Microalgae as Functional Ingredients in Savory Food Products: Application to Wheat Crackers

Crackers are widely consumed snack foods and there is an increasing trend in adding functional ingredients to their composition. In the present work, the dried biomasses of four microalgae strains—Arthrospira platensis F&M-C256, Chlorella vulgaris Allma, Tetraselmis suecica F&M-M33, and Phaeodactylum tricornutum F&M-M40—were used as a source of proteins, antioxidants, and other bioactive molecules in artisanal wheat crackers. Two incorporation levels were tested: 2% (w/w) and 6% (w/w). The impact of microalgae addition was evaluated in terms of physical properties, biochemical composition, antioxidant activity, in vitro digestibility, and sensory characteristics. Microalgae crackers presented stable color and texture throughout eight weeks of storage. Microalgae crackers were slightly thinner and lighter than the control but presented a similar density in agreement with scanning electron microscope images, indicating that gas retention was not greatly affected by microalgae addition. Regarding biochemical composition, 6% A. platensis and C. vulgaris crackers presented a significantly higher protein content (13.2–13.5%), for which they could be claimed to be a “source of protein” according to the Regulation (EC) No. 1924/2006. A. platensis crackers showed the highest antioxidant activity and attained better sensory analysis scores. T. suecica and P. tricornutum crackers showed high phenolic content and antioxidant activity but attained low sensory scores mainly because of their unattractive fishy off-flavor.


Introduction
The world snack food market has been projected to reach USD 639 billion by 2023, growing at a compound annual growth rate (CAGR) of 5.8% from 2018 to 2023 [1]. In recent years, a major market trend has been the introduction of new products containing health-promoting ingredients associated with a growing demand for healthy snacks due to the consumer mind set changing taste and preference. In particular, savory snacks, namely, salted crackers, are growing at a greater rate than sweet snacks [2]. Crackers are considered popular snack products which have appreciable demand amongst consumers [3,4]. Crackers are generally defined as dry, thin, and crisp bakery products [5] usually made of wheat flour, fat (or shortening), salt, and leavening agents (yeast, chemical leaveners, or a combination of both). In general, there are three major types of crackers: saltine (soda crackers), Table 1. Biochemical composition of the four microalgae biomasses used in the experiments (%, dry weight). Results are expressed as average ± standard deviation (n = 3).

Cracker Preparation
Cracker samples were prepared according to the following control formulation (w/w): 60.5% commercial all-purpose wheat flour T55 (76.7% carbohydrates, 7.8% protein, 2.9% fiber, and 1% lipids), 1.5% baking powder (corn starch, sodium diphosphate, and sodium bicarbonate), 1% salt (NaCl), 1% sugar (sucrose), 7.5% vegetable oil, and 28.5% distilled water. Microalgae biomass was added to the same formulation at 2% and at 6% (w/w) incorporation levels by replacing a corresponding amount of wheat flour. The ingredients were weighed based on a 150 g batch and kneaded for 1 minute at speed 4 in a food processor (Bimby, Vorwerk, Germany) to obtain a homogeneous cohesive dough. For replications, three separate batches of each cracker formulation were made. The dough was further laminated into thin sheets which were passed through a pasta roller machine under three different gauge positions (the dough was passed through three times for each position). A square mold (56 × 56 mm) was used to cut the laminated dough into pieces, which were then slightly perforated. The shaped cracker dough was left to proof for 10 min at room temperature. The crackers were then baked in a forced-air convection oven (Unox, Italy) at 180 • C for 10 min. Baked samples were dried at 60 • C for 30 min, cooled at ambient temperature for 30 min, and stored in non-transparent closed plastic containers. Physical analyses were performed after 24 h ("week 0") and after four-and eight-week storage (color, texture, and a w ). Part of the cracker batches were immediately crushed to powder (using an electric mill) and frozen to be used for biochemical composition, antioxidant capacity, and in vitro digestibility analyses.

Dimensions
The characteristic dimensions of the crackers were measured using a digital caliper model 684132 (Lee Tools, Houston, TX, USA). The individual width (W) and thickness (T) of 20 crackers from each formulation type were measured; spread ratio (W/T) was calculated accordingly. The weight of the cracker samples was also measured and the corresponding densities calculated (weight (g)/volume (cm 3 )). All these analyses were carried out 24 h after cracker preparation.

