the darkness, rancidity, roasted flavor, and hardness of the cookies. While the pea protein
emerged as an intriguing choice for shell material in spray-drying of linseed oil, also fortification with non-
encapsulated linseed oil resulted in favorable outcomes based on sensory perception and volatile profiles of
the food models.
1. Introduction
The essential omega-3 polyunsaturated fatty acid (n-3 PUFA),
α-linolenic acid (ALA, 18:3n-3), contributes to e.g. healthy cholesterol
level, blood pressure and cardiovascular health (Bork et al., 2022; Sal-
a-Vila et al., 2022; Wei et al., 2018). However, the intake of ALA, and
dietary n-3 PUFA in general, is universally below recommendations
(Guesnet et al., 2019; Sioen et al., 2017). Main sources of ALA are plant
seed oils (Wang et al., 2021; Yuan et al., 2022), among which especially
linseed oil, also known as flaxseed oil, stands out with its superior ALA
content (Kajla et al., 2015; Mueed et al., 2022). However, the high ALA
content of linseed oil makes it susceptible to oxidation, which can lead to
a reduction in both nutritional and sensory quality (Brühl et al., 2007;
Frankel, 2012; Mozuraityte et al., 2016; Wang et al., 2021).
Microencapsulation, most commonly by spray-drying, may offer
protection against oxidation. Sensitive core agent is capsulated in a
protective shell material, resulting in a spray-dried emulsion (SDEM)
powder (Delshadi et al., 2020; Ruiz Ruiz et al., 2017). In addition to oils
and fats as such, pastries and cakes serve as a source of ALA (Guesnet
et al., 2019; Redruello-Requejo et al., 2023). Hence, incorporating
SDEMs into these common food products could provide convenient
means of elevating the recommended daily intake of ALA. In the EU,
foods containing 0.3 g of ALA per 100 g and per 100 kcal may be labeled
as a source of n-3 fatty acids (FAs). Additionally, products may be
labeled as a food high in n-3 FAs when they contain at least 0.6 g of ALA
per 100 g and per 100 kcal. (Regulation (EC) No 1924/2006 of the
European Parliament and of the Council of December 20, 2006 on
Nutrition and Health Claims Made on Foods, 2014.)
Polymers typically used as coating materials include carbohydrates,
dairy proteins, and their mixtures (Chang & Nickerson, 2018; Per-
ez-Palacios et al., 2022). According to Chang and Nickerson (2018)
plant-based proteins present a sustainable advancement in the field of
microencapsulation. Previously, for example, soy protein isolate
(Tambade et al., 2020), rice protein (Gomes & Kurozawa, 2021), lentil
protein isolate (Wang et al., 2022), and pea protein isolate (Bajaj et al.,
2015; Damerau et al., 2022a) have been used in encapsulation. Despite
the advancements, further investigation is needed to enhance the quality
of microencapsulated oils. For instance, Wang et al. (2022) observed
oxidation in microencapsulated oils after an 8-week storage period.
Furthermore, sensory properties have a significant importance in the
* Corresponding author.
E-mail address: kaisa.linderborg@utu.fi (K.M. Linderborg).
Contents lists available at ScienceDirect
LWT
journal homepage: www.elsevier.com/locate/lwt
https://doi.org/10.1016/j.lwt.2024.117305
Received 11 November 2024; Received in revised form 20 December 2024; Accepted 28 December 2024
LWT - Food Science and Technology 216 (2025) 117305 
Available online 29 December 2024 0023-6438/© 2024 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY license ( http://creativecommons.org/licenses/by/4.0/ ). 
consumer acceptance of foods (Baker et al., 2022). Regrettably, to the
authors’ best knowledge, there is a limited amount of research con-
cerning the enrichment of foods with encapsulated linseed oil, especially
with sustainable plant proteins.
In this study, linseed oil was encapsulated using maltodextrin, along
with either pea, soy, or whey protein, and incorporated within dark
chocolate and shortbread cookies. The main objectives were to study the
chemical and sensory quality of the SDEMs and the food models. The
hypotheses of this research were that plant proteins are suitable wall
materials to replace whey protein in spray drying of linseed oil and that
encapsulation will enhance the oxidative stability and preserve the
sensory quality of n–3 PUFAs in fortified foods during storage.
2. Materials and methods
2.1. Materials
2.1.1. Spray-dried emulsions
SDEMs were produced as described in Damerau et al., (2022a) using
cold-pressed linseed oil, LO, (Olejarnia Swiecie, Swiecie nad Osą,
Poland) combined with maltodextrin, M, (DE 14–22, Edpol Food &
Innovation company, Łom _za, Poland) and either whey protein, W, (WPC
80, Ostrowia Company,Warszawa, Poland), pea protein, P, (pea protein,
Roquette, Lestrem, France) or soy protein, S, (soy protein, Barentz Food
& Nutrition, Netherlands), keeping the ratio of LO, maltodextrin and
protein as 2:1:1. Aqueous emulsions (2 L) were formed using Thermomix
(Vorwerk, Wuppertal, Germany), and then homogenized either in
two-steps (25 MPa and 5 MPa) using a high-pressure laboratory valve
homogenizer (Panda 2 K, GEA Niro Soavi, Parma, Italy). The emulsions
were spray dried in a Production Minor Spray Dryer (A/S Niro Atomizer,
Copenhagen, Denmark) equipped with a disc spray system. The feed
flow rate was maintained at 77 mL/min, while inlet and outlet air
temperatures were 125 C and 75 C, respectively. Thus, the following
SDEMs were produced: WM (whey protein and maltodextrin), PM (pea
protein and maltodextrin), and SM (soy protein and maltodextrin).
2.1.2. Food models
Either SDEM or non-encapsulated LO was incorporated into choco-
late (Table 1), prepared according to Damerau et al. (2022a), to meet
>0.3 g ALA per 100 g and 100 kcal. Thus, the following chocolate
models were produced: W (enriched with WM capsules), P (enriched
with PM capsules), S (enriched with SM capsules), O (enriched with
non-encapsulated LO), and a reference chocolate without LO as a con-
trol, R.
The cookie models (Table 1) were prepared as described in (Damerau
et al., 2022a) with minor modifications. 34.5 g of margarine (Keiju 70%,
Bunge Finland Plc, Finland), 15.2 g sugar (Aito Hienosokeri, VD-Group
Plc, Belgium) and 50.2 g wheat flour (Myllyn Paras Plc, Hyvinkaa,
Finland) were mixed. As a result, the following cookie models were
baked: W (enriched with WM capsules), P (enriched with PM capsules),
S (enriched with SM capsules), O (enriched with non-encapsulated LO),
and a reference cookie without LO, R. An addition of either SDEM or
non-encapsulated LO to the cookies resulted in>0.6 g ALA per 100 g and
100 kcal (marked with 1) or double of this concentration (marked with
2). In the cookies fortified by SDEM, an equal proportion of flour and
margarine was replaced by SDEMs, whereas in LO fortified cookies part
of margarine was replaced by non-encapsulated LO. The cookies (ca. 20
g) were baked at 175 C for 17 min. They weighed thereafter 17.8 g
(0.13 g).
2.1.3. Storage trial of food models
The chocolate was stored at room temperature (23.6 C) at humidity
of 39.8 % for 11 weeks and the cookies at 50 C for 31 days (Table 1). 1 g
 0.03 g of finely cut chocolate and finely crushed cookies were
weighted in triplicate in 20 mL headspace vials. The fresh samples were
flushed with nitrogen to prevent further oxidation. Samples in storage
trial were stored under air. For sensory analysis and lipid extraction, the
samples were stored as whole; the chocolates were wrapped in waxed
paper and stored in resealable plastic bags, and the cookies were stored
in small aluminum foil containers with lids. After the trial the cookies
were stored at  80 C for chemical and at  20 C for sensory analysis.
Chocolate was stored at 4 C. Prior to the sensory assessments, a new
batch of chocolate was prepared for the fresh samples.
2.2. Methods
2.2.1. Scanning electron microscopy
SDEMs were attached to the microscope state by using two-sided
adhesive tape, mounted on scanning electron microscopy (SEM) tubs,
and coated with palladium in a sputter coater. SEM Quanta 200 (FEI
Company, Hillsboro, OR) operating an accelerating voltage of 30 kV and
 400 magnifications was used.
2.2.2. Total and surface oil content and encapsulation efficiency
Surface oil (easily extractable oil) and total oil contents of SDEMs
were determined as previously described in Takeungwongtrakul et al.
