Role of Fibre Morphology in Some Quality Features of Fibre-Enriched Biscuits

Abstract

The effect of replacing 5 and 10% of the flour in a biscuit formulation with two wheat fibres (of different lengths) and apple fibre (differences in morphology and proportion of soluble fraction) was studied. All the fibres decreased the flour pasting properties. The longer wheat fibre produced the greatest increase in the G′ and G″ viscoelastic moduli of the dough compared to the control (no fibres added). The biscuit texture properties were measured using the 3-point break test and cone penetrometry: wheat fibre biscuits were more resistant to breaking while apple fibre produced a crumbly biscuit. Sensory analysis by a trained panel showed minor changes in the apple fibre biscuits compared to the control and greater hardness in the wheat fibre biscuits.

Keywords:

• Biscuit
• Fibres morphology
• Physical and sensory properties

INTRODUCTION

Nowadays, consumers are increasingly interested in healthy food products. In the 1980s, dietary fibre was identified as an important component of a healthy diet and the food industry began to look for palatable ways to increase the fibre content of its products.[1] In the past decade, research has documented connections between dietary fibre intake and decreased risks of chronic diseases.[2] For this reason, fibre has received much attention in both scientific and informative literature. However, the intake of fibre and fibre-containing foods remains low in many populations worldwide. Interest in foods with high fibre contents has increased in recent decades and the importance of this food constituent has led to the development of a large market for fibre-rich ingredients to be used in baked products, such as snacks, muffins, or a number of types of biscuits that are convenient to consume at breakfast.[3] Because of their diversity, bakery products are used as a source of different nutritionally-rich ingredients.[4] Biscuits are a very popular bakery item, which is consumed by nearly all levels of society. Some of the reasons for such wide popularity are their ready-to-eat nature, affordable cost, good nutritional quality, availability in different tastes, and longer shelf life.[5]

Dietary fibres can be classified into insoluble or soluble and each fraction provides specific physiological functions and nutritional benefits: insoluble fibre promotes the movement of material through the digestive system and soluble fibre helps to lower blood cholesterol and regulate blood glucose levels.[6] Both types have been used in biscuits to increase the daily intake of fibre, mainly using cereal brans, which are rich in insoluble fibre, and gums, such as pectin, which are soluble fibres.

Cereal bran as a source of fibre to replace flour in biscuits has been studied by many authors. Brewers’ spent grain has been studied in sugar biscuits[7] and in wire cut biscuits.[8] Gujral et al.[9] replaced part of the wheat flour with wheat bran and coarse wheat flour in biscuits; the latter replacement increased the sensory scores and lowered the fracture strength. Leelavathi and Rao[10] studied the replacement of flour by raw and toasted wheat bran, reporting that the former, up to 30%, could be used as a substitute for flour in the preparation of high-fibre biscuits without affecting the overall biscuit quality. Recently, Ellouze-Ghorbel et al.[11] used different sources of wheat bran (Triticum aestivium and Triticum durum) to enrich biscuits. Sudha et al.[4] reported on the influence of different cereal brans on the sensory quality of biscuits; they obtained acceptable biscuits by incorporating 30% of oat bran or 20% of barley bran into the formulation.

Some studies have been carried out using fruit fibres: apple fibre,[12] banana,[13] and mango dietary fibre.[5] Bilgiçli et al.[14] replaced wheat flour with apple fibre, lemon fibre, wheat fibre, and wheat bran; however, they did not compare the effect of the different solubilities of the fibres used. More recently, a new generation of fibres has been studied, including inulin, oat beta-glucan enriched fraction (BGEF), potato fibre, and resistant starch.[15] Aparicio-Sanguilan et al.[1] obtained biscuits with resistant starch from lintnerized banana and Tuohy et al.[16] added guar gum, obtaining biscuits with a prebiotic effect.

All of the papers mentioned above studied the physicochemical changes in the dough or the biscuit (as well as physiological changes) as being dependent on the source of fibre employed; in general, they did not mention the morphology of the fibre as a factor to take into consideration. Prentice et al.[7] and Örztürk et al.[8] stated that fibre length could be expected to have an important effect on the properties of the dough and the biscuits, but on comparing these two studies, the final conclusion about the influence of different fibre sizes in biscuits was not clear. The aim of the present work was to study the influence of dose, fibre length, and proportion of soluble fraction on short dough biscuits using an apple fibre and two wheat fibres of different lengths to replace part of the flour.