Color Analysis
The color of the cracker samples was measured instrumentally using a Minolta CR-400 (Japan) colorimeter with standard illuminant D65 and a visual angle of 2 • . The results were expressed in terms of L*, lightness (values increasing from 0% to 100%); a*, redness to greenness (60 to −60 positive to negative values, respectively); and b*, yellowness to blueness (60 to −60 positive to negative values, respectively) according to the CIELab system. Chroma, C* ab (saturation) and hue angle, hº ab , were also calculated as defined by C* ab = [(a* 2 + b* 2 )] 1/2 ; hº ab = arctan(b*/a*). The total color difference between sample crackers throughout the storage time (up to eight weeks) was determined using average L*a*b* values according to ∆E* = [(∆L*) 2 + (∆a*) 2 + (∆b*) 2 ] 1/2 . The measurements were conducted under the same light conditions using a white standard (L* = 94.61, a* = −0.53, and b* = 3.62) under artificial fluorescent light at room temperature, and measurements were replicated ten times for each formulation sample (one measurement per cracker) as well as for the control 24 h, 4 weeks, and 8 weeks after preparation.

Texture analysis
Instrumental texture analysis was carried out in a TA.XTplus (Stable Micro Systems, Godalming, UK) texturometer. Crackers were submitted to a "three-point bending" or "snap" test, at 3 mm/s probe speed, with a 5 kg load cell, and at a controlled (20 ± 2 • C) room temperature. Hardness was calculated as the force peak (N) in the force versus time texturogram; N corresponds to the maximum force that is needed to break or "snap" the cracker. Measurements were repeated ten times for each formulation sample (one measurement per cracker) as well as for the control 24 h, 4 weeks, and 8 weeks after preparation.

Microscopy Analysis
Crackers with 2% (w/w) microalgae biomass (and the control) were analyzed using a table top scanning electron microscope (TM3000, Hitachi, Tokyo, Japan) under low vacuum  conditions and Secondary Electrons/Back-Scattered Electrons (SE/BSE) dual signal mode (combined topography/morphology) with 40×, 100×, and 250× magnification. Cracker samples were cut into small pieces (approximately 2 × 4 × 4 mm) and placed in a sample holding chamber in a cross-sectional position.

Total Water Content and Water Activity (a w ) Determination
Crackers samples were analyzed for total water content and water activity. Water content was determined gravimetrically using an automatic moisture analyzer (PMB 202, aeADAM, UK) at 130 • C until a constant weight was achieved. Water activity (a w ) was determined using a thermohygrometer (HygroPalm HP23-AW, Rotronic AG, Bassersdorf, Switzerland) at 20 ± 1 • C. Measurements were performed during storage time (24 h, 4 weeks, and 8 weeks after preparation) by crushing the samples to powder and using an electric mill immediately before analysis (in triplicate). The crackers' proximate biochemical composition was analyzed on powdered samples in terms of total ash content, crude protein [48], crude fat [49], and total dietary fiber [50] by standard methods, as detailed in Batista et al. [17]. Chemical analyses were performed in triplicate, except for total dietary fiber, where a single test was carried out according to the AOAC 991.43 [50] Standard Method. Carbohydrates were calculated by difference and energy value was calculated using the conversion factors designated in Annex XIV of Regulation (EU) No. 1169/2011 [51]. Sodium content was also analyzed, in triplicate, by atomic absorption spectrometry in a certified external laboratory [52].

Phenolics and Antioxidant Capacity Determination
A total phenolic content assay was carried out according to Ganesan, Kumar, and Bhaskar, [53] using the Folin Ciocalteu assay. Samples of 0.1 g of lyophilized crackers were dissolved in 6 mL deionized water and sonicated (MicrosonTM XL2000, Misonix Inc., Farmingdale, NY, USA) for 30 min at maximum power (frequency 20 kHz and power 130 W) while maintaining the temperature below 30 • C by immersing the sample flask in an ice bath. To 100 µL aliquots of each sample was added 2 mL of 2% sodium carbonate (Sigma-Aldrich, St. Louis, MO, USA) in water. After 2 min, 100 µL of 50% Folin Ciocalteu reagent (Sigma-Aldrich, St. Louis, MO, USA) was added. The reaction mixture was incubated in darkness at 25 • C for 30 min. The absorbance of each sample was measured at 720 nm. Results were expressed in gallic acid equivalents (mg GAE g −1 ) through a calibration curve (gallic acid: 0 to 150 µg mL −1 , R 2 = 0.9907) (Sigma-Aldrich, St. Louis, MO, USA).
To evaluate the radical scavenging capacity of the cracker samples, a DPPH radical scavenging assay was carried out according to Rajauria et al. [54]. Sample extracts were prepared by adding 0.2 g of lyophilized crackers to 20 mL of a 1:5 methanol:water solution and sonicating for 30 min (the same conditions as described above for total phenols extracts). The DPPH assay was performed by mixing 100 µL of DPPH radical solution (165 µM in methanol, Sigma-Aldrich, St. Louis, MO, USA) and 100 µL of sample extract. The reaction mixtures were incubated in darkness at 30 • C for 30 min and the absorbance was measured at 517 nm. The antioxidant capacity of the samples was expressed in terms of µg of vitamin C equivalent antioxidant capacity (VCEAC) per milligram of sample (ascorbic acid calibration curve: 0 to 10 µg mL −1 , R 2 = 0.9918) and corresponding radical scavenging activity (RSA). Two blank assays, one without samples and another without reagents, were also performed.
Analyses were repeated in triplicate and performed on powdered cracker samples.