(2015) and Damerau et al. (2022b) using hexane (Sigma-Aldrich,
Poznan, Poland) or a chloroform/methanol (Sigma_Aldrich, Poznan,
Poland) mixture (2:1, v/v), respectively. The oil content was determined
gravimetrically after evaporation of solvent. Encapsulation efficiency
(EE) (%) was calculated by using equation (1), where TOC indicates total
oil content and SOC indicates surface oil content:
EE …%†ˆ
…TOC  SOC†
TOC
 100 (1)
2.2.3. Fatty acid analysis
Lipids from SDEMs were extracted in triplicate as described by
Damerau et al. (2022b). Lipids from the food models were extracted
using a modified Folch extraction method (Folch et al., 1957). 0.3 g of
food model and 1.23 mL 2-propanol (Honeywell International Inc.,
Germany) were mixed with an ultra turrax for 30 s at 800 rpm. 3.75 mL
hexane (Honeywell International Inc., Germany) was added and mixed
as previously described. The mixture was shaken with 1.5 mL 0.8%
potassium chloride in MQ-water and centrifuged at 1700g for 2 min.
The extraction was repeated with 3.75 mL hexane. The organic phases
Table 1
Sample materials and their codes including storage periods for collecting sam-
ples for chemical and sensory analysis.
Chocolate models Cookie models
Samples a R, O, W, S, P R, O1, O2, W1,W2, S1, S2, P1, P2
Time point
b
Storage
period (d)
Samples
collected c
Storage
period (d)
Samples
collected c
0 0 VOC, FA, PV,
GDA
0 VOC, FA, PV,
GDA
1 14 VOC 7 VOC
2 35 VOC, FA, PV 14 VOC, FA, PV
3 56 VOC 24 VOC, FA, PV,
GDA
4 76 VOC, FA, PV,
GDA
31 VOC
a R ˆ not fortified food model, O ˆ fortified with non-encapsulated LO, W ˆ
fortified with whey protein concentrate and maltodextrin capsules, S ˆ fortified
with soy protein and maltodextrin capsules, P ˆ fortified with pea protein and
maltodextrin capsules, 1 ˆ fortified with SDEM or non-encapsulated linseed oil
to gain 0.6 g ALA per 100 g and 100 kcal, or 2 ˆ fortified to gain 1.2 g ALA per
100 g and 100 kcal.
b In the sample abbreviations, the timepoint is indicated after the underscore.
c Abbreviations of conducted analyses: VOC ˆ volatile oxidation compounds
by HS-SPME-GC-MS, FA ˆ fatty acids by GC-FID, PV ˆ peroxide value, and GDA
ˆ general descriptive analysis.
S.A. Hakanen et al. LWT 216 (2025) 117305 
2 
were collected. Acid-catalyzed FA esterification method was employed
(Christie, 2010). The FA methyl esters were analyzed with a Shimadzu
GC-2030 with an AOC-20i autoinjector, a flame ionization detector
(Shimadzu Corporation, Kyoto, Japan), and DB-23 column (60 m 0.25
mm  0.25 μm, Agilent Technologies, J.W. Scientific, Santa Clara, CA,
USA) under previously reported conditions (Damerau et al., 2022a).
Peaks were identified with Supelco 37 Component FAME mix (Supelco,
Inc. St. Louis, MO, USA), 68D and GLC-490 (Nu-Check-Prep, Elysian,
MN, USA) as external standards. The quantification of FAs was based on
internal standard (Triheptadecanoin (Larodan Fine Chemicals AB,
Malmo, Sweden)) and correction factors.
2.2.4. Oxidative stability index
Oxidative stability index (OSI, expressed as hours) was determined
using A 743 Rancimat (Methrom, Herisau, Switzerland) eight-channel
oxidative stability instrument (Damerau et al., 2022b). 2.5 g of each
SDEM or food sample was placed in a capped reaction vessel in a ther-
mostatic electric heating block. The temperature was set at 110 C and
the airflow rate at 20 L/h.
2.2.5. Physical properties
Hardness was measured using a breaking test of an Instron universal
testing machine type 4301 (Canton, MA, USA) with Instron Series IX
AMTS ver. 8.04 software at a temperature of 21 C (two and four
measurements for the chocolate and cookie models, respectively), using
dimensions 20 mm  15 mm  8 mm (length  width  thickness) for
chocolate and 50 mm  10 mm (diameter  thickness) for cookies.
Measurements of color parameters L*, a*, and b* of the prepared
chocolate and cookie samples were done using a digital image analysis
at a temperature of 21 C (three and four measurements for the choco-
late and cookie models, respectively). The set consisted of a computer, a
high-resolution, low-noise charge-coupled device (CCD) Nikon DXM-
1200 color camera (Nikon Inc., Melville, USA), a Kaiser RB-5004-HF
High Frequency Daylight Copy Light Set with 4  36 W fluorescent
light tubes (color temperature 5400 K) (Kaiser Fototechnik GmbH & Co.
KG, Buchen, Germany), and Laboratory Universal Computer Image
Analysis (LUCIA) G v. 4.8 software (Ogrodowska et al., 2017). Addi-
tionally, total color differences (ΔE) of the fortified samples compared to
the control sample were calculated from equation (2) (Razavizadeh &
Tabrizi, 2021):
ΔEˆ


L*f  L
*
c
2
‡

a*f  a
*
c
2
‡

b*f  b
2
r
(2)
where, L*, a*, b* are color parameters for the fortified (f) and control
samples (c).
2.2.6. Peroxide value
The peroxide value (PV) of lipid extracts (equal to those used in FA
analysis) was determined with a modified ferric thiocyanate method
(Lehtonen et al., 2011) using iron (III) chloride standard line and
expression mEq/kg food model.
2.2.7. Volatile oxidation compounds
The volatile oxidation products (VOCs) of food models were
analyzed with solid-phase microextraction gas chromatography-mass
spectrometry (SPME-GC-MS). TriPlus RSH autosampler coupled with
TRACE 1310 GC and ISQ 7000 MS detector (Thermo Scientific, Reinach,
Switzerland) and equipped with DVB/CAR/PDMS-fiber 50/30 μm
(Supelco, Bellefonte, PA, USA) and SPB®-624 capillary column (60 m 
0.25 mm  1.4 μm, Supelco, Bellefonte, PA, USA), and GC-MS condi-
tions described in Damerau et al. (2022a) were used. Identification of
compounds was performed using the NIST MS Search library (version
2.4, National Institute of Standards and Technology, Gaithersburg, MD,
USA).
2.2.8. Sensory assessments
Two sensory panels, consisting of the students and the staff of the
University of Turku and the University of Applied Sciences of Turku,
were recruited to evaluate chocolates (n ˆ 11) and cookies (n ˆ 11)
using the generic descriptive analysis (GDA). All assessors acknowl-
edged an informed consent statement to participate in the study. They
were informed about the research objectives, and participants’ rights
and privacy in compliance with the GDPR, the Instructions of Finnish
National Board on Research Integrity, and the University of Turku. At
the time of the sensory evaluation, pre-ethical permission for sensory
evaluations was not required by our institution.
Three and four training sessions were held prior to the chocolate and
the cookie assessments, respectively. The panel agreed on 19 attributes
for cookies and 22 for chocolate (Supplementary Table 1 and Supple-
mentary Table 2). The food models were evaluated in duplicate as fresh
and after the storage period. The assessments were conducted in a
sensory laboratory in accordance with the ISO 8589 standard. Cookies
(ca. 6.3 g) and chocolate (ca. 5 g) were served at room temperature in
covered glass bowls labeled with three-digit numerical random codes.
Sample order was randomized using the Latin Squared design. Carbon
filtered water and water biscuits (Carr’s Table Water, Carr’s of Carlisle
Ltd, Great Britain) were provided for palate cleansing during forced time
delays (30 s) between samples. Compusense Cloud (version 21.0,
Compusense Inc., Guelph, ON, Canada) was used for data collection.
2.3. Statistical analysis
The evaluation of the sensory panels and the assessors’ performance
was conducted using PanelCheck, version 1.4.2 (Nofima, Tromsø, Nor-
way; Tomic et al. (2010)). Analysis of variance (ANOVA) was performed
using IBM SPSS Statistics 27 (IBM Corporation, Armonk, NY, USA) with
p < 0.05. One-way ANOVA was used to compare ALA content, OSI,
hardness, and color while a two-way mixed ANOVA was used for PV
value analysis. Two-way ANOVA models were carried out for sensory
attributes (main effects and sample  assessor interactions; samples as
fixed factors, assessors as random factors). Tukey’s HSD tests or Bon-
ferroni post hoc tests were conducted after the ANOVA models.
The data of the volatile compounds and the consensus sensory data,
averaged over the assessors and the replicates, were subjected to the
principal component analysis (PCA) executed by Unscrambler® X,
version 11.0 (Camo Analytics AS., Oslo, Norway). The mean centered
sensory data and auto-scaled volatile data underwent full and random
cross validation, respectively, using NIPALS.
3. Results and discussion
3.1. Characterization of SDEMs
SDEMs were of relatively spherical shape (Fig. 1). However, SM
consisted of particles with higher uniformity and larger diameters, while
in WM and PM, a greater number of particles with small diameters were
found. Previously, the usage of soy proteins as emulsifiers in encapsu-
lation of essential oils resulted in particles with significantly higher
mean diameters than those with pea protein concentrate (Francisco
et al., 2020). Higher solubility of soy proteins may promote a higher
viscoelasticity of the atomized emulsion droplets, thus, greater swelling
during the drying step, resulting in larger particles. In this study, some
particles in SM had visible surface folds indicating softer structures and
greater susceptibility to deformation. Additionally, greater shrinkage of
particles has been reported in samples consisting of sunflower oil and
soy proteins compared to particles containing pea proteins and the same
oil (Le Priol et al., 2019). The particle shrinkage may be related to rapid
formation of wall crust layer on the particle surface followed by the high
amount of water evaporation throughout the drying process (Gong et al.,
2016).
