Introduction

Cookie dough is a complex system composed mainly of flour, fat and sugar and low amounts of water. Quality characteristics of cookies mainly depend on their texture, which results from the interactions between ingredients. Fat (e.g. margarine, shortenings or butter) plays an important role in shortness, richness and tenderness, thus improving impacting the sensory quality of the product (Sudha et al. 2014).

The substitution of fat in cookies and biscuit formulations has been reported in the literature. O’Brien et al. (2003) microencapsulated hydrogenated vegetable fat by first producing an emulsion of the molten fat stabilised by whey protein isolate under high-pressure homogenisation, which was then spray-dried. The microparticles, so-called high-fat powder, were subsequently applied in short dough biscuits. Sudha et al. (2014) microencapsulated a commercially-available bakery fat, also, by spray drying, and incorporated it into biscuit dough.

In recent years, there was a trend by consumers to avoid foods containing partially hydrogenated fats, such as margarine and shortenings, due to the presence of trans fatty acids and the concerns about their function in the human body (Ullah et al. 2016). Chia (Salvia hispanica L.) seed oil is rich in n-3 and n-6 polyunsaturated fatty acids (PUFAs), and the recent consumption increase is mainly due to its enhanced nutritional value and higher sensory quality when compared to fish oil. The positive correlation between PUFA consumption and protection against several metabolic malfunctions, such as cardiometabolic diseases, is reported in the literature (Martin et al. 2016). Thus, consumers would benefit from access to a wide range of PUFA-enriched foods. Microencapsulation is a way to improve oxidative stability of the chia seed oil (de Almeida et al. 2018; Guimarães-Inácio et al. 2018; Souza et al. 2017), and also to allow its application as fat replacer.

Cookies are an easy to prepare, long shelf life foodstuff and present good acceptance by the population (Rodrigues et al. 2014). The drawback is the high amount of fat (mainly hydrogenated fat) that is usually used in the formulation. In this sense, the partial substitution of this fat by microencapsulated PUFA-rich oils could be an alternative to improve cookie quality. In fact, cookies containing inulin and erythritol (Laguna et al. 2013), hulless barley (Škrbić and Cvejanov 2011), microencapsulated grape seed extracts (Davidov-Pardo et al. 2012), microencapsulated garden cress (Lepidium sativum) oil (Umesha et al. 2015) and sour cherry pomace extract (Šaponjac et al. 2016) have been already developed.

The present work aimed to apply microencapsulated chia seed oil as a margarine substitute in cookie formulation. Proximate composition, texture, colour, fatty acid composition, physical characteristics, microstructure, hygroscopic and thermal behaviour, as well as sensory acceptability of the cookies were determined.

Material and Methods

Materials

Chia oil was purchased from Veris Óleos Vegetais Ltda. Carnauba wax (Sigma–Aldrich) was used as the encapsulating agent, and sodium caseinate (Sigma–Aldrich) was used as a stabiliser, in the microparticle preparation. Chloroform and methanol (both from Vetec, P.A.) were used to extract the oil from the cookies, and iso-octane (2,2,4-trimethylpentane) (HPLC grade, Sigma–Aldrich) was used in the lipid oxidation evaluation. Ammonium chloride (Vetec, 99.5%) and sulphuric acid (Vetec, 95%) were used in the transesterification reaction. For the cookie production, the following ingredients were acquired at a local market: wheat flour, margarine (80% lipid), sugar and baking soda. Distilled water was also used.

Chia Oil Microencapsulation

Chia seed oil-loaded microparticles were obtained according to the hot homogenisation technique described by de Almeida et al. (2018). Briefly, the aqueous phase was prepared by dissolving 55 mg sodium caseinate in 50 g water at 95 °C. Separately, carnauba wax (3.350 g) was placed in a borosilicate vessel connected to a thermostatic bath at 95 °C. Then, chia oil (1.650 g) was added under magnetic stirring for 1 min, to produce the molten mixture. Next, the aqueous phase was added to the vessel and homogenised (16,000 rpm, 5 min) using an Ultra-Turrax disperser (IKA, T25). The obtained mixture was poured into an ice bath vessel for cooling and particle formation. Finally, the dispersed particles were freeze-dried (Liotop, L101), and the resultant fat powder was stored at − 18 °C, protected from light.