In Vitro Digestibility Tests
The in vitro dry matter digestibility (IVDMD) of the microalgae samples was assessed by the method described in Niccolai et al. [55]. This method reproduces the chemical-enzymatic digestion (by gastric and pancreatic juices) which occurs in the proximal tract of the monogastric digestive system. After a two-step enzymatic digestion with porcine pepsin and pancreatin (AppliChem GmbH, Darmstadt, Germany), the IVDMD (%) was calculated from the difference between the initial sample biomass (1 g) and the undigested biomass residue (after correction for the blank assay). A control assay with casein standard (Sigma-Aldrich, St. Louis, MO, USA), used as reference material with 100% digestibility, was also carried out in order to assess the validity of the method (IVDMD = 97.2 ± 0.8%).
The protein content of the undigested biomass residue was analyzed using Lowry's [56] method to calculate the in vitro protein digestibility (IVPD, % = total protein content − protein content after digestion) ÷ total protein × 100).
The same procedure was applied to analyze the IVDM and IVPD of the cracker samples produced in this work and of six commercial cracker samples (Table 2) in order to assess the range of variation of these parameters in commercial products.
All analyses were repeated in triplicate and performed on powdered samples.

Sensory Analysis
Cracker samples with 2% microalgal biomass, as well as the control sample, were tested by an untrained sensory analysis panel (n = 30, age: 19-38, gender: 8 M, 22 F). The cracker samples were evaluated in terms of color, smell, taste, texture, and global appreciation (six levels from "very pleasant" to "very unpleasant"), as well as salt perception (five levels from "very unsalted" to "very salty"). The buying intention was also assessed, from "would certainly buy" to "certainly wouldn't buy" (five levels). The assays were conducted in a standardized sensory analysis room according to the standard EN ISO 8589 [57].

Statistical Analysis
Statistical analysis of the experimental data was performed using STATISTICA from StatSoft (version 8.0, Dell, Round Rock, TX, USA), through variance analysis (one-way ANOVA), and by the Scheffé test-post hoc comparison at a significance level of 95% (p < 0.05). All results have been presented as average ± standard deviation.

Cracker Colour
Microalgae crackers presented unusual and visually attractive colors, as can be observed from the photos shown in Figure 1. The evolution of color parameters throughout storage time (0, 4, and 8 weeks) in terms of lightness (L*), greenness (a*), yellowness (b*), chroma (C*), and hue (hº) is represented in Figure 2.

Sensory Analysis
Cracker samples with 2% microalgal biomass, as well as the control sample, were tested by an untrained sensory analysis panel (n = 30, age: 19-38, gender: 8 M, 22 F). The cracker samples were evaluated in terms of color, smell, taste, texture, and global appreciation (six levels from "very pleasant" to "very unpleasant"), as well as salt perception (five levels from "very unsalted" to "very salty"). The buying intention was also assessed, from "would certainly buy" to "certainly wouldn't buy" (five levels). The assays were conducted in a standardized sensory analysis room according to the standard EN ISO 8589 [57].

Statistical Analysis
Statistical analysis of the experimental data was performed using STATISTICA from StatSoft (version 8.0, Dell, Round Rock, TX, USA), through variance analysis (one-way ANOVA), and by the Scheffé test-post hoc comparison at a significance level of 95% (p < 0.05). All results have been presented as average ± standard deviation.