The total oil content of SDEMs was in the range of 45.96 (PM) to
S.A. Hakanen et al. LWT 216 (2025) 117305 
3 
49.15% (WM) (Table 2), similar as introduced in the emulsion (50%).
The significantly different result of the PM sample may be related to the
higher surface oil content (15.32%) compared to the WM (6.40%) and
SM (8.57%) samples. Generally, low surface oil content and thus high EE
are indicators of the success of microencapsulation. Aberkane et al.
(2014) reported EE to be influenced by the type of coating material.
They did not indicate pea protein without the addition of any poly-
saccharide as a suitable encapsulating material. In our study, the com-
bination of P with M resulted in the highest surface oil content of the
studied LO microcapsules. Carneiro et al. (2013), using the same ma-
terials, obtained a much lower EE (62.3%), which is related to the total
solid concentration (30%) in emulsion and wall-to-core ratio of 1:2. In
our study, total solid concentration was 20% and the ratio of wall and
core ingredients was 1:1. However, despite the low EE, samples showed
the highest oil protection against oxidation (Carneiro et al., 2013),
which is consistent with our results (Table 2). Francisco et al. (2020)
found higher EE for encapsulated orange essential oils using soy protein,
likely due to larger particle size, which is consistent with the present
study. Similarly, Le Priol et al. (2019) reported higher EEs for sunflower
oil with soy protein (91%) compared to pea protein (88%), indicating a
more efficient encapsulation process with soy protein.
The FA profile of LO consisted of ALA (57.5%), oleic (19.4%), lino-
leic (14.8%), palmitic (5.1%) and stearic acid (2.6%), with other FAs
being less than 1%, which is in comparable range with our earlier studies
(Tanska et al., 2016; Ogrodowska et al., 2024). ALA content of SDEMs
was marginally but significantly lower than of LO (Table 2).
OSI showed that all SDEMs were more stable than LO (Table 2). WM
was the most stable followed by PM. SM was least stable but still two
times more stable than LO. This is consistent with the results of encap-
sulation of fish oil with the same wall material combinations and spray-
drying conditions (Damerau et al., 2022a), indicating that wall material
selection seems to have a big impact on the OSI.
3.2. Physical and chemical quality of food models
3.2.1. Hardness and color of the food models
The texture, especially hardness, of chocolates and cookies affects
their perceived quality. 30% less force was needed to break the choco-
late with non-encapsulated oil than the control sample (Table 3). The
addition of SDEMs, most strongly SM, increased the breaking force
needed. Also, Hadnađev et al. (2023) reported that breaking force was
higher for samples of chocolate with microcapsules, and equal to in our
study, samples with whey protein were softer compared to samples with
plant proteins. An opposite relationship was observed for the cookie
model, where samples with whey proteins (W1, W2) were harder than
those with plant proteins (Table 3). Furthermore, it was observed that
the incorporation of twice the amount of each SDME into the cookies
resulted in a noticeable increase in their hardness.
All fortified chocolates had a significantly lighter color, lower blue-
ness and lower yellowness compared to control (Table 3). Hadnađev
et al. (2023) observed differences in color parameters in chocolates with
microcapsules, regardless of the coating protein type. Also, Razavizadeh
and Tabrizi (2021) reported that chocolates with microencapsulated
chia seed oil were lighter than the control. The influence of SDEM
addition on cookie color was significantly lower than in the case of
chocolate, as shown by the lower ΔE values for enriched samples. The
exception was only samples with twice the concentration of whey and
pea proteins, which were characterized by a higher yellowness.
Fig. 1. SEM for spray-dried emulsions formulated with M ˆ maltodextrin, W ˆ
whey protein concentrate, S ˆ soy protein or P ˆ pea protein as wall material.
Table 2
Total and surface oil contents, encapsulation efficiency, α-linolenic acid (ALA, 18:3n-3) content and oxidative stability index (OSI) of spray-dried emulsions (n ˆ 2).
SDEMa Total oil [%] Surface oil [%] Encapsulation efficiency [%] ALA content of oil content [%] OSI [h]b
WM 49.15  0.95b 6.40  0.23a 86.98  0.21c 56.22  0.01a 13.08  0.22c
SM 49.29  0.30b 8.57  0.33b 82.62  0.78b 56.93  0.01b 5.05  0.23a
PM 45.96  0.58a 15.32  0.60c 66.68  0.88a 56.06  0.04a 11.73  0.22b
a Spray-dried emulsion (SDEM) with W ˆ whey protein concentrate, S ˆ soy protein or P ˆ pea protein and M ˆ maltodextrin as wall material. Different letters on
the same column indicate a statistically significant (p < 0.05) difference between samples.
b OSI – oxidative stability index at 110 C, for non-encapsulated oil was 2.39  0.11 h.
S.A. Hakanen et al. LWT 216 (2025) 117305 
4 
3.2.2. ALA content and oxidative stability
The ALA level of fresh fortified chocolate ranged from 2.4 g to 2.8 g
per 100 g (Table 3), which is 0.39 g–0.47 g per 100 kcal, and fulfilled the
claim “source of n-3 fatty acids”. S had a significantly higher amount of
ALA compared to other chocolate samples. However, this could be
related to the initial nutritional calculations indicating that the required
amount of SDEM addition was higher than other models. In fresh for-
tified cookies, the ALA concentration varied from 4.4 to 4.8 g and
8.1–8.6 g per 100 g, which is 0.89–0.97 g and 1.57–1.63 g per 100 kcal,
respectively. In addition to the SDEMs, ALA originated from also the
margarine. In cookies containing lower concentrations of ALA, the
amount of ALA was significantly higher in P compared to S and W
(Table 3). Conversely, in cookies with higher ALA concentration, the
ALA level was highest in S.
PV value was initially low in both food models but increased during
the storage period (Fig. 2). In the chocolates, the amount of primary
oxidation compounds was slightly but significantly lower in P followed
by S and W (Fig. 2A). The PV level of the stored cookies appears
consistent with the findings of Srivastava and Mishra (2021) who for-
tified cookies with vegetable oil. The hydroperoxide level of S and P
cookies was the same or lower compared to R and O, while W exhibited
the least effective protection against oxidation (Fig. 2B). Surprisingly, W
showed the lowest stability of all models fortified with SDEM, while of
the SDEMs WM was the most oxidative stable. Lack of correspondence
between the stability of SDEMs and food models has been also previ-
ously detected (Damerau et al., 2022a). The differences in behaviour for
W compared to P and S might be explained by fact that amino acid
composition of PM and SM is more similar than one of WM, which could
have caused differences in interactions of wall material of SDEMs with
the food matrix and therefore on any possible protecting effects of the
food matrix.
The fortification with non-encapsulated LO and SM lowered the
stability of chocolate measured by OSI while addition of PM and WM
increased it (Table 3). Highest stability was observed for W. Compared
to the PV the order of stability was similar as observed for SDEMs with
OSI. The addition of non-encapsulated LO in the higher concentration
showed a lower OSI than R (Table 3). S1 had a significantly higher OSI
than other fortified cookies. This result was in line with PV findings for
S1. Overall, the order of stability based on PV and OSI was dissimilar for
the food models. Differences were less pronounced with OSI than with
PV, possibly because PV measures primary oxidation while OSI mea-
sures time until complete oxidation. However, results of both methods
combined show a clear trend that fortification with SDEMs improved
stability of food models compared to direct addition of oil. Further, the
findings indicate that interactions with food matrix play a big role in
stability of SDEMs in foods. Therefore, stability of SDEMs needs to be
studied in foods and cannot be estimated by stability of SDEMs.
3.2.3. Volatile profile of the food models
A total of 73 and 96 volatile compounds were identified in the
chocolate and cookie models, respectively (Table 4). The most abundant
chemical groups were acids and aldehydes. Acetic acid, 3-methylbuta-
noic acid, and hexanal were the major volatiles in chocolate. In our
previous study (Damerau et al., 2022a, Damerau et al., 2022b), acetic
acid and 3-methylbutanoic acid were also the highest detected com-
pounds in dark chocolate models. The source of acids in volatile profile
of dark chocolate is mainly from the fermentation process of cocoa bean
(Afoakwa et al., 2009). Among the identified compounds in cookies,
hexanal, pentanal, 3-methylbutanal, and hexanoic acid dominated. They
are oxidation indicators previously found from oxidized LO
(Gomez-Cortes et al., 2015).