Cookie Production

Cookie formulations (Table 1) were based on de Almeida et al. (2018), with minor modifications. A control formulation was prepared without microparticles added (no margarine substitution). Initially, wheat flour (40.00 g), sugar (17.92 g), salt (0.38 g), baking soda (0.90 g) and chia seed oil-loaded microparticles (in the formulations with the partial fat replacement) were mixed for 3 min. Then, margarine was added, and the mixture was homogenised for 2 min. Finally, water was added according to the dough characteristics and the water amount was adjusted in the other formulations, to achieve the same dough consistency, based on visual observation.

Table 1 Cookie formulations with the partial substitution of margarine by chia oil-loaded microparticles (M15: 15%wt substitution of margarine by microparticles; M30: 30%wt substitution of margarine by microparticles) and the control formulation (no microparticles added)
Full size table

Doughs were spread and thickness was standardised to 5 mm and cut into 25-mm diameter circles and oven-baked (Tedesco, FTT 240E) at 200 °C for 15 min. After cooling to room temperature, the cookies were packed in bi-oriented polypropylene (BOPP) plastic bags and stored at 25 °C. Bags containing 10 cookies each (five bags per formulation) were stored at 25 °C. Texture and fatty acid composition of the samples were analysed at 0 and 30 days.

Cookie Characterisation

The proximate composition of the produced cookies was established according to the method described by the Adolfo Lutz Institute (2008). Cookie samples were crushed separately and packed in polyethylene packages until the analyses. For the moisture content determination, samples were oven-dried at 105 °C until constant weight. The residual material of the moisture analysis was incinerated in a muffle at 600 °C for 12 h, and ash percentage was calculated gravimetrically. Protein content was measured by the semi-micro Kjeldahl method determining total nitrogen (factor 6.25 was used for the conversion to total protein). Lipid content was verified by the Soxhlet method. The percentage of carbohydrates was calculated by difference. All determinations were carried out in triplicate.

For colour analysis, parameters L*, a* and b* were obtained with a MiniScan EZ apparatus (HunterLab) for 10 samples of each treatment. Chroma (C*) and hue angle (H°) were calculated as follows (Mcguire 1992):

$$ {C}^{\ast }=\sqrt[2]{a^{\ast 2}+{b}^{\ast 2}} $$
(1)
$$ \theta =\left[\raisebox{1ex}{${\tan}^{-1}\left(\frac{b\ast }{a\ast}\right)$}\!\left/ \!\raisebox{-1ex}{$6.2832$}\right.\right]\times 360 $$
(2)

With:

$$ {\mathrm{a}}^{\ast }>0\ \mathrm{and}\ \mathrm{b}\ge 0\ \mathrm{then}\ {\mathrm{H}}^{{}^{\circ}}=\uptheta $$
$$ {\mathrm{a}}^{\ast }<0\ \mathrm{and}\ \mathrm{b}\ge 0\ \mathrm{then}\ {\mathrm{H}}^{{}^{\circ}}=180+\uptheta $$
$$ {\mathrm{a}}^{\ast }<0\ \mathrm{and}\ \mathrm{b}<0\ \mathrm{then}\ {\mathrm{H}}^{{}^{\circ}}=180+\uptheta $$
$$ {\mathrm{a}}^{\ast }>0\ \mathrm{and}\ \mathrm{b}<0\ \mathrm{then}\ {\mathrm{H}}^{{}^{\circ}}=360+\uptheta $$