Cracker Colour
Microalgae crackers presented unusual and visually attractive colors, as can be observed from the photos shown in Figure 1. The evolution of color parameters throughout storage time (0, 4, and 8 weeks) in terms of lightness (L*), greenness (a*), yellowness (b*), chroma (C*), and hue (hº) is represented in Figure 2.   In terms of lightness, it is evident that the addition of microalgae biomass resulted in much darker crackers in relation to the control (L* 77.7), with 2% microalga samples showing values around 39-49 and 6% microalga samples values around 30-37. Regarding a*, microalgae samples showed negative values in the green domain. However, when increasing microalgae concentration to 6% the green chromaticity decreased, reaching a* values which were close to zero or even positive in some cases ( Figure 2). Similar behavior was observed in terms of yellow chromaticity, with b* and C* values decreasing when algae concentration increased from 2% to 6%. Similar effects have been reported in the literature for microalgae cookies [15][16][17] and other baked products. These results suggest that the reaction kinetics of pigment degradation, namely green chlorophylls, upon high temperature baking might be dependent on the initial pigment concentration. Additionally, the formation of brown-colored chlorophyll degradation products such as pheophorbides and pyropheophorbide can also impact the visual color perception and consequently the positioning of the sample within the L*a*b* tridimensional space. Besides this, moisture loss and volume changes occurring upon baking can also greatly influence the crackers' color appearance [58]. As will be discussed later, 6% microalgae samples presented lower dimensions (Table 3) and lower water contents (Table 4), which can influence color perception and accelerate surface browning [58]. Similarly, the formation of brown-colored Maillard reaction compounds can also be favored for 6% microalgae samples which have higher protein contents (Table 5).
Nevertheless, hue angle remained relatively constant for each microalgae sample, independently of the concentration, and was very stable throughout the eight-week storage time. This means that the crackers' color was deeply related to the microalgae's pigment profile while its intensity or chromaticity might have been affected by the baking procedure.
P. tricornutum samples showed the highest b* values (up to 26) and hue angle close to pure yellow (90-96 • ), which should be related to the presence of fucoxanthin [59]. On the other hand, C. vulgaris and T. suecica showed the highest a* values (in modulus) and hue angles closer to pure green (116-118 • ), which is in agreement with the high chlorophyll content of these chlorophytes [60].
All cracker samples showed good color stability throughout the eight-week storage period, with variations being barely detected by the human eye and with ∆E* values being lower than 5except for the A. platensis crackers, which showed an ∆E* of 9.3 (for 2% alga).

Cracker Dimensions
Characteristic dimensions of the microalgae crackers are presented in Table 3. In general, crackers with 2% algal biomass did not show significant differences (p > 0.05) in terms of dimensions (width, thickness, and spread ratio) in relation to the control. However, for the higher microalgae incorporation level (6%) it was observed that there was a significant (p < 0.05) reduction in the crackers' width and particularly thickness (3.5-3.8 versus 4.5 mm) when compared to the control. These differences led to higher spread ratios for the microalgae crackers which were statistically significant (p < 0.05) for 6% T. suecica and 6% P. tricornutum. The average weight of the crackers was also significantly lower for the 6% algae samples (3.8-4.7 g) in relation to the control (5.8 g), which might also be related to lower moisture retention upon baking. Nevertheless, the crackers' density remained statistically (p > 0.05) constant upon microalgae addition (Table 3), even if smaller values were reported for the 6% samples. This is a very important quality parameter, since low-weight and less dense crackers are usually desired by consumers [61]. This seems to indicate that gas retention in puffed crisp crackers is not hindered by the presence of microalgae. Table 3. Characteristic dimensions of crackers with 2% and 6% (w/w) microalgae biomass incorporation. Results are expressed as average ± standard deviation (n = 20). Different letters in the same column correspond to significant differences (p < 0.05).

Cracker Texture and Microstructure
Texture is one of the most important quality attributes in crackers since consumers highly appreciate a crisp and crunchy texture [62]. One of the most suitable instrumental analysis tests for assessing the texture of these types of brittle food samples is the "three-point bending" or "snap" test, in which the cracker is placed upon two support beams while a third moves down (parallel) into the sample's middle point causing the sample to fracture into two pieces [63]. Figure 3 presents the hardness results of the cracker samples as determined by the snap test. In general, crackers with 2% algal biomass did not show significant differences (p > 0.05) in terms of dimensions (width, thickness, and spread ratio) in relation to the control. However, for the higher microalgae incorporation level (6%) it was observed that there was a significant (p < 0.05) reduction in the crackers' width and particularly thickness (3.5-3.8 versus 4.5 mm) when compared to the control. These differences led to higher spread ratios for the microalgae crackers which were statistically significant (p < 0.05) for 6% T. suecica and 6% P. tricornutum. The average weight of the crackers was also significantly lower for the 6% algae samples (3.8-4.7 g) in relation to the control (5.8 g), which might also be related to lower moisture retention upon baking. Nevertheless, the crackers' density remained statistically (p > 0.05) constant upon microalgae addition (Table 3), even if smaller values were reported for the 6% samples. This is a very important quality parameter, since low-weight and less dense crackers are usually desired by consumers [61]. This seems to indicate that gas retention in puffed crisp crackers is not hindered by the presence of microalgae.