Many volatiles were only formed during storage, while some others
reacted further and decreased during storage. The total amount of vol-
atiles increased over the storage period in food models fortified with
SDEMs (Table 4). In the beginning of the storage trial, the chocolate
fortified with SDEMs demonstrated lower levels of volatiles compared to
R and O. In comparison to chocolate, the shell material of SDEMs
exhibited distinct behavior in cookies from the start of storage trial.
Further, the fortification concentration influenced the content of vola-
tiles, as noted also previously (Damerau et al., 2022a).
In the PCA model for the chocolates (Fig. 3A and B), majority of
volatile compounds are clustered on the positive side of PC1 (with 53%
of total variation), while only 2-acetylpyrrole is negatively loaded. PC1
predominantly discriminates the samples according to the storage
period, with those stored longer positioned on the positive side of PC1.
This alignment correlates with most volatiles, a pattern particularly
evident in SDEM fortified chocolate compared to the others. Regarding
PC2 (16% of variation), it exhibits a positive loading associated with a
cluster of volatiles, including 2-methylpropanoic acid, acetic acid, 2-
methylbutanoic acid, and alpha-pinene. Additionally, PC2 exhibits a
moderate negative loading concerning nonanoic acid and dihydro-3-
hydroxy-4,4-dimethyl-2(3H)-furanone. PC2 discriminates between the
samples based on both their storage time and the fortification type.
Considering the first two PCs, the fresh samples exhibit a cohesive
cluster, characterized by negative loadings of PC1 and PC2. In contrast,
the stored P and O samples closely resemble the references, while S and
Table 3
Hardness (nˆ5), color (CIEL*a*b*) (nˆ5), oxidative stability index (OSI) (nˆ2) and α-linolenic acid (ALA, 18:3n-3) content (nˆ2) of the prepared chocolate and cookie
models.
Food model Sample type a Hardness [N] L* a* b* ΔE OSI [h] b ALA [g per 100 g]b
Chocolate model R 61.7  2.6a 43.9  3.5a  5.3  1.8a 35.8  1.4b - 77.0  1.4b 0.11  0.01a
O 47.4  0.6b 59.4  6.2b,c  3.5  0.2a,b 23.0  4.3a 81.2 72.5  0.7a 2.38  0.07b
S 76.5  0.4d 65.1  4.0c  1.2  0.3b 25.5  2.3a 84.3 70.1  0.3a 2.84  0.08b
P 73.7  2.8c,d 65.5  0.6c  1.6  0.2b 24.6  2.8a 84.9 80.8  0.4b,c 2.36  0.04b
W 68.5  0.6b,c 55.8  0.8b  1.1  0.4b 29.5  0.8a,b 76.2 82.0  1.4c 2.35  0.02c
Average 65.5 ± 11.0 57.9 ± 8.8 ¡2.5 ± 1.8 27.7 ± 5.1 - 76.5 ± 4.9 -
Cookie model R 10.6  1.9a 83.6  0.1b,c  7.2  0.1c 31.6  0.6a - 15.3  0.5b 3.64  0.14a
O1 16.0  4.0a,b 83.9  0.3c  7.7  0.1a,b,c 34.1  0.8a 1.8 15.0  0.0b 4.73  0.18c
O2 11.9  2.7a,b 84.5  0.6c  8.1  0.1a 34.8  1.2a 8.0 11.0  0.4a 8.16  0.02d
S1 14.1  1.1a,b 84.3  0.6c  7.5  0.1b,c 34.6  2.7a 3.1 16.9  0.2c 4.37  0.06b
S2 20.3  2.4b,c 83.6  0.7b,c  7.3  0.2c 36.2  2.3a 4.7 15.8  0.3b,c 8.55  0.16e
P1 12.2  2.3a,b 84.7  0.6c  7.5  0.1b,c 33.2  1.4a 2.0 16.5  0.5b,c 4.78  0.04c
P2 27.0  7.4c 79.7  1.4a  7.2  0.5c 48.8  3.3c 17.7 16.1  0.2b,c 8.26  0.09d
W1 20.2  2.8b,c 84.1  1.0c  7.9  0.2a,b 35.9  1.9a 4.4 16.1  0.1b,c 4.41  0.04b
W2 30.1  6.3c 82.0  0.8b  7.9  0.2a,b 42.4  2.4b 11.0 16.1  0.7b,c 8.09  0.17d
Average 17.4 ± 7.3 83.4 ± 1.6 ¡7.6 ± 0.3 36.8 ± 5.4 - 15.4 ± 1.7 -
a Rˆ not fortified food model, Oˆ fortified with non-encapsulated LO, Wˆ fortified withWM SDEMs, Sˆ fortified with SM SDEMs, Pˆ fortified with PM SDEMs, 1
ˆ fortified with SDEM or non-encapsulated linseed oil to gain 0.6 g ALA per 100 g and 100 kcal, or 2 ˆ fortified to gain 1.2 g ALA per 100 g and 100 kcal.
b Different letters (a-c) on the same column separately for each food model indicate a statistically significant (p < 0.05) difference between samples.
S.A. Hakanen et al. LWT 216 (2025) 117305 
5 
W exhibit more deviations from the Rs. Based on loadings from PC-1 the
ascending order regarding volatiles secondary oxidation compounds
was R, O, P, S and W. In our previous publication with fish oil (Damerau
et al., 2022b) the rising order was R, O, P, W and S. The higher oxidative
stability of LO, differences in FA compositions of the oils and differences
in EE could contribute to different order regarding formation of volatile
oxidation products during storage. The results of volatile oxidation
products were in line with PV results but different from OSI result (see
3.2.2.). The OSI is most likely more affected by the food matrix itself
than the PV and volatile oxidation product formation as also other
compounds than lipids like proteins contribute to a total oxidation
measurement.
In the PCA (Fig. 3C and D), most of the cookie samples exhibited
compact clustering, with only a few extreme samples such as O2_3,
O2_4, S2_3, and S2_4. Notably, most of the volatiles aligned positively
with PC1 (with 60% of the total variation). This was comprehensive as
PC1 primarily distinguished samples based on their storage period.
Fresh samples clustered towards the negative end of PC1 axis, while
stored cookies exhibit a tendency toward the positive end. PC2 (with 12
% of variation) primarily discriminated between the cookies fortified
with higher concentration of SDEMs and the remaining samples grouped
closely together in moderately distinct subclusters. Based on the PCA,
the oxidative stability order (lowest to highest) was R, P, O, W, and S for
lower concentration and R, P, W, O, and S for the higher one. A direct
comparison cannot be made with our previous publication on fortifi-
cation with fish oil (Damerau et al., 2022a) because here vegan cookies
were prepared. However, in both studies, the cookie model fortified with
SDEM with pea protein exhibited the lowest formation of volatile
compounds during storage compared to models with other SDEMs. In
both studies an increased fortification concentration had a negative ef-
fect on the oxidative stability. Compared to chocolate, the results of
cookie models regarding oxidative stability based on volatile secondary
oxidation products were not in line with either PV or OSI. This could be
based on those differences in measurement as discussed in 3.2.2.
Interestingly, the Omodels with direct oil additions performed better
for fish oil (Damerau et al., 2022a) than for LO regarding oxidative
stability. This difference may be attributed to higher EE for LO SDEMs
compared to fish oil ones and dissimilarities in FA composition, which
resulted in a better oxidative stability and less volatile formation.
Overall, P samples resembled most of the R samples as fresh and stored
Fig. 2. Peroxide values of chocolate (A.) and cookie (B.) models. For chocolate, the storage period of 0 day is marked in light grey, 36 days in grey, and 77 days in
dark grey. Corresponding storage periods for cookies were 0 day, 15 days, and 25 days. Different letters indicate significant difference at p < 0.05, with capital letters
(A–E) indicating distinctions between samples at a specific timepoint, and lowercase letters (a–c) denoting variations between timepoints for a specific sample. R ˆ
not fortified food model, O ˆ fortified with non-encapsulated linseed oil, W ˆ whey protein concentrate, S ˆ soy protein, P ˆ pea protein as shell material. The
number refers to fortification concentration of SDEM or non-encapsulated linseed oil to gain either 1 ˆ 0.6 g or 2 ˆ 1.2 g ALA per 100 g and 100 kcal.
S.A. Hakanen et al. LWT 216 (2025) 117305 
6 
Table 4
The retention times and matches with NIST library of the volatile compounds identified in the chocolate and cookie models, and the peak area magnitudes in fresh
references (R_0) and in the most oxidized samples in the final timepoint, 4, in both chocolate (W_4) and cookie models (S2_4). Sample codes refer to Table 1.