For the evaluation of cookie physical characteristics, diameter (D) and thickness (T) were measured with a Pantec digital external micrometre (0–25 mm) and a digital calliper (Ford, 0–150 mm), according to Rao et al. (2016) and Sozer et al. (2014). Weight loss (WL) (Rodríguez-García et al. 2014), spread ratio (S) and thickness ratio were calculated using the following equations, where WL is the weight loss during baking, Wdough is the weight (g) of cookie dough before baking, Wcookie is the weight (g) of the cookies after baking, Tdough represents the thickness of the dough before baking, and Dcookie and Tcookie are the diameter and thickness of cookies after baking (mm), respectively. Results were determined in sextuplicate for each formulation:

$$ WL\left(\%\right)=\left(\frac{W_{dough}-{W}_{cookie}}{W_{dough}}\right)\times 100 $$
(3)
$$ S=\left(\frac{D_{cookie}}{T_{cookie}}\right) $$
(4)
$$ Thickness\kern0.34em ratio=\left(\frac{T_{cookie}}{T_{dough}}\right) $$
(5)

Texture profile analysis (TPA) of dough samples (pre-baking) was performed with 10 replicates of each formulation in a texturometer (TA-XT, Express Enhanced, Stable Micro Systems) equipped with a 10-kg load cell and using a 25-mm cylindrical probe (P/25) at 25 °C. Dough samples (3 cm height × 4.5 cm diameter) were compressed by two cycles with a 5-s interval and speed test of 5 mm s−1. Dough samples were compressed at 60% of their height, and the results evaluated were hardness (N), adhesiveness (N s−1), springiness (−), resilience and cohesiveness. For the baked cookies, puncture force (N) was determined as described by Chen and Opara (2013) with minor modifications. For each treatment [control, M15 (15%wt substitution of margarine by microparticles) and M30 (30%wt substitution of margarine by microparticles)], ten baked cookies were analysed immediately after baking (0 day) and after storage for 30 days. Cookies were compressed to 50% depth of thickness with a cylindrical probe of 2 mm diameter (P/2) at a test velocity of 1 mm s−1.

For the evaluation of cookie microstructural characteristics, samples were initially stored in a desiccator for 1 week to remove adsorbed moisture. Afterwards, samples were fractured, gold-coated and then visualised in a scanning electron microscope (Carl Zeiss EVO-MA15 at 20 kV) using backscattered electron detector. The oil-loaded microparticles were also analysed by scanning electron microscopy (SEM).

Cookies (5–10 mg) were subjected to differential scanning calorimetry (DSC; Perkin Elmer, 4000) in aluminium pans. Samples were scanned from 0 to 300 °C at 20 °C min−1 under a nitrogen gas flow at 50 mL min−1.

Fatty acid methyl esters (FAME) were prepared by methylation of triacylglycerols, according to Hartman and Lago (1973), as described by Milinsk et al. (2008). Chromatographic analysis was performed in a gas chromatograph equipped with a flame ionisation detector (FID; Shimadzu GC-2010 Plus) and a capillary column (BPX-70, 60 m × 0.25 mm i.d. × 0.25 μm film thickness) with 70% cyanopropyl polysilphenylene-siloxane. A sample volume of 1.0 μL was injected, with a split ratio of 1:50. Column temperature program was as follows: 160 °C was kept for 2 min and then increased at 2 °C min−1 to 170 °C (2 min), increased at 4 °C min−1 to 180 °C (4 min), and finally heated at 10 °C min−1 to 235 °C (9 min). The carrier gas was hydrogen at 1.24 mL min−1 at 35.4 cm s−1. The make-up gas was nitrogen at 30 mL min−1. The flame of the FID was produced with hydrogen (30 mL min−1) and synthetic air (300 mL min−1). Fatty acid identification was performed by comparison with the retention time of authentic FAME standards (Sigma–Aldrich), and analyses were performed in triplicate.