Cracker Texture and Microstructure
Texture is one of the most important quality attributes in crackers since consumers highly appreciate a crisp and crunchy texture [62]. One of the most suitable instrumental analysis tests for assessing the texture of these types of brittle food samples is the "three-point bending" or "snap" test, in which the cracker is placed upon two support beams while a third moves down (parallel) into the sample's middle point causing the sample to fracture into two pieces [63]. Figure 3 presents the hardness results of the cracker samples as determined by the snap test. Due to the extremely brittle nature of these samples, the response to breakage was highly variable, and, consequently, the hardness results had relatively high experimental errors (5-13%).
For all the samples, a significant variation in cracker hardness throughout the eight-week study was not observed (p > 0.05), which means that the texture properties remained stable, a relevant quality in terms of sensory acceptance by the consumer.
The crackers with A. platensis, C. vulgaris, and T. suecica did not show significant (p > 0.05) hardness differences in relation to the control crackers, although there seemed to be a tendency for hardness to increase with microalgae addition, especially for the 6% samples. These results are also in agreement with previous studies with microalgae cookies in which structural reinforcement has been observed upon microalgae addition (e.g., [15][16][17]64]).
In the case of the P. tricornutum crackers the inverse behavior was observed, i.e., there was a tendency for hardness to decrease, which was statistically significant (p < 0.05) for the 6% sample. In fact, P. tricornutum 6% dough samples were more difficult to manipulate since the dough was very elastic. As a result, these crackers became much thinner and lighter (showing a higher spread ratio) ( Table 3) as well as less resistant to breakage.
The samples with 2% microalga incorporation, as well as the control, were also analyzed by SEM. From the microscope images ( Figure 4) it seems that the control sample presents a denser and more continuous structure, while crackers with microalgae seem to have a slightly more porous and heterogeneous microstructure.  It seems reasonable to admit that the addition of microalgae might have influenced gluten matrix formation and/or starch gelatinization mechanisms, although to a limited extent. Batista et al. [65] have reported that Spirulina biomass impairs starch gelatinization in model gel systems, namely by increasing the gelatinization temperature, which is probably due to competition for water binding zones during the hydration of starch granules. In fact, during dough mixing, the available free water is partitioned between microalgae and wheat flour hydrophilic sites, which might also impair gluten protein network development. It is possible that the formation of a weaker It seems reasonable to admit that the addition of microalgae might have influenced gluten matrix formation and/or starch gelatinization mechanisms, although to a limited extent. Batista et al. [65] have reported that Spirulina biomass impairs starch gelatinization in model gel systems, namely by increasing the gelatinization temperature, which is probably due to competition for water binding zones during the hydration of starch granules. In fact, during dough mixing, the available free water is partitioned between microalgae and wheat flour hydrophilic sites, which might also impair gluten protein network development. It is possible that the formation of a weaker gluten network might result in the collapse of small gas cells into larger cavities, which might influence gas and water trapping during baking.

Cracker Water Content and Water Activity
Water content and/or water activity (a w ) are critical quality parameters in low moisture foods which strongly influence their crispiness and sensory acceptance. Above a critical a w value, typically around 0.5, the food materials get softer and stale, losing their crispiness [66,67]. Table 4 presents the microalgae crackers' water content and a w value evolution throughout the storage period (at 24 h, 4 weeks, and 8 weeks), and, as expected, these results seem well correlated. Table 4. Evolution of total water content and water activity (a w ) of crackers with 2% and 6% (w/w) microalgae biomass incorporation across time. Results are expressed as average ± standard deviation (n = 3). Different letters (a to h) in the same column correspond to significant differences (p < 0.05) between samples. Samples marked with x, y, or z in the same line correspond to significant (p < 0.05) differences from week 0 to week 8.