Match with NIST20 Chocolate areaa Cookie areaa
No Compound RT (min) chocolate cookie R_0 W_4b R_0 S2_4b
acids
1 formic acid 15.59 ​ 905 ​ ​ ​ ‡‡
2 acetic acid 17.18 925 959 ‡‡‡ ‡‡‡ ‡‡ ‡‡‡
3 propanoic acid 21.80 946 ​ ‡‡‡ ‡‡‡ ‡‡ ‡‡‡
4 2-methylpropanoic acid 24.42 951 ​ ‡‡‡ ‡‡‡ ​ ​
5 pentyl ester formic acid 25.42 ​ 910 ​ ​ ​ ‡‡‡
6 butanoic acid 25.75 934 965 ‡ ‡‡ ‡‡ ‡‡‡
7 2,2-dimethylpropanoic acid 26.21 863 911 ‡ ​ ‡‡ ‡‡‡
8 3-methylbutanoic acid 28.16 943 905 ‡‡‡ ‡‡‡ ​ ‡‡‡
9 2-methylbutanoic acid 28.46 913 815 ‡‡‡ ‡‡‡ ​ ‡‡
10 pentanoic acid 29.88 852 931 ‡‡ ‡‡‡ ‡‡ ‡‡‡
11 2-pentenoic acid 31.78 ​ 882 ​ ​ ​ ‡‡‡
12 4-methylpentanoic acid 32.61 908 ​ ‡‡ ‡‡ ​ ​
13 hexanoic acid 33.73 952 955 ‡‡‡ ‡‡‡ ‡‡ ‡‡‡
14 heptanoic acid 37.27 ​ 808 ​ ​ ‡ ‡‡‡
15 octanoic acid 40.73 841 923 ‡‡ ‡‡ ‡ ‡‡‡
16 2-phenylethyl ester acetic acid 43.48 865 ​ ‡‡ ‡‡ ​ ​
17 nonanoic acid 43.98 921 903 ‡‡ ​ ‡ ‡‡
alcohols
18 ethanol 8.17 867 935 ‡‡ ‡‡‡ ‡ ‡‡‡
19 2-propanol 9.80 899 903 ‡ ‡‡ ‡ ‡
20 propanol 12.99 949 ​ ‡ ‡ ​ ​
21 1,3-butanediol 14.86 805 792 ‡ ‡ ​ ‡
22 butanol 18.80 889 ​ ‡‡ ‡‡‡ ​ ​
23 1-penten-3-ol 19.31 918 930 ‡ ‡‡‡ ‡ ‡‡‡
24 pentanol 23.76 ​ 937 ‡‡ ‡‡‡ ‡‡ ‡‡‡
25 2-penten-1-ol 24.07 831 ​ ‡ ‡‡‡ ​ ​
26 2,3-butanediol 26.93 904 ​ ‡ ‡‡ ​ ​
27 2-furanmethanol 28.84 ​ 900 ​ ​ ‡ ‡‡‡
28 3-octen-1-ol (Z) 34.51 ​ 810 ​ ​ ‡ ‡‡‡
29 phenylethyl alcohol 39.72 888 ​ ‡‡ ‡‡ ​ ​
aldehydes
30 acetaldehyde 5.80 903 917 ‡‡ ‡‡‡ ‡‡ ‡‡‡
31 2-propenal 8.84 ​ 952 ​ ​ ‡ ‡‡
32 propanal 9.13 966 946 ‡ ‡‡‡ ​ ‡‡‡
33 2-methylpropanal 12.01 956 960 ‡ ‡‡ ​ ‡‡
34 2-methyl-2-propenal 12.74 960 ​ ​ ‡‡ ​ ​
35 butanal 13.88 952 966 ‡ ‡‡ ​ ‡‡‡
36 3-methylbutanal 17.16 ​ 955 ​ ​ ‡‡ ‡‡‡
37 2-methylbutanal 17.68 917 911 ‡ ‡‡ ‡‡ ‡‡‡
38 2-butenal 17.69 ​ 913 ​ ​ ​ ‡‡‡
39 pentanal 19.51 890 874 ‡‡ ‡‡‡ ‡‡ ‡‡‡
40 2-pentenal 23.19 892 923 ​ ‡‡ ​ ‡‡‡
41 hexanal 24.70 967 972 ‡‡‡ ‡‡‡ ‡‡‡ ‡‡‡
42 2-methyl-2-pentenal 26.32 944 934 ‡‡ ‡‡‡ ​ ‡‡‡
43 furfural 27.59 ​ 946 ​ ​ ‡ ‡‡‡
44 2-hexenal 27.94 912 927 ​ ‡‡ ‡ ‡‡‡
45 heptanal 29.39 871 940 ‡‡ ‡‡‡ ‡‡ ‡‡‡
46 methional 30.81 ​ 900 ​ ​ ​ ‡‡
47 2,4-hexadienal 30.92 886 895 ‡ ‡‡ ​ ‡‡‡
48 2-heptenal (E) 31.20 ​ 864 ​ ​ ​ ‡‡
49 2-heptenal (Z) 32.50 899 957 ‡ ‡‡ ‡‡ ‡‡‡
50 benzaldehyde 33.21 903 922 ‡‡‡ ‡‡‡ ‡‡ ‡‡‡
51 octanal 33.65 954 974 ‡‡ ‡‡ ‡ ‡‡‡
52 2,4-heptadienal (E,E) 34.72 862 920 ‡‡ ‡‡‡ ‡‡ ‡‡‡
53 2-octenal (E) 36.50 ​ 962 ​ ​ ‡‡‡ ‡‡‡
54 nonanal 37.55 875 942 ‡‡‡ ‡‡‡ ‡‡‡ ‡‡‡
55 4-oxohex-2-enal 37.93 905 897 ​ ‡‡‡ ‡ ‡‡‡
56 2,6-nonadienal 40.13 ​ 942 ​ ​ ‡ ‡‡
57 2-nonenal (E) 40.20 ​ 942 ​ ​ ‡‡ ‡‡
58 decanal 41.06 ​ 892 ​ ​ ‡‡ ‡‡‡
59 2,4-nonadienal 42.47 ​ 907 ​ ​ ​ ‡‡
60 2-decenal (E) 43.71 ​ 958 ​ ​ ‡ ‡‡
61 2,4-decadienal (E,E) 46.17 ​ 931 ​ ​ ‡‡ ‡‡
62 2-undecenal 47.56 ​ 905 ​ ​ ‡ ‡‡
esters
63 methyl acetate 10.31 845 877 ‡ ‡‡ ‡ ‡
64 2-methyl-1-butyl acetate 24.22 ​ 798 ​ ​ ‡‡ ‡‡‡
65 ethyl octanoate 39.96 840 ​ ‡‡ ‡ ​ ​
heterocyclic compounds
66 furan 8.57 ​ 918 ​ ​ ‡ ‡‡
67 2-methylfuran 13.45 895 911 ​ ‡ ‡ ‡‡
(continued on next page)
S.A. Hakanen et al. LWT 216 (2025) 117305 
7 
for both food models. The P samples also performed similar or better
than R samples in terms of oxidative stability by PV or OSI although PM
had the lowest EE of all SDEMs. This shows again that the quality of
SDEMs does not directly transfer to performances of an SDEM in the food
models. More knowledge on interactions of wall material of SDEMs and
food matrix is needed to predict the performance of SDEMs in food
fortification.
3.3. Sensory quality of food models
3.3.1. The performance of the assessors and the panels
Both panels effectively discriminated between the samples
(Supplementary Tables 3 and 4). Strong consensus among the appear-
ance and the texture attributes contributed to effective discrimination
and repeatability. Certain assessors encountered challenges with some
odor or flavor attributes, but because omitting any participants did not
noticeably influence the results, all assessors were retained in analyses.