The quality of the lipid fraction was evaluated through the atherogenicity index (AI) and thrombogenicity index (TI), according to Ulbricht and Southgate (1991) and, also, by the hypocholesterolaemic-to-hypercholesterolaemic fatty acid ratio (HH) (Santos-Silva et al. 2002) using Eqs. 6 to 8 where C12:0 is lauric acid, C14:0 is myristic acid, C16:0 is palmitic acid, C18:0 is stearic acid, C8:1n-9cis is oleic acid, C18:2n-6 is linoleic acid, C20:4n-6 is arachidonic acid, C18:3n-3 is α-linolenic acid, C20:5n-3 is eicosapentaenoic acid, C22:5n-3 is eicosapentaenoic acid, C22:6n-3 is docosahexaenoic acid, ΣMUFA is the sum of mono unsaturated fatty acids and Σn-6 and Σn-3 are the sum of omega-6 and omega-3, respectively:

$$ \mathrm{AI}=\frac{\mathrm{C}12:0+4\times \mathrm{C}14:0+\mathrm{C}16:0}{\Sigma \mathrm{MUFA}+\Sigma \mathrm{n}-6+\Sigma \mathrm{n}-3} $$
(6)
$$ \mathrm{TI}=\frac{\mathrm{C}14:0+\mathrm{C}16:0+\mathrm{C}18:0}{0,5\times \Sigma \mathrm{MUFA}+0,5\times \Sigma \mathrm{n}-6+3\times \Sigma \mathrm{n}-3}\kern1.25em $$
(7)
$$ \mathrm{HH}=\frac{\mathrm{C}8:1\mathrm{n}-9\ \mathrm{cis}+\mathrm{C}18:2\mathrm{n}-6+\mathrm{C}20:4\mathrm{n}-6+\mathrm{C}18:3\mathrm{n}-3+\mathrm{C}20:5\mathrm{n}-3+\mathrm{C}22:5\mathrm{n}-3+\mathrm{C}22:6\mathrm{n}-3}{\mathrm{C}14:0+\mathrm{C}16:0} $$
(8)

The moisture sorption isotherms of the cookies were determined at 25 °C (Aquasorp, Decagon Devices). Initially, samples were kept in a desiccator with silica for 1 week to equilibrate humidity. Experimental data were then fitted to the Guggenhein–Anderson–de Boer (GAB) (Eq. 9), Peleg (Eq. 10) and Kühn (Eq. 11) models, respectively, as follows, where Xw is the equilibrium moisture content (%dry matter) at a given water activity (aw), mo is the monolayer water content, C is the Guggenheim constant (represents the sorption heat of the first layer) and K is the sorption heat of the multilayer; B, k, z, k1, k2, n1 and n2 are the constants. All parameters were determined using non-linear regression with Statistica 7.0 software (StatSoft, USA). Fitting was evaluated by the coefficient of determination (R2) and root mean square error (RMSE, Eq. 12),

$$ {X}_w=\frac{\left({m}_0.C.K.{a}_w\right)}{\left(1-K.{a}_w\right).\left(1-K.{a}_w+C.K.{a}_w\right)} $$
(9)
$$ {X}_{\mathrm{w}}={k}_1{a_w}^{n1}+{k}_2{a_w}^{n2} $$
(10)
$$ {X}_w=K.{{}^{\left(\frac{1}{a_w}\right)}}^{-z}-B $$
(11)
$$ RMSE={\left[\frac{1}{N}.\sum \limits_{i=1}^N{\left({X}_{w_{\mathrm{exp}}}-{X}_{w_{calc}}\right)}^2\right]}^{\raisebox{1ex}{$1$}\!\left/ \!\raisebox{-1ex}{$2$}\right.} $$
(12)

Texture, flavour, appearance, odour and overall acceptance of all cookie formulations were evaluated by sensory acceptance tests, using a nine-point hedonic scale from 9 (extremely like) to 1 (extremely dislike) (Meilgaard et al. 1999). Untrained adult panellists (n = 53), aged 18 to 45 years, occupied individual and non-communicating cabins with fluorescent lighting. Before starting the test, panellists received the instructions so that they could choose to participate or not in the test. Samples were presented in monadic and randomised order and coded with three random digits. This work was approved by the Ethics Committee in Research of the Federal Technological University of Parana (UTFPR), Brazil, under protocol number 67425317.0.0000.5547.