Water Content (%, w/w) Water Activity (a w )
Week The control cracker presented an initial a w of 0.128 and adding 2% microalgae biomass caused an increase (p < 0.05) of up to 0.15-0.16 (except for C. vulgaris), which still corresponds to highly crispy products [67]. On the other hand, higher microalgae biomass incorporations (6%) led to significantly (p < 0.05) lower a w values of 0.07-0.08 (except for T. suecica). The same behavior was observed in terms of water content, although the differences detected were not statistically significant (p > 0.05); samples with 2% microalga had initial moisture content (3.3-4.3%) similar to the control (3.7%), while 6% microalgae crackers had lower initial moisture content (2.1-2.9%). A similar behavior in terms of a w was observed by Batista et al. [17] in short-dough biscuits with the addition of 2% and 6% biomass of the same microalgae. As discussed in the previous sections, it is possible that the addition of high concentrations of microalgae biomass could lead to a weaker gluten network which is unable to efficiently trap gas bubbles and water molecules. Hence, a higher moisture loss for the 6% microalgae crackers could have led to a lower moisture content (Table 4) and lower weight (Table 3) for these samples. There was a tendency for total moisture and a w to increase (p < 0.05) with time, which was the same pattern undergone by the control (Table 4). This means that, in general, the addition of microalgae biomass does not seem to have a negative impact on the crackers' shelf life with respect to a w evolution (and consequent crispness loss). In fact, after 8 weeks' storage, the majority of the samples showed lower a w than the control, with the lowest values observed for P. tricornutum (0.10-0.12). For T. suecica 2%, a pronounced increase was observed with time with moisture values increasing from 3.3% up to 7.7% and a w values from 0.15 to 0.52. Table 5 presents the proximate biochemical composition of the crackers prepared with microalgae biomass incorporation. Table 5. Proximate biochemical composition (g/100g) of crackers with 2% and 6% (w/w) microalgae biomass incorporation. Results are expressed as average ± standard deviation (n = 3) except for total dietary fiber (n = 1). Different letters in the same column correspond to significant differences (p < 0.05). Regarding mineral content, no significant differences (p > 0.05) were found in terms of total ash content (except for C. vulgaris 6%) and sodium contents. Nevertheless, it should be noted that there seems to be a tendency for mineral content to increase for higher microalgae biomass incorporations. In terms of sodium content, the highest value was found in T. suecica 6% crackers, which presented 1.1% Na, corresponding to 2.6% NaCl (control: 1.9% NaCl). This is in agreement with the higher sodium content in the biomass of this marine microalga (Table 1). These values are comparable to those of commercial wheat crackers (Table 2, 1.8-2.2 g salt/100 g).

Cracker Proximate Biochemical Composition
In terms of fat content, there were also no significant (p > 0.05) differences between the control and microalgae crackers, with crude fat contents ranging from 11.5% to 13.2% (Table 5), similarly to commercial wheat crackers (Table 2, 6.3-18.0%).
Microalgae-enriched cracker samples showed significantly higher (p < 0.05) protein contents than the control (9.8%) ( Table 5). Crackers with 2% microalgae ranged from 10.4 to 11.4% protein while 6% microalgae crackers ranged from 12.1 to 14.6% protein, with the highest values obtained for A. platensis and C. vulgaris samples. This is in agreement with the high protein content of these microalgae (Table 1) and to previous studies in the literature for microalgae-enriched bakery food products (e.g., [17,20,21,23,64]). In fact, these microalgae-enriched crackers can be regarded as a very interesting protein-fortified bakery product. For 6% A. platensis and 6% C. vulgaris crackers, the protein content was found to correspond to 13.2 and 13.5%, respectively, of the total energy content of the sample, so these products can be claimed as "source of protein" according to Regulation (EC) 1924/2006 [68].
Regarding the total dietary fiber content, microalgae cracker samples presented average values ranging from 4.4 to 6.7% in the same order of magnitude as the control crackers (5%) (commercial wheat crackers: 2.5-5.0%, Table 2). The highest value was reported for the A. platensis 6% sample. Previous studies have also reported higher fiber content in A. platensis-enriched biscuits [64,69]. The total carbohydrate value was always lower than the control, while the energy values remained relatively constant (430-442 kcal/100 g) and similar to those of commercial wheat crackers ( Table 2, 410-461 kcal/100 g). Figure 5 represents the total phenolic content of the cracker samples. The addition of 2% microalgae biomass did not promote significant (p > 0.05) differences in phenolic content in relation to the control (1.4 mg GAE g −1 ). However, when increasing microalgae biomass content to 6%, phenolic content significantly (p < 0.05) increased to 1.6-1.7 mg GAE g −1 for A. platensis and C. vulgaris and to 2.3-2.5 mg GAE g −1 for T. suecica and P. tricornutum, which corresponds to a net total phenolics increase of about 0.3 and 1 mg GAE g −1 compared to the control, respectively. Other authors have highlighted the high phenolic compound content of Phaeodactylum and Tetraselmis [70,71].