Table 4 (continued )
Match with NIST20 Chocolate areaa Cookie areaa
No Compound RT (min) chocolate cookie R_0 W_4b R_0 S2_4b
68 2-ethylfuran 18.57 941 943 ‡ ‡‡‡ ​ ‡‡‡
69 dihydro-2-methyl-3(2H)-furanone 25.49 ​ 916 ​ ​ ​ ‡‡‡
70 methylpyrazine 25.86 850 915 ‡ ‡‡ ‡ ‡‡‡
71 2,5-dimethylpyrazine 29.69 ​ 825 ​ ​ ‡ ‡‡‡
72 2,3-dimethylpyrazine 30.22 905 ​ ‡ ‡‡ ​ ​
73 acetylfuran 30.89 ​ 882 ​ ​ ​ ‡‡‡
74 2-pentylfuran 32.18 919 936 ‡‡ ‡‡‡ ‡‡ ‡‡‡
75 2-(2-pentenyl)furan 32.72 ​ 877 ​ ​ ‡‡ ‡‡‡
76 2(5H)-furanone 33.36 ​ 922 ​ ​ ‡‡ ‡‡‡
77 trimethylpyrazine 16.74 888 ​ ‡‡ ‡‡ ​ ​
78 5-methyl-2(5H)-furanone 34.10 ​ 919 ​ ​ ​ ‡‡‡
79 tetramethylpyrazine 36.56 954 ​ ‡‡‡ ‡‡‡ ​ ​
80 linalool oxide 37.19 830 ​ ‡‡ ‡‡ ​ ​
81 2-acetylpyrrole 37.95 860 ​ ‡‡ ​ ​ ​
82 5-ethyldihydro-2(3H)-furanone 38.54 903 942 ​ ‡‡ ‡ ‡‡‡
83 dihydro-3-hydroxy-4,4-dimethyl-2(3H)-furanone 38.70 883 ​ ‡‡ ‡‡ ​ ​
84 tetrahydro-6-pentyl-2H-pyran-2-one 39.54 ​ 847 ​ ​ ‡ ‡‡
85 maltol 39.74 ​ 962 ​ ​ ‡‡ ‡‡‡
86 tetrahydro-6-methyl-2H-pyran-2-one 40.40 ​ 945 ​ ​ ​ ‡‡
87 dihydro-5-propyl-2(H)-furanone 42.12 ​ 867 ​ ​ ‡ ‡
88 5-butyldihydro-2(3H)-furanone 45.86 ​ 857 ​ ​ ​ ‡‡
89 tetrahydro-6-propyl-2H-pyran-2-one 47.36 ​ 919 ​ ​ ​ ‡‡
90 dihydro-5-pentyl-2(3H)-furanone 50.29 ​ 879 ​ ​ ‡ ‡
hydrocarbons
91 pentane 7.84 862 918 ‡ ‡‡ ‡ ‡‡
92 2-methylpentane 10.60 950 903 ​ ​ ‡ ​
93 3-methylpentane 11.34 937 921 ‡ ​ ‡‡‡ ‡‡
94 heptane 17.42 ​ 946 ​ ​ ‡‡‡ ‡‡‡
95 1-nonene 27.06 ​ 900 ​ ​ ​ ‡‡
96 2,4-nonadiene 30.30 ​ 786 ​ ​ ​ ‡‡‡
97 1-decene 31.39 ​ 891 ​ ​ ​ ‡‡
98 decane 31.59 936 ​ ‡ ‡‡ ​ ​
99 2,6-dimethyldecane 35.34 901 ​ ‡‡ ‡‡‡ ​ ​
100 undecane 35.47 ​ 932 ​ ​ ‡‡ ‡‡
101 dodecane 36.39 890 ​ ‡‡ ‡‡ ​ ​
ketones
102 2,3-butanedione 14.04 883 ​ ‡ ‡‡ ​ ​
103 2-butanone 14.44 941 952 ‡ ‡‡‡ ​ ‡‡‡
104 1-penten-3-one 19.05 892 870 ​ ‡‡ ​ ‡‡‡
105 2-pentanone 19.19 980 978 ‡ ‡‡ ​ ‡‡‡
106 3-penten-2-one 22.32 ​ 944 ​ ​ ​ ‡‡‡
107 2-hydroxy-3-pentanone 26.45 873 819 ​ ‡‡ ‡ ‡‡‡
108 3-hexen-2-one 27.19 ​ 915 ​ ​ ​ ‡‡‡
109 2-heptanone 29.02 798 917 ‡‡ ‡‡ ‡ ‡‡‡
110 3-hydroxy-3-methylbutanone 30.64 ​ 869 ​ ​ ​ ‡‡
111 6-methyl-5-hepten-2-one 33.09 ​ 876 ​ ​ ‡‡ ‡‡‡
112 2-octanone 33.21 ​ 839 ​ ​ ‡‡ ‡‡‡
113 3-octen-2-one 35.59 ​ 905 ​ ​ ‡ ‡‡‡
114 4-nonanone 36.17 ​ 861 ​ ​ ​ ‡‡
115 3,5-octadien-2-one 37.48 924 919 ‡‡ ‡‡‡ ‡ ‡‡‡
116 2-nonanone 37.13 873 932 ‡‡ ‡‡ ‡ ‡‡
terpenes
117 pinene, alpha- 29.79 938 ​ ‡‡‡ ‡‡‡ ​ ​
118 pinene, beta- 32.00 869 ​ ‡‡ ‡‡ ​ ​
119 3-carene 33.07 949 ​ ‡‡‡ ‡‡‡ ​ ​
120 D-limonene 33.79 926 849 ‡‡‡ ‡‡‡ ‡‡ ‡
121 linalool 37.25 821 ​ ‡‡ ‡‡ ​ ​
a Samples marked with "‡‡‡" belong to the top 75th percentile, samples with "‡‡" belong to the 25th to 75th percentile, and samples with "‡" belong to the bottom
25th percentile when considering all volatile compounds across all samples of either the chocolate or cookie model.
b The sample exhibiting the greatest dissimilarity from the reference sample based on the PCA model (Fig. 3).
S.A. Hakanen et al. LWT 216 (2025) 117305 
8 
3.3.2. Sensory profile of dark chocolate model
The first principal component (Fig. 4A and B; PC1 with explained
variance of 80%) separates the samples into two distinct groups ac-
cording to their storage time. PC1 is positively loaded with meltability,
smoothness, brownness, and sweetness, which are characteristic of the
fresh samples. In contrast, the stored samples, negatively loaded on PC1,
are associated with greyness, hardness, and uneven surface color.
Moreover, significant differences were observed in these attributes be-
tween fresh and stored samples (Supplementary Table 5). However,
Razavizadeh and Tabrizi (2021) observed no statistical differences in the
liking of sensory characteristics of milk chocolate fortified with chia
seed oil encapsulated with soy protein, maltodextrin, and inulin,
possibly due to the panelists’ inability to discriminate the samples. The
inconsistency may be due to the higher core to wall ratio of SDEMs in
Fig. 3. PCAs of identified volatile compounds in chocolate (A: samples and B: volatiles) and cookie (C: samples and D: volatiles). R ˆ not fortified food model, O ˆ
fortified with non-encapsulated LO, W ˆ fortified with WM SDEMs, S ˆ fortified with SM SDEMs, P ˆ fortified with PM SDEMs, 1 ˆ fortified with SDEM or non-
encapsulated linseed oil to gain 0.6 g ALA per 100 g and 100 kcal, or 2 ˆ fortified to gain 1.2 g ALA per 100 g and 100 kcal. The number following the underscore
refers to the timepoint.
Fig. 4. Consensus PCA scores and correlation loadings plots of the descriptive sensory profiles of dark chocolate (A: samples and B: sensory attributes) and cookie (C:
samples and D: sensory attributes) models. In scores plots: R ˆ not fortified food model (black), O ˆ fortified with non-encapsulated LO (blue), W ˆ whey protein
concentrate (red), S ˆ soy protein (yellow), P ˆ pea protein (green) as shell material. In 4C, the first number refers to fortification concentration of SDEM or non-
encapsulated linseed oil either 1 ˆ to gain 0.6 g ALA per 100 g and 100 kcal (circles) or 2 ˆ to gain 1.2 g ALA per 100 g and 100 kcal (triangles). The number
following the underscore refers to the timepoint.
S.A. Hakanen et al. LWT 216 (2025) 117305 
9 
our study leading to noticeable sensory differences.
The influence of shell material was more evident in the stored sam-
ples (Supplementary Table 5). Sample P was rated as harder, smoother,
and less grainy with a more uneven surface color than R, whereas they
did not differ in the odor or the flavor attributes. At the same time, S and
W exhibited closer resemblances to R in appearance and texture, but
they diverged in the flavor attributes. The S and W were rated signifi-
cantly higher in the LO flavor, the rancid flavor, and the musty flavor
compared to P and R. Additionally, rancidity significantly increased in S
and W during storage, a change not observed in the other samples.
Taken together, the pea protein offers more effective protection against
lipid oxidation and exhibits better flavor masking capabilities in choc-
olate compared to soy or whey proteins, which is in accordance with our
previous research (Damerau et al., 2022a). However, soy protein
emerged also as a suitable encapsulation material in the previous study.
3.3.3. Sensory profile of shortbread cookie model
In the PCA model (Fig. 4C and D), the samples are clustered in four
groups based on the oil/SDEM concentration and the storage time. PC1
primarily separates the cookies in accordance with the concentration.
The cookies fortified with lower oil/SDEM concentration are collectively
linked with attributes such as crumbly, sweet, and dark spotted.
Conversely, the samples with higher concentration are typified as hard,
uneven color, crispy, roasted flavor, and dark. The results seem to be
consistent with the findings of our previous fish oil study (Damerau
et al., 2022a). The PC2 further distinguished between the two clusters,
separating the samples into fresh and stored ones. Fresh chocolates
correlated with sweet odor and moderately with nutty odor, while the
stored were more associated with fatty mouthfeel and rancid odor.
Rancidity increased significantly in fortified cookies due to storage
time and oil/SDEM concentration (Supplementary Table 6). Highest
rancidity was observed in stored S2 and W2, respectively. Among the
encapsulation materials, pea protein showed the most effective protec-
tion against rancid flavor and odor, followed by whey and soy, respec-
tively. The results appear to corroborate our earlier findings in Damerau
et al. (2022a) where flavor masking capacity was most prominent in pea
protein samples.