Statistical Analysis

Proximate composition, fatty acid composition, texture, physical characteristics and sensory analyses data were evaluated by the Student’s t test or analysis of variance (ANOVA), at 5% significance level and by Tukey’s test. The colour parameters (L*, C* and H°) were analysed by a factorial ANOVA, to identify any interaction between the storage time and the treatments, also at 5% significance level, using Statistica 7.0 software (StatSoft, USA).

Results and Discussion

Cookie Dough Characterisation

The evaluation of dough’s texture is important since it is associated with the dough machinability in automated breadmaking processes (Armero and Collar 1997; Collar et al. 1998). From the TPA data (Table 2), the cookies with partial substitution of margarine by the microencapsulated chia oil (M15 and M30) had 2.5-fold less hardness than the control sample (p < 0.05). The plasticity and mechanical resistance of the dough increase and the dough became softer but more resistant mechanically as a consequence. O’Brien et al. (2003) observed an increase in dough hardness when hydrogenated vegetable fat was replaced by the microencapsulated fat, in biscuit formulations. In that study, a constant water quantity was added to the cookie formulation, in contrast to the present work, where the amount of water added was defined by the dough spread characteristics. Vrignaud (1977) related the observed softening effect to the water-binding ability of milk proteins, such as sodium caseinate (used in the microparticles as stabilising agent), and the plasticising properties of gluten, which affect the dough properties.

Table 2 Texture profile analysis of the cookie dough (before baking): control; M15 (15%wt substitution of margarine by microparticles) and M30 (30%wt substitution of margarine by microparticles)
Full size table

Samples containing the microparticles presented significantly higher springiness and resilience (p < 0.05) when compared to the control sample. This means that the partial substitution of margarine by the microparticles produced springy doughs; however, they were more resilient presenting the tendency to recover their original shape after being spread. Furthermore, the samples containing microparticles had higher cohesiveness compared to the control (p < 0.05). Collar et al. (1999), evaluating the texture properties of wheat doughs by using TPA noticed that the cohesiveness parameter was negatively correlated with hardness. Namely, when subjected to the same strain, a harder dough would suffer greater permanent damage to the internal structure than a less hard dough.

Microparticles and Cookie Characterisation

The solid lipid microparticles presented a spherical shape, as well as a regular size distribution (Fig. 1). Microstructural analysis of the cookies revealed the presence of starch granules with a variable size from 5 to 25 μm, in all samples, but to a greater abundance in the M30 cookies. Similar sized starch granules were observed by Ng et al. (2017) in the incorporation of dietary fibre-rich oyster mushroom powder into biscuits. According to Sanz et al. (2017), biscuit microstructure is formed by a protein-sugar matrix, in which starch granules are dispersed. During the baking process, fat melts and the flour particles are coated, making it difficult to hydrate the granules and form new bonds. Cookie formulations usually contain a large amount of sugar and insufficient water. Consequently, many of the starch granules do not gelatinise and remain visible in the microstructure, as seen in Fig. 1. This phenomenon was more evident in sample M30 (30%wt substitution of margarine by the microparticles) than the other cookie formulations, with the formation of a thick layer around the starch granules, probably due to the microparticles that melted during the baking process.