MATERIALS AND METHODS

Ingredients

The three fibres used (Vitacel, J. Rettenmaier & Söhne, Germany, kindly supplied by IRS Iberica, Barcelona, Spain) were wheat fibre (WF) of two different sizes (200 μm and 101 μm) and apple fibre (AF). The main fibre characteristics provided by the supplier are shown in Table 1. The flour employed was plain soft wheat flour suitable for biscuits (Golden Dawn, Allied Mills, Tilbury, Essex, UK). The flour data composition provided by the supplier was moisture 14.4 g/100 g flour, protein 9.2 g/100 g flour, and ash 0.6 g/100 g flour used in the different recipes were 100, 95, or 90 g, corresponding to replacement of 0 (no replacement), 5, or 10 g of the flour with the different fibres.

Table 1 Fibre characteristics as provided by the supplier

Fibre	Bulk density (g/l)	WBC* (g H20/g d.s.)	Average particle size (% max)	IDF/SDF* (%)
WF-200	75–100	8.3	>250 μm: 5%	94.5/2.5
			>75 μm: 5–30%
			>32 μm: 50–80%
WF-101	260–335	4.8	>250 μm: 1%	94.5/2.5
			>75 μm: 30%
			>32 μm: 50%
AF-401	450–465	5	>400 μm: max 0.5%	75/25
			>150 μm: max 40%
			>32 μm: max 80%
*WBC: Water binding capacity; IDF: insoluble dietary fibre; SDF: soluble dietary fibre.

The amount of flour used in the different recipes was 100.00, 95.00, or 90.00 g, corresponding to fibre replacement of 0 (no replacement), 5.00, or 10.00 g of the flour with the different fibres.

The codes employed to identify the samples were control (no flour replacement, no fibre) and 5 or 10% followed by the fibre code (WF-101, WF-200, or AF), meaning that 5.00 or 10.00% of the flour had been replaced with one fibre (Table 1); for example, 10%WF-200 means that 10.00% of the flour in the formulation was replaced with WF-200 fibre. A total of seven samples were prepared: control, 5%WF-101, 5%WF-200, 5%AF, 10%WF-101, 10%WF-200, and 10%AF. The remaining ingredients were (flour weight basis): (a) shortening 32.15%: non-hydrogenated vegetable fat for frying (Bako, Cullompton, Devon, UK); (b) powdered sugar 29.45% (British Sugar Plc, Bury St. Edmunds, Suffolk, UK); (c) milk powder 1.75% (Dairy Crest, Rayleigh, Essex, UK); (d) salt 1.05%; (e) sodium bicarbonate 0.35%; (f) ammonium hydrogen 0.2%; and (g) tap water was adjusted for each formula by a Farinograph in preliminary tests (control: 11.00%; 5%WF101: 11.60%; 5%WF200: 11.73%; 5%AF: 12.84%; 10%WF101: 12.17%; 10%WF200: 12.46%; and 10%AF: 13.36%). Water absorption range was between 50.2% for the wheat flour and 60.4% for 10%AF.

Biscuit Preparation

The flour and fibre were pre-blended for 15 min in a double cone mixer (KEK-Gardener, Macciesfield, Cheshire, UK). The fat, sugar, milk powder, leaving agents, salt and water were mixed in a mixer (Fit Hobart, Troy, OH, USA) for 30 s at low speed (60 rpm), the bowl was scraped down and they were mixed again for 3 min at a higher speed (255 rpm). The flour or the flour/fibre mix was then added and mixed in for 20 s at 60 rpm then, after scraping down the bowl once more, for a further 40 s at 60 rpm. After a resting period in a plastic bag, the dough was sheeted and moulded to 6.5 cm diameter × 0.5 cm thick (with docking and logo) in a single step using a RTech Minilab sheeting line (Rtech Ltd., Warrington, UK); 15 biscuits were placed on each 48 × 21.5 cm perforated tray and baked in a tunnel oven (Spooner, Ilkley, West Yorkshire, UK) with two sections at different temperatures, 220 and 200°C, for a total of 5.3 min.