4. Conclusions
Encapsulation by spray-drying, utilizing pea, soy, and whey protein
in conjunction with maltodextrin, significantly improved the oxidative
stability of LO capsules without significant loss of the n-3 PUFAs. The
most effective protection against oxidation was achieved by combining
whey protein with maltodextrin. However, the versatile multidisci-
plinary approach used in our study, spanning from chemical analyses to
sensory evaluations, demonstrated that this behaviour did not translate
to their use in high fat food models, dark chocolate and shortbread
cookie. Volatile profiles exhibited distinguishable differences based on
fortification concentration and encapsulation materials in both food
models. Storage led to an increase in volatiles, particularly pronounced
in soy and whey containing samples. Food models fortified with pea
protein capsules demonstrated a closer resemblance to the control than
the other spray-dried samples. Increasing oil content had a detrimental
effect on oxidative stability. Surprisingly, the volatile profile of sample
with non-encapsulated LO resembled closely that of the control sample
in both food models.
The addition of LO or spray-dried emulsions in chocolate increased
the graininess and decreased the meltability of chocolate models. The
chocolate flavor decreased in soy and whey samples during storage
while LO, musty and rancid flavors increased. The pea sample resembled
most closely to the control sample. Interestingly, soy exhibited most
similar appearance and texture properties to the control sample.
Encapsulation influenced the appearance, odor, flavor, and texture of
both fresh and stored cookies. Especially, the SDEM addition increased
darkness, rancid flavor, roasted flavor, and hardness. Similarly, as in
chocolate samples, the pea sample was the closer to the control sample
than other encapsulation samples.
While certain hypothesized plant protein formulations demonstrated
promise, non-encapsulated LO fortification also yielded favorable out-
comes based on sensory perception and volatile profiles in these high-fat
food models. Pea protein emerged as an intriguing choice for shell ma-
terial in spray-drying of LO. Further research is necessary to elucidate its
performance in different foods with varying fat and water content.
CRediT authorship contribution statement
Sari A. Hakanen: Writing – review & editing, Writing – original
draft, Visualization, Supervision, Methodology, Investigation, Formal
analysis. Annelie Damerau: Writing – review & editing, Writing –
original draft, Supervision, Conceptualization. Dorota Ogrodowska:
Writing – review & editing, Writing – original draft, Visualization,
Investigation, Conceptualization. Annalisa Seubert:Writing – review&
editing, Investigation. Waldemar Brandt: Resources, Methodology.
Oskar Laaksonen: Writing – review & editing, Supervision, Method-
ology. Małgorzata Tanska: Writing – review & editing, Writing –
original draft, Investigation. Kaisa M. Linderborg: Writing – review &
editing, Supervision, Resources, Project administration, Funding
acquisition, Conceptualization.
Declaration of generative AI and AI-assisted technologies in the
writing process
During the preparation of this work the authors used ChatGPT 3.5 in
order to improve readability and language. After using this service, the
authors reviewed and edited the content as needed and take full re-
sponsibility for the content of the published article.
Funding
The work was carried out with financial support from the Academy
of Finland project “Omics of oxidation – Solutions for a better quality of
docosahexaenoic and eicosapentaenoic acids” (Decision No. 315274).
Personal grants to Sari A. Hakanen from the Turku University founda-
tion, the Finnish Cultural Foundation, and the Magnus Ehrnrooth
foundation are also acknowledged. Further, personal financial grant to
Annelie Damerau from the Finnish Cultural Foundation is
acknowledged.
Declaration of competing interest
The authors declare that they have no known competing financial
interests or personal relationships that could have appeared to influence
the work reported in this paper.
Acknowledgements
The authors would like to thank the participants of the sensory
evaluation.
Appendix A. Supplementary data
Supplementary data to this article can be found online at https://doi.
org/10.1016/j.lwt.2024.117305.
Data availability
Data will be made available on request.
S.A. Hakanen et al. LWT 216 (2025) 117305 
10 
References
Aberkane, L., Roudaut, G., & Saurel, R. (2014). Encapsulation and oxidative stability of
PUFA-rich oil microencapsulated by spray drying using pea protein and pectin. Food
and Bioprocess Technology, 7(5), 1505–1517. https://doi.org/10.1007/S11947-013-
1202-9/FIGURES/7
Afoakwa, E. O., Paterson, A., Fowler, M., & Ryan, A. (2009). Matrix effects on flavour
volatiles release in dark chocolates varying in particle size distribution and fat
content using GC–mass spectrometry and GC–olfactometry. Food Chemistry, 113(1),
208–215. https://doi.org/10.1016/j.foodchem.2008.07.088
Bajaj, P. R., Tang, J., & Sablani, S. S. (2015). Pea protein isolates: Novel wall materials
for microencapsulating flaxseed oil. Food and Bioprocess Technology, 8(12),
2418–2428. https://doi.org/10.1007/s11947-015-1589-6
Baker, M. T., Lu, P., Parrella, J. A., & Leggette, H. R. (2022). Consumer acceptance
toward functional foods: A scoping review. International Journal of Environmental
Research and Public Health, 19(3), 1217. https://doi.org/10.3390/ijerph19031217
Bork, C. S., Lundbye-Christensen, S., Stine, Venø, K., Lasota, A. N., Schmidt, E. B., &
Overvad, K. (2022). Plant n-3 PUFA intake may lower the risk of atherosclerotic
cardiovascular disease only among subjects with a low intake of marine n-3 PUFAs.
European Journal of Nutrition, 61, 557–559. https://doi.org/10.1007/s00394-021-
02581-5
Brühl, L., Matthaus, B., Fehling, E., Wiege, B., Lehmann, B., Luftmann, H., …
Hofmann, T. (2007). Identification of bitter off-taste compounds in the stored cold
pressed linseed oil. Journal of Agricultural and Food Chemistry, 55(19), 7864–7868.
https://doi.org/10.1021/JF071136K/ASSET/IMAGES/LARGE/JF-2007
-01136K_0005.JPEG.
Carneiro, H. C. F., Tonon, R. V., Grosso, C. R. F., & Hubinger, M. D. (2013). Encapsulation
efficiency and oxidative stability of flaxseed oil microencapsulated by spray drying
using different combinations of wall materials. Journal of Food Engineering, 115(4),
443–451. https://doi.org/10.1016/J.JFOODENG.2012.03.033
Chang, C., & Nickerson, M. T. (2018). Encapsulation of omega 3-6-9 fatty acids-rich oils
using protein-based emulsions with spray drying. Journal of Food Science and
Technology, 55(8), 2850–2861. https://doi.org/10.1007/s13197-018-3257-0
Christie, W. W. (2010). Lipid analysis: Isolation, separation, identification, and structural
analysis of lipids (4th ed.). The Oily Press.
Damerau, A., Mustonen, S. A., Ogrodowska, D., Varjotie, L., Brandt, W., Laaksonen, O.,
… Linderborg, K. M. (2022a). Food fortification using spray-dried emulsions of fish
oil produced with maltodextrin, plant and whey proteins - effect on sensory
perception, volatiles and storage stability. Molecules, 27(11), 3553. https://doi.org/
10.3390/MOLECULES27113553
Damerau, A., Ogrodowska, D., Banaszczyk, P., Dajnowiec, F., Tanska, M., &
Linderborg, K. M. (2022b). Baltic herring (Clupea harengus membras) oil
encapsulation by spray drying using a rice and whey protein blend as a coating
material. Journal of Food Engineering, 314, Article 110769. https://doi.org/10.1016/
j.jfoodeng.2021.110769
Delshadi, R., Bahrami, A., Tafti, A. G., Barba, F. J., & Williams, L. L. (2020). Micro and
nano-encapsulation of vegetable and essential oils to develop functional food
products with improved nutritional profiles. Trends in Food Science and Technology,
104, 72–83. https://doi.org/10.1016/j.tifs.2020.07.004
European Parliament, & Council. (2007). Corrigendum to Regulation (EC) No 1924/
2006 of the European Parliament and of the Council of 20 December 2006 on
nutrition and health claims made on foods. Official Journal of the European Union, 12
(3), 16. https://eur-lex.europa.eu/legal-content/EN/ALL/?uriˆCELEX:02006R1924
-20141213.
Folch, J., Lees, M., & Stanley, G. H. S. (1957). A simple method for the isolation and
purification of total lipides from animal tissues. Journal of Biological Chemistry, 226
(1), 497–509. https://doi.org/10.1016/s0021-9258(18)64849-5
Francisco, C. R. L., de Oliveira Júnior, F. D., Marin, G., Alvim, I. D., & Hubinger, M. D.
(2020). Plant proteins at low concentrations as natural emulsifiers for an effective
orange essential oil microencapsulation by spray drying. Colloids and Surfaces A:
Physicochemical and Engineering Aspects, 607, Article 125470. https://doi.org/
10.1016/J.COLSURFA.2020.125470
Frankel, E. N. (2012). Methods to determine extent of oxidation. In E. N. Frankel (Ed.),
Lipid oxidation (2nd ed., pp. 99–127). Woodhead Publishing. https://doi.org/
10.1533/9780857097927.99.