Fig. 1
figure 1

Scanning electron micrographs at × 1500 magnification: a microparticles; b cookie control formulation; c M15 cookies (15%wt substitution of margarine by microparticles); d M30 cookies (30%wt substitution of margarine by microparticles). Magnification: × 100: e microparticles; f control; g M15 cookies; h M30 cookies

Full size image

Glass transition temperature (Tg) of starch from wheat flour was found in all samples at 60 °C (Ng et al. 2017). Also, in all thermograms, the melting temperature of sucrose was approximately 186 °C (Lu et al. 2017). Microparticles presented an endothermic peak at 83 °C related to the melting temperature of the lipid phase (Rosiaux et al. 2015) (Fig. 2). This was also found in the cookie samples with 15 and 30%wt margarine substitution by the oil-loaded microparticles (M15 and M30, respectively) at 80 °C, and the intensity of such peaks was proportional to the quantity of microparticles added to the cookies as expected. It can also be noted that degradation occurred at approximately 300 °C, meaning no degradation is expected during the cookie baking that is often carried out at 180 °C.

Fig. 2
figure 2

Differential scanning calorimetry thermograms of: a microparticles; b cookie control; c M15 cookies (15%wt substitution of margarine by microparticles); d M30 cookies (30%wt substitution of margarine by microparticles)

Full size image

There was a significant difference (p < 0.05) in the percentages of moisture, lipids and carbohydrates between the samples (Table 3). Among all the cookies, the M30 samples had higher contents of moisture (5.22%) and carbohydrates (82.68%) and lower lipids (8.85%), in agreement with the formulation used (Table 1). No significant difference was found in the protein content, which was expected, given that white wheat flour was the source of protein and its concentration was kept constant in all cookie formulations.

Table 3 Cookie proximate composition; cookie physical properties (weight loss (WL%), thickness ratio and spread ratio) and puncture force (N) of baked cookies (0 and 30 days of storage): control, M15 (15%wt substitution of margarine by microparticles) and M30 cookies (30%wt substitution of margarine by microparticles)
Full size table

Regarding the cookie physical characteristics (Table 3), the weight loss (WL) as a function of water evaporation from dough during the baking step was equivalent in all samples (p > 0.05), despite the greater amount of water added to M15 and M30 formulations compared to control cookies. This result is in agreement with the proximate composition, as M30 cookie samples had the highest moisture content. A significantly higher (p < 0.05) thickness ratio was detected for the cookies with 30%wt microparticles than the control cookies. Furthermore, the M30 cookies presented a smaller S ratio than the control and M15 samples (p < 0.05), probably due to a higher increase in thickness during the baking process. This behaviour can be related to the presence of the sodium caseinate, used as a stabiliser for the microparticle production. Jayasena and Nasar-Abbas (2011) also noted an increase in the thickness of cookies after baking when high protein sources were added into the formulation.

The puncture forces decreased (p < 0.05) as storage time progressed (Table 3) but comparable values were recorded among the treatments for the same day (p > 0.05). Sudha et al. (2014) noticed that cookies produced with fat substitution by sodium caseinate microparticles or skimmed milk powder microparticles with encapsulated vegetable fat had higher hardness values (15–21 N) than the control sample (12 N).

Fatty Acid Profile and Lipid Fraction Quality

Palmitic (16:0) and stearic (18:0) were the main saturated fatty acids (SFAs) found in cookies, oleic (18:1n-9) was the major monounsaturated fatty acid (MUFA) and PUFAs or linoleic (18:2n-6) also predominated (Table 4). Shimizu et al. (2011) compared the fatty acid profiles of samples from four Brazilian margarine brands, being the profile similar when compared to the control sample of the present work. In the current study, SFA, MUFA, PUFA and total fatty acid contents in the control sample were comparable to those presented by Garsetti et al. (2016), for margarine samples marketed in the USA.