Cracker Phenolic Content and Antioxidant Capacity
The results obtained for T. suecica crackers are higher than those attained by Batista et al. [17] in sweet cookies prepared with the same microalgae strain. Similar behavior has been reported by Abraços [72] in Salvia sclaeroides-enriched sweet cookies and crackers (ten times higher phenolic content in the crackers) which were prepared exactly under the same thermal processing conditions as the microalgae crackers and biscuits studied in this work.
The antioxidant capacity of microalgae-enriched crackers was tested by the DPPH method ( Figure 6). Other authors have highlighted the high phenolic compound content of Phaeodactylum and Tetraselmis [70,71].
The results obtained for T. suecica crackers are higher than those attained by Batista et al. [17] in sweet cookies prepared with the same microalgae strain. Similar behavior has been reported by Abraços [72] in Salvia sclaeroides-enriched sweet cookies and crackers (ten times higher phenolic content in the crackers) which were prepared exactly under the same thermal processing conditions as the microalgae crackers and biscuits studied in this work.
The antioxidant capacity of microalgae-enriched crackers was tested by the DPPH method ( Figure 6).
Similarly to the total phenolic results ( Figure 5) for the 2% microalgae crackers, there were no significant (p > 0.05) improvements in the antioxidant capacity (0.55-0.60 µg VCEAC mg −1 ) in relation to the control sample (0.54 µg VCEAC mg −1 , 53% RSA). The addition of 6% microalgae led to a significant (p < 0.05) antioxidant capacity increase (up to 0.62-0.65 µg VCEAC mg −1 ) for the T. suecica and P. tricornutum samples. Moreover, 6% A. platensis crackers showed the highest antioxidant capacity value of 0.70 µg VCEAC mg −1 , which corresponds to 68% RSA. Other authors [17,64,73] have also reported high antioxidant activity for A. platensis-enriched biscuits, which might be related to the presence of phycocyanin. sweet cookies prepared with the same microalgae strain. Similar behavior has been reported by Abraços [72] in Salvia sclaeroides-enriched sweet cookies and crackers (ten times higher phenolic content in the crackers) which were prepared exactly under the same thermal processing conditions as the microalgae crackers and biscuits studied in this work.
The antioxidant capacity of microalgae-enriched crackers was tested by the DPPH method ( Figure 6).   Figure 7 presents the in vitro digestibility of microalgae crackers, in terms of total dry matter (IVDMD) (Figure 7a) and protein (IVPD) (Figure 7b) digestibility. The lower and upper limits for these parameters, determined for six commercial crackers, are also represented in Figure 7.

In Vitro Digestibility
Microalgae crackers showed IVDMD values ranging from 82% to 86%, values which were slightly lower than that of the control sample (87.7%). This reduction was only statistically different (p < 0.05) for the 6% C. vulgaris and 6% P. tricornutum samples. It is worth nothing that all the values fall within the range of values for the commercial crackers analyzed (78-93%).
Regarding IVPD, there was a wider dispersion of results, with commercial crackers' values ranging from 26% to 78% and the control sample having 75% IVPD. Crackers with 2% microalgae biomass showed IVPD values ranging from 50 to 66%, i.e., a bit lower than that of the control but still within the commercial samples range. However, the samples with 6% microalgal biomass showed higher IVPD values, particularly A. platensis (83%). On the other hand, 6% C. vulgaris sample showed the lowest IVPD (42%); the value was still within the range of that of the commercial samples, however. By comparing IVPD values with the initial protein content of the samples (Table 5), it can be observed that, for example, the 6% A. platensis sample, which has 14.3% protein content and 83% IVPD, yields 11.9 g digestible protein per 100 g cracker, which is considerably higher than that of the control (7.3 g digestible protein per 100 g cracker). This high protein availability reinforces the interest of this sample as a "protein source" food.
Several authors have previously reported that A. platensis biomass shows 10% to 20% higher IVDMD [17,74] and up to 50% higher IVPD [32] than C. vulgaris biomass. The different IVDMD and IVPD values between A. platensis and C. vulgaris biomass are probably related to differences in cell wall structure, considering that while cyanobacteria have thin cell walls composed by peptidoglycan, green algae such as C. vulgaris have thick and rigid cellulosic cell walls [75].  Microalgae crackers showed IVDMD values ranging from 82% to 86%, values which were slightly lower than that of the control sample (87.7%). This reduction was only statistically different (p < 0.05) for the 6% C. vulgaris and 6% P. tricornutum samples. It is worth nothing that all the values fall within the range of values for the commercial crackers analyzed (78-93%).