Gomes, M. H. G., & Kurozawa, L. E. (2021). Influence of rice protein hydrolysate on lipid
oxidation stability and physico-chemical properties of linseed oil microparticles
obtained through spray-drying. LWT, 139, Article 110510. https://doi.org/10.1016/
j.lwt.2020.110510
Gomez-Cortes, P., Sacks, G. L., & Brenna, J. T. (2015). Quantitative analysis of volatiles
in edible oils following accelerated oxidation using broad spectrum isotope
standards. Food Chemistry, 174, 310–318. https://doi.org/10.1016/J.
FOODCHEM.2014.11.015
Gong, K.-J., Shi, A.-M., Liu, H.-Z., Liu, L., Hu, H., Adhikari, B., et al. (2016). Emulsifying
properties and structure changes of spray and freeze-dried peanut protein isolate.
Journal of Food Engineering, 170, 33–40. https://doi.org/10.1016/j.
jfoodeng.2015.09.011
Guesnet, P., Tressou, J., Buaud, B., Simon, N., & Pasteau, S. (2019). Inadequate daily
intakes of n-3 polyunsaturated fatty acids (PUFA) in the general French population
of children (3–10 years) and adolescents (11–17 years): The INCA2 survey. European
Journal of Nutrition, 58(2), 895–903. https://doi.org/10.1007/s00394-018-1694-1
Hadnađev, M., Kalic, M., Krstonosic, V., Jovanovic-Ljeskovic, N., Erceg, T., Skrobot, D., &
Dapcevic-Hadnađev, T. (2023). Fortification of chocolate with microencapsulated
fish oil: Effect of protein wall material on physicochemical properties of
microcapsules and chocolate matrix. Food Chemistry X, 17, Article 100583. https://
doi.org/10.1016/j.fochx.2023.100583
Kajla, P., Sharma, A., & Sood, D. R. (2015). Flaxseed—a potential functional food source.
Journal of Food Science and Technology, 52(4), 1857–1871. https://doi.org/10.1007/
s13197-014-1293-y
Le Priol, L., Dagmey, A., Morandat, S., Saleh, K., El Kirat, K., & Nesterenko, A. (2019).
Comparative study of plant protein extracts as wall materials for the improvement of
the oxidative stability of sunflower oil by microencapsulation. Food Hydrocolloids,
95, 105–115. https://doi.org/10.1016/J.FOODHYD.2019.04.026
Lehtonen, M., Lampi, A. M., Ollilainen, V., Struijs, K., & Piironen, V. (2011). The role of
acyl moiety in the formation and reactions of steryl ester hydroperoxides. European
Food Research and Technology, 233(1), 51–61. https://doi.org/10.1007/s00217-011-
1492-y
Mozuraityte, R., Kristinova, V., Standal, I. B., Carvajal, A. K., & Aursand, M. (2016).
Oxidative stability and shelf life of fish oil. In M. Hu, & C. Jacobsen (Eds.), Oxidative
stability and shelf life of foods containing oils and fats (pp. 209–231). AOCS Press.
https://doi.org/10.1016/B978-1-63067-056-6.00005-7.
Mueed, A., Shibli, S., Korma, S. A., Madjirebaye, P., Esatbeyoglu, T., & Deng, Z. (2022).
Flaxseed bioactive compounds: Chemical composition, functional properties, food
applications and health benefits-related gut microbes. Foods, 11(20), 3307. https://
doi.org/10.3390/foods11203307
Ogrodowska, D., Damerau, A., Banaszczyk, P., Tanska, M., Konopka, I. Z., Piłat, B.,
Dajnowiec, F., & Linderborg, K. M. (2024). Native and pregelatinized potato and rice
starches and maltodextrin as encapsulating agents for linseed oil ethyl esters –
comparison of emulsion and powder properties. Journal of Food Engineering, 364,
Article 111799. https://doi.org/10.1016/j.jfoodeng.2023.111799
Ogrodowska, D., Tanska, M., & Brandt, W. (2017). The influence of drying process
conditions on the physical properties, bioactive compounds and stability of
encapsulated pumpkin seed oil. Food and Bioprocess Technology, 10(7), 1265–1280.
https://doi.org/10.1007/s11947-017-1898-z
Perez-Palacios, T., Ruiz-Carrascal, J., Solomando, J. C., De-la-Haba, F., Pajuelo, A., &
Antequera, T. (2022). Recent developments in the microencapsulation of fish oil and
natural extracts: Procedure, quality evaluation and food enrichment. Foods, 11(20),
3291. https://doi.org/10.3390/foods11203291
Razavizadeh, B. M., & Tabrizi, P. (2021). Characterization of fortified compound milk
chocolate with microcapsulated chia seed oil. LWT, 150, Article 111993. https://doi.
org/10.1016/j.lwt.2021.111993
Redruello-Requejo, M., Samaniego-Vaesken, M. de L., Puga, A. M., Montero-Bravo, A.,
Ruperto, M., Rodríguez-Alonso, P., Partearroyo, T., & Varela-Moreiras, G. (2023).
Omega-3 and omega-6 polyunsaturated fatty acid intakes, determinants and dietary
sources in the Spanish population: Findings from the ANIBES Study. Nutrients, 15(3),
562. https://doi.org/10.3390/NU15030562
Ruiz Ruiz, J. C., Ortiz Vazquez, E. D. L. L., & Segura Campos, M. R. (2017). Encapsulation
of vegetable oils as source of omega-3 fatty acids for enriched functional foods.
Critical Reviews in Food Science and Nutrition, 57(7), 1423–1434. https://doi.org/
10.1080/10408398.2014.1002906
Sala-Vila, A., Fleming, J., Kris-Etherton, P., & Ros, E. (2022). Impact of α-linolenic acid,
the vegetable ω-3 fatty acid, on cardiovascular disease and cognition. Advances in
Nutrition, 13(5), 1584–1602. https://doi.org/10.1093/ADVANCES/NMAC016
Sioen, I., van Lieshout, L., Eilander, A., Fleith, M., Lohner, S., Szommer, A., Petisca, C.,
Eussen, S., Forsyth, S., Calder, P. C., Campoy, C., & Mensink, R. P. (2017). Systematic
review on N-3 and N-6 polyunsaturated fatty acid intake in European countries in
light of the current recommendations - focus on specific population groups. Annals of
Nutrition and Metabolism, 70(1), 39–50. https://doi.org/10.1159/000456723
Srivastava, S., & Mishra, H. N. (2021). Development of microencapsulated vegetable oil
powder based cookies and study of its physicochemical properties and storage
stability. LWT, 152. https://doi.org/10.1016/j.lwt.2021.112364
Takeungwongtrakul, S., Benjakul, S., & H-kittikun, A. (2015). Wall materials and the
presence of antioxidants influence encapsulation efficiency and oxidative stability of
micro-encapsulated shrimp oil. European Journal of Lipid Science and Technology, 117
(4), 450–459. https://doi.org/10.1002/ejlt.201400235
Tambade, P. B., Sharma, M., Singh, A. K., & Surendranath, B. (2020). Flaxseed oil
microcapsules prepared using soy protein isolate and modified starch: Process
optimization, characterization and in vitro release behaviour. Agricultural Research, 9
(4), 652–662. https://doi.org/10.1007/s40003-020-00461-8
Tanska, M., Roszkowska, B., Skrajda, M., & Dąbrowski, G. (2016). Commercial cold
pressed flaxseed oils quality and oxidative stability at the beginning and the end of
their shelf life. Journal of Oleo Science, 65(2), 111–121. https://doi.org/10.5650/jos.
ess15243
Tomic, O., Luciano, G., Nilsen, A., Hyldig, G., Lorensen, K., & Næs, T. (2010). Analysing
sensory panel performance in a proficiency test using the PanelCheck software.
European Food Research and Technology, 230(3), 497–511. https://doi.org/10.1007/
s00217-009-1185-y
Wang, Y., Ghosh, S., & Nickerson, M. T. (2022). Microencapsulation of flaxseed oil by
lentil protein isolate-κ-carrageenan and-ι-carrageenan based wall materials through
spray and freeze drying. Molecules, 27(10), 3195. https://doi.org/10.3390/
molecules27103195
Wang, J., Han, L., Wang, D., Sun, Y., Huang, J., & Shahidi, F. (2021). Stability and
stabilization of omega-3 oils: A review. Trends in Food Science and Technology, 118,
17–35. https://doi.org/10.1016/j.tifs.2021.09.018. Elsevier.
Wei, J., Hou, R., Xi, Y., Kowalski, A., Wang, T., Yu, Z., Hu, Y., Chandrasekar, E. K.,
Sun, H., & Ali, M. K. (2018). The association and dose–response relationship between
dietary intake of α-linolenic acid and risk of CHD: A systematic review and meta-
analysis of cohort studies. British Journal of Nutrition, 119(1), 83–89. https://doi.org/
10.1017/S0007114517003294
Yuan, Q., Xie, F., Huang, W., Hu, M., Yan, Q., Chen, Z., Zheng, Y., & Liu, L. (2022). The
review of alpha-linolenic acid: Sources, metabolism, and pharmacology.
Phytotherapy Research, 36(1), 164–188. https://doi.org/10.1002/PTR.7295
S.A. Hakanen et al. LWT 216 (2025) 117305 
11