Table 4 Fatty acid composition (%) in the lipid fraction of cookies and indices of the nutritional quality of the lipid fraction: control, M15 cookies (15%wt substitution of margarine by microparticles) and M30 cookies (30%wt substitution of margarine by microparticles)
Full size table

The incorporation of the chia oil-loaded microparticles in the cookie dough caused a significant decrease (p < 0.05) in the SFA and MUFA fractions and increased the PUFAs. PUFA/SFA ratio was higher for the M30 sample, considering that the recommended value for this parameter should be greater than 0.45 (Wood et al. 2004). According to Simopoulos (2011), a 1:1–2:1 n-6/n-3 ratio is ideal for a healthy diet. Here, this ratio was 54.14% lower for the M30 sample than the control. The α-linolenic acid concentration increased 48.5% and 110.4% in formulations M15 and M30, respectively, relative to the control. This increase demonstrates the efficiency of the incorporation of the microparticles containing chia oil into the cookie dough. AI and TI decreased (p < 0.05) with the addition of the microparticles (Table 4), implying an increased concentration of anti-atherogenic and anti-thrombogenic fatty acids. The M30 formulation had the lowest values, of 0.290 (AI) and 0.400 (TI). Lower values indicate relatively healthier lipids, as they represent the potential for stimulating platelet formation (Fuchs et al. 2013; Ulbricht and Southgate 1991). For the HH ratio, there was a significant increase for the enriched formulations, particularly for the M30 sample (5.45%). From a comparative perspective, larger values for this ratio indicate nutritionally more adequate lipids for feeding (Santos-Silva et al. 2002).

Cookie Colour

It is possible to observe brown spots on the surface of the cookies after baking (Fig. 3a), notably on the control in comparison to the M15 and M30 cookies. Nishibori and Kawakishi (1992) attribute the formation of the spots to the Maillard reaction, involving the interaction of reducing sugars and proteins. Also, according to Chevallier et al. (2002), during baking, starch dextrinisation and sugar caramelisation also lead to a darkening effect. It is possible that the presence of microparticles had hindered one or more of these darkening processes.

Fig. 3
figure 3

a Scanning electron micrographs of cookie doughs (before baking) and after baking: c control; (M15) 15%wt substitution of margarine by microparticles; (M30) 30%wt substitution of margarine by microparticles; b colour luminosity (L*); c chroma (C*); and d hue angle (H°). a,b Bars marked with different letters present significant difference (p < 0.05) by Tukey’s test

Full size image

Immediately after baking (0 days), control cookies presented a similar luminosity (L*) to the M15 samples (Fig. 3b). Likewise, M15 had comparable L* values to M30. There was a remarkable increase in L* value as the partial substitution of margarine increased. At 30 days of storage, all samples presented a reduction in L* value; however, a statistical difference (p < 0.05) was only apparent for M30 sample relative to M15 and the control. Despite the decrease in L* value at 30 days of storage, the M30 exhibited a final L* not dissimilar to its initial value (0 days), resulting in the lowest variation in L* of all the formulations.

The calculated C* data (Fig. 3c) demonstrated that the sample with the highest percentage of margarine substitution by microparticles (M30) also showed the lowest variation over storage time; however, there was a decreasing pattern as the partial margarine substitution percentage increased. Conversely, for H° (Fig. 3d), M15 presented the lowest variation with storage duration. Furthermore, at both time intervals, H° of the control was similar to M15 at day 0, whereas the M30 values were different (p < 0.05) from M15 but comparable to the control formulation.

Moisture Sorption Isotherms

Type III isotherms were found for all cookie samples (Fig. 4), according to the Brunauer–Emmett–Teller classification (Brunauer et al. 1938). Also, when comparing the samples with margarine substitution (M15 and M30), the aw value increased with increasing microparticle concentration.