Sensory Evaluation
Sensory analysis trials were carried out on samples with 2% microalgae biomass. Figure 8 represents the average scores of the sensorial parameters as evaluated by the panel.
peptidoglycan, green algae such as C. vulgaris have thick and rigid cellulosic cell walls [75].

Sensory Evaluation
Sensory analysis trials were carried out on samples with 2% microalgae biomass. Figure 8 represents the average scores of the sensorial parameters as evaluated by the panel. The control sample showed high sensory scores (>4) and was preferred over the microalgae crackers, which was as expected for this type of highly innovative product. The texture of the microalgae crackers presented good sensorial scores which were similar to the control, which can be related to the instrumental texture, density, and aw/water content results, indicating a crisp and aerated texture. The A. platensis sample was preferred in terms of taste, smell, and global appreciation, while crackers with marine microalgae (T. suecica and P. tricornutum) showed the lowest scores for these parameters, as well as for color. The control sample showed high sensory scores (>4) and was preferred over the microalgae crackers, which was as expected for this type of highly innovative product. The texture of the microalgae crackers presented good sensorial scores which were similar to the control, which can be related to the instrumental texture, density, and a w /water content results, indicating a crisp and aerated texture. The A. platensis sample was preferred in terms of taste, smell, and global appreciation, while crackers with marine microalgae (T. suecica and P. tricornutum) showed the lowest scores for these parameters, as well as for color.
Interestingly, the addition of microalgae seems to affect the perception of saltiness, with a higher frequency of "unsalted" classification ascribed to algae crackers (20-33%) in relation to the control (10%) (Figure 9). Interestingly, the addition of microalgae seems to affect the perception of saltiness, with a higher frequency of "unsalted" classification ascribed to algae crackers (20-33%) in relation to the control (10%) (Figure 9). Regarding intent to buy (Figure 10), 40% of the panelists declared that they would certainly or probably buy the A. platensis samples. Regarding intent to buy (Figure 10), 40% of the panelists declared that they would certainly or probably buy the A. platensis samples. Figure 9. Responses of the sensory analysis panel tasters (n = 30) in terms of salt perception of crackers enriched with 2% microalgal biomass, as well as the control sample.
Regarding intent to buy (Figure 10), 40% of the panelists declared that they would certainly or probably buy the A. platensis samples. Figure 10. Responses of the sensory analysis panel tasters (n = 30) in terms of buying intention for crackers enriched with 2% microalgal biomass, as well as the control sample.
The lower sensory scores attributed to T. suecica and P. tricornutum crackers was agreed upon by more than 60% of the panelists, who declared that they would probably or certainly not buy this product. Batista et al. [17] have also reported a preference for A. platensis-enriched cookies in Figure 10. Responses of the sensory analysis panel tasters (n = 30) in terms of buying intention for crackers enriched with 2% microalgal biomass, as well as the control sample.
The lower sensory scores attributed to T. suecica and P. tricornutum crackers was agreed upon by more than 60% of the panelists, who declared that they would probably or certainly not buy this product. Batista et al. [17] have also reported a preference for A. platensis-enriched cookies in relation to C. vulgaris. Fradique et al. [76] have also reported lower sensory scores, including "strange flavor" detection for pasta enriched with the marine microalgae Isochrysis galbana and Diacronema vlkianum.

Conclusions
In this work, microalgae biomass was successfully used as an ingredient in artisanal wheat crackers. The microalgae crackers presented a stable and attractive crispy and aerated texture that was similar to the control and appreciated by sensory analysis panelists. Crackers with 6% A. platensis biomass seem particularly interesting due to their high protein content (making them eligible for a "source of protein" claim) and high antioxidant activity, while the 2% formulation presented high digestibility and attained high sensory analysis scores. The use of marine microalgae strains such as T. suecica and P. tricornutum is interesting due to the high phenolic content and antioxidant activity of the crackers, but low sensory scores and novel food regulation constraints postpone their utilization.