Flour Pasting Properties

The pasting properties of the flour and the flour/fibre mixes were measured using a Rapid Visco Analyser (RVA-4, Newport Scientific, Warriewood, New South Wales, Australia). Flour or a flour/fibre mix (3.5 g) was added to 25 mL of water. Aliquots of 3.28 and 3.22 g of flour alone (representing the removal of 5 and 10% of the flour, respectively) in 25 mL of water were also tested.

Rapid initial stirring was carried out by applying a 960 rpm stirring step for the first 10 s of the test, followed by decreasing and stabilizing the stirring to 160 rpm for the rest of the test. The temperature profile consisted of an initial holding time of 1 min at 25°C, raising the temperature to 95°C at a rate of 14°C/min, a holding stage at 95°C for 3 min and lowering the temperature to 25°C at a rate of 14°C/min, followed by a final holding period of 2 min at 25°C. The paste viscosity was expressed in centipoises (cP; 12 cP ≅ RVU [rapid viscosity units]). The following parameters were determined: peak viscosity, through, breakdown, final viscosity, setback, peak temperature, and pasting temperature. At least two determinations were carried out in each sample.

Dough Rheology Properties

Strain sweep tests were performed using a strain-controlled ARES rheometer (TA Instruments, Crawley, West Sussex, UK) fitted with parallel plates (25 mm diameter, 2 mm gap). The measurements were performed at 25°C. The elastic modulus (G′), viscous modulus (G″), and critical strain (γ _c ) were obtained. Also, the complex modulus, |G*|□, was calculated as follows:

The rheological analysis was always carried out just after moulding the dough. Each formulation was prepared twice, on different days, and four samples of each preparation were measured. The results were expressed as the average of the eight determinations performed per formulation.

Biscuit Properties

Texture and sound emissions

The texture of the biscuits was measured using a Texture Analyzer TA.TX.plus (Stable Micro Systems, Godalming, UK). Data management was performed using Texture Exponent software (version 2.0.7.0; Stable Microsystems, Godalming, UK). Twenty replicates of each formulation were conducted.

Breaking strength

The biscuits were broken using the three-point bending rig probe (A/3PB). The experimental conditions were: supports 50 mm apart, a 20-mm probe travel distance, and a trigger force of 20 g. The max force (N), and the displacement at rupture (mm) were measured. An acoustic envelope detector (AED) coupled to the texturometer was used for sound recording; the experimental conditions were adapted from Varela et al.[17] The gain of the AED was set at 1. A Bruel and Kjaer free-field microphone (8-mm diameter) was calibrated using a Type 4231 Acoustic calibrator (94 and 114 dB SPL-1,000 Hz), was placed in a frontal position in order to obtain a better acoustic signal, at a distance of 4 cm and an angle of 45° to the sample. A built-in low pass (anti-aliasing) filter set the upper calibrated and measured frequency at 16 kHz. Ambient acoustic and mechanical noises were filtered by a 1 kHz high-pass filter. The AED operates by integrating all the frequencies within the band pass range, generating a voltage proportional to the sound pressure level (SPL). The data acquisition rate was 500 points/s for both force and acoustic signals. All tests were performed in a laboratory with no special soundproofing facilities at an ambient temperature of 22 ± 2°C. The texture and sound parameter measured for each formulation was the number of sound peaks and SPLmax (dB). Each sound graph was simultaneously displayed with the correspondent force/displacement graph and after choosing the real peaks (rather than those due to external noise), the statistics were calculated.

Cone penetrometry

The test was performed under the following experimental conditions: a test speed of 5 mm/s, a trigger force of 5 g, and a cone traveling a distance of 12 mm. The parameters obtained from the cone penetrometry probe were area under the curve (N.s), number of peaks at one second and force at one second (N).

Biscuits’ photographs

Two biscuits from each formulation were cut on a horizontal plane and the surface of the biscuit was removed. The exposed surface was photographed with a C-Cell imaging system (Calibre Control International, Campden & Chorleywood Food Research Association, Appleton, Warrington, UK).