Fig. 4
figure 4

Moisture sorption isotherms at 25 °C of control, M15 cookies (15%wt substitution of margarine by microparticles) and M30 cookies (30%wt substitution of margarine by microparticles)

Full size image

Based on the fitting criterion (R2 and RMSE), the analysed mathematical models presented good fitting for the experimental data of all samples, except the control sample and the Kühn model (Table 5). From further examination of the GAB parameters, M15 and M30 samples had a higher m0 relative to the control, which could be related to the presence of sodium caseinate. The parameter C is a constant defined as the ratio of the partition function of the first adsorbed molecule on a site and the partition function of adsorbed molecules beyond the first molecule in the multilayer. Thus, the larger the C value, the stronger the water is bound in the monolayer and the larger the difference in enthalpy between the monolayer molecules and multilayer molecules (Quirijns et al. 2005). The results for the control, M15 and M30 cookies, were − 5.298 × 104, 0.975 and 4.427 × 104, respectively. As the amount of sodium caseinate is increased in the cookie formulation, due to the increased amount of microparticles, the C value is also increased, again suggesting the influence of caseinate on this parameter.

Table 5 Data for model parameters of the moisture sorption isotherms of cookies: control; M15 cookies (15%wt substitution of margarine by microparticles) and M30 cookies (30%wt substitution of margarine by microparticles)
Full size table

The parameter k provides a measure of the interactions between the multilayers with the food matrix, and its molecular energy value tends to decrease for the interaction between the monolayer and liquid water. Parameter k is called a correction factor, as it corrects the properties of the multilayer molecules relative to the bulk liquid. All cookie samples presented a k > 0.9. When k approaches one, there is almost no distinction between multilayer molecules and liquid molecules. In this context, in all the cookies produced in the current study, the water molecules presented the same characteristics as the molecules in the bulk liquid (Quirijns et al. 2005). Still, according to Bastıoğlu et al. (2017), a k > 0.9 indicates a behavior of liquid water of molecules.

Cookie Sensory Acceptability

For all the sensory parameters assessed, M15 and control cookies were similar to each other (p > 0.05) (Fig. 5). In contrast, the cookies with the highest substitution of margarine by the oil-loaded microparticles (M30) presented the lowest results for all scores.

Fig. 5
figure 5

Sensory scores for control, M15 cookies (15%wt substitution of margarine by microparticles) and M30 cookies (30%wt substitution of margarine by microparticles)

Full size image

Panellists were able to differentiate sample M30 from the other formulations when texture was evaluated (p < 0.05). It is worth noting that sensory perception involves several steps outside and inside the mouth of the panellist, from the first bite through to mastication, swallowing and residual feel in the mouth and throat (Szczesniak 2002). Thus, a different evaluation between the instrumental and sensory analysis can be expected.

The appearance of the control and M15 samples was also preferred by the panellists over the M30 cookies. Overall, it may be concluded that there was good acceptability of the cookies produced with the substitution of margarine by 15%wt of solid lipid microparticles containing chia oil.

Conclusions

Cookies with partial substitution of margarine (15 and 30%wt) by microencapsulated chia oil were produced successfully. α-Linolenic acid concentration increased 48.49 and 110.41%, respectively, when compared with the control sample, and also the atherogenicity and thrombogenicity indexes were significantly reduced. Microparticles melted during baking creating a thick surface around the starch granules. Formulations containing the oil-loaded microparticles had significant differences in hardness, cohesiveness, springiness and resilience when compared to the control, probably due to the presence of the microparticles, sodium caseinate (stabilising agent in the microparticles) and the higher amount of water required to obtain the doughs. On the other hand, no significant difference was detected in the puncture test. Cookies without margarine substitution (control) were the brownest but the samples with 30%wt substitution presented the highest colour stability during the storage duration (30 days). Water sorption isotherms presented good fitting results with GAB, Peleg and Kühn equations, respectively. Evaluation of the GAB parameters showed an increase in the m0 with the increase in the microparticle amount in the formulation and, consequently, in the sodium caseinate content. An increase in the C parameter was also found, indicating that water is more strongly bound as a function of the increase in margarine partial substitution. The cookies produced with the substitution of margarine by 15%wt oil-loaded microparticles were well-accepted sensorially.