Scanning electron microscope (SEM)

The fibres and biscuit samples were examined without any further preparation using a Carl Zeiss EVO 60 scanning electron microscope (Cambridge, UK); the pressure was controlled at 60 Pa. The biscuits were broken in two to obtain a fracture surface; then, the sample was trimmed behind the fracture surface to produce a strip of biscuit a few millimetres wide for examination. The biscuit sample, with the fractured surface uppermost, was stuck to an aluminium SEM stub with silver ‘DAG’ cement.

Dimensions

Biscuit thickness (height) was measured by stacking 10 biscuits vertically against the biscuit thickness ruler, sliding the gauge to rest on top of the pile and recording the average thickness (AACC 10-50D). Biscuit width (diameter) was measured by arranging the biscuits along the length ruler with the stamped word parallel to its long edge and recording the average length; the biscuits were then rearranged with the written word perpendicular to the long edge on the ruler and the average ‘width’ was measured. Both measurements were expressed in cm (the measurements corresponding to 10 biscuits was divided by 10); two replicates were performed.

Moisture and fibre determination

The moisture content (%) of the flour, flour/fibre mixtures and biscuits was determined according to Approved Method 44-15.02.[18] The total dietary fibre content of the doughs and biscuits was determined in three replicates by Approved Method 32-07.01,[18] using a FOSS Fibretec E 1023 filtration module and Shaking Water Bath 1024 system (FOSS, Hilleroed, Denmark).

Sensory Analysis

Selection of terms and panel training

A panel of eight assessors (between 25 and 38 years old) skilled in quantitative descriptive analysis (QDA) was trained to select the descriptors using the checklist method. Terms were selected and discussed in an open session with the panel leader. The assessors were given a brief outline of the procedures and a list of attributes and representative samples, and were asked to choose and write down the most appropriate attributes to describe all the sensory properties of the biscuits, or to suggest new ones. The panel leader collected and wrote all the attributes on a board. The panel then discussed the appropriateness of the selected attributes, their definitions, and the procedures for assessing them. At the end of the session a consensus on the list of attributes (colour; thickness; flour, butter and toast odour; visual structure from compact to flacky; manual hardness; crumbliness; fragility; hardness; crunchiness; doughy, flour, butter and apple taste) and procedures had been chosen; this procedure was proposed by Stone and Sidel[19] in order to obtain a complete sensory description of a product. The panelists attended 12 training session of 1 h each. Training involved two stages: in the first stage, different samples were tested by the panelists to gain a better understanding of all the descriptors and different tastings were carried out until the panel was homogeneous in its assessments, as explained below under Statistical Analysis. In the second stage, the panelists used 10-cm unstructured scales to score the intensity of the selected attributes. The assessors were instructed to score the external appearance first, followed by odour, then the manual properties of the biscuit, then in-mouth texture and, finally, taste. The panel's performance was evaluated by principal component analysis, using the Pearson correlation matrix, until there were no outliers in the group between the different training sessions.

Formal assessment

A balanced complete block experimental design was carried out in duplicate (two sessions) to evaluate the samples. The intensities of the sensory attributes were scored on a 10 cm unstructured line scale. Seven samples were evaluated per session. In each session, the samples were randomly selected from each cooking batch and served in random order, each on a separate plastic tray identified with random three-digit codes. The panelists were instructed to rinse their mouths with water between sample evaluations. Testing was carried out in a sensory laboratory equipped with individual booths. Data acquisition was performed using Compusense five release 5.0 software (Compusense Inc., Guelph, Ontario, Canada).

Statistical Analysis

Analysis of variance (one-way ANOVA) was applied to study the differences between formulations; least significant differences were calculated by the Tukey test and the significance at p < 0.05 was determined. These analyses were performed using SPSS for Windows Version 12 (SPSS Inc., Chicago, IL, USA). For each descriptor, two-way ANOVA was applied to check panel performance considering assessors, samples, and their interaction as factors. Analysis of variance (one-way ANOVA) was applied to the trained panel in order to study the effect of formulation; least significant differences were calculated by Tukey's test (p < 0.05).

RESULTS AND DISCUSSION

Fibre Morphology

The morphology of the different fibres was studied by SEM. The two wheat fibres both had an elongated shape but were different in length, WF-200 being longer than WF-101 (Figs. 1a and 1b, respectively); the apple fibre had a completely different morphology, as its particles were shorter and rounded in shape (Fig. 1c). These observations were in accordance with the size data provided by the supplier (Table 1). Mongeau and Brassard[20] studied the water holding capacity of insoluble dietary fibres in relation to particle size and affirmed that at a lower particle size, fewer pores and holes remained in the wheat fibre to hold water; in accordance with this, the longer wheat fibre (WF-200) should be able to bind more water than the shorter wheat fibre. Chaplin[21] investigated the interaction between dietary fibre structure and water binding and affirmed that fibre can bind water directly in a number of ways. For instance, polysaccharide single chains are able to interact with other chains to form junction zones, which enclose large amounts of water, strongly bound and fairly static. They found that these junction zones were denser and less flexible than single chain zones. This situation effectively increases the fibre diameter of the junction zones, reducing the pore size, increasing the capillarity and reducing the tendency for the water to exit the cavities. That would indicate that WF-200 has a higher capacity to bind water than WF-101, as the long chains could form more junctions, entrapping the water better.

Figure 1 Microphotographs of the different fibres: (a) wheat fibre 101; (b) wheat fibre 200; (c) apple fibre. (Scale = 20 μm.)

The morphology of the apple fibre (AF-401) indicated a completely different scenario. It does not contain long or long-shaped fibres but rounded shape particles. However, it is necessary to consider that while wheat fibre is almost totally insoluble, apple fibre contains 25% soluble fibre (pectin), which is highly hydrophilic[6] and this would certainly affect its performance in the dough. Thibault and Ralet[22] reported that the hydration properties of fibres depend on their chemical and physical structures (surface area and particle size) and on their processing history.

In turn, all of these factors that could affect the water status in the dough would affect its technological characteristics (such as handling properties) and the final properties of the biscuits. While water is a minor component in a biscuit dough formula, it plays an important role during dough and biscuit preparation.[23] The water level used in the recipe affects gluten development in the dough, biscuit spread during baking, moisture retention, and the eating quality of the finished product.[24] In the present study, the fibres absorbed high amounts of water preventing proper hydration of the ingredients and making the dough powdery or causing it to fall apart.

Flour Pasting Properties

The parameters generated by the RVA describe starch swelling, gelatinization, and gelling ability. During one RVA cycle (heating–temperature maintenance–cooling), the starch granules in excess water swell and absorb moisture, leading to disruption of the hydrogen bonds and leaching of amylose from the starch granules. Finally, during the cooling cycle stage, the starch molecules reassociate and form a gel.[15] The changes in rhelogical properties observed during starch gelatinization depend on the presence of swollen starch granules in a dispersed amylose-amylopectin matrix and on the amylose-amylopectin interaction,[25] which would also be affected by the addition of fibre.

The maximum peak viscosity was obtained with flour alone (100% = 3.5 g). In the samples containing 5 and 10% less flour (3.28 and 3.11 g of flour, respectively, in the same amount of water), the peak viscosity values decreased as expected. When these percentages of flour were replaced with any of the fibres, the values recovered to some extent, varying according to the type of replacement (Table 2). The closest values to the 100% flour sample corresponded to 5%WF-200, where they were higher than for the 5%-less flour-only sample, so this fibre makes the mix more viscous, probably by absorbing part of the available water. This sample was followed by 5%AF and 5%WF-101. While the composition of WF-200 and WF-101 was the same, their differing physical shapes led to different behaviour. The samples with 10% replacement followed the same trend (WF-200 > AF > WF-101), but in this case they all presented higher peak viscosity values than the flour-only sample with 10% less flour. All the pasting parameter values indicated that the effects on these properties of replacing part of the flour with fibres could be explained by the different WBCs of the fibres, which make less water available for the flour. Brennan and Samyue[15] studied flour replacement with resistant starch (RS) and inulin; as the level of these two fibres rose, the peak viscosity values fell, an effect that these authors attributed to a reduction in the amount of gelatinizable starch. In the present case, when the flour was replaced with fibre, the pasting values were slightly higher than those for the reduced flour alone samples (5 or 10% less), so although the starch content was lower, the fibre helped to recover (and in some cases exceed) the viscous properties of the starch (flour) that had been replaced.

Table 2 Pasting properties of the flour and flour/fibre mixtures

Ingredients	Flour/fibre (total = 3.5 g)	Peak viscosity* (cp)	Trough* (cp)	Breakdown* (cp)	Final viscosity* (cp)	Setback* (cp)	Peak T* (°C)	Pasting T* (°C)
Flour	3.5/0.00	133.75^e (0.6)	64.46^de (2.0)	69.29^e (1.4)	136.71^cd (1.9)	72.25^cd (0.01)	5.60^a (0.01)	85.18^a (0.67)
Flour (5% less)	3.28/0.00	120.79^d (0.9)	60.17^bc (1.2)	60.63^d (0.3)	133.67^c (2.8)	73.50^d (1.65)	5.60^a (0.09)	86.40^b (0.01)
Flour (10% less)	3.11/0.00	97.96^a (0.3)	52.00^a (0.1)	45.96^a (0.3)	113.88^a (0.2)	61.88^a (0.18)	5.50^a (0.05)	87.60^c (0.57)
5%WF-200/flour	3.28/0.17	124.29^e (0.9)	65.58^e (0.1)	58.71^d (0.9)	136.17^cd (0.1)	70.58^bcd (0.01)	5.57^a (0.05)	86.45^bc (0.07)
5%WF-101/flour	3.28/0.17	117.63^cd (1.4)	59.00^b (0.6)	58.63^d (0.8)	126.21^b (1.5)	67.21^b (0.88)	5.60^a (0.01)	86.35^ab (0.01)
5%AF/flour	3.28/0.17	120.00^cd (0.7)	62.00^bcd (0.6)	58.00^d (0.1)	130.25^bc (1.3)	68.25^bc (0.71)	5.43^a (0.05)	86.28^ab (0.04)
10%WF-200/flour	3.11/0.34	117.00^c (0.6)	63.67^cde (0.1)	53.33^c (0.7)	130.21^bc (2.5)	66.54^b (2.42)	5.47^a (0.01)	86.38^b (0.04)
10%WF-101/flour	3.11/0.34	101.04^a (1.0)	53.46^ab (1.1)	47.58^ab (0.1)	112.54^a (1.2)	59.08^a (0.12)	5.47^a (0.01)	87.20^bc (0.14)
10%AF/flour	3.11/0.34	109.21^b (0.5)	59.08^b (0.2)	50.13^b (0.8)	126.88^b (0.3)	67.79^b (0.06)	5.47^a (0.09)	86.35^ab (0.01)
Values in parentheses are standard deviations. Means in the same column without a common letter differ (p < 0.05) according to the Tukey test.
*Peak viscosity: the maximum viscosity developed during the heating portion of the test; trough: minimum viscosity after the peak normally occurring around the start of sample cooling; breakdown: peak viscosity minus trough viscosity; final viscosity: the viscosity at the end of the peak test; setback: final viscosity minus trough viscosity; peak temperature: temperature at which peak viscosity occurred; pasting temperature: temperature at which viscosity first increased by at least 25 cp over 20 s.

Dough Properties

Dough properties depend on the contributions of the different ingredients, such as starch, proteins, and the water present, which in turn influences the handling properties. If the dough is too firm or too soft, it is not easy to handle; the dough must be sufficiently cohesive to hold together during the different processing steps and viscoelastic enough to separate cleanly when cut by the mould.[9]

Rheological Properties

The G′ and G″ values were independent of the applied strain up to a critical value ( _c ), which defines the onset of non-linear response. In all the doughs, G′ was always higher than G″ (Table 3). The replacement of flour by all the fibres resulted in an increase in both G′ and G″ values, the highest increase being found with 10% replacement by WF-200; 5%WF-200, 10%WF-101, and AF doughs showed similar elasticity (G′) results. Therefore, the main difference was that the longest fibre (WF-200) gave the dough the most elasticity. A decrease in tan delta (values closer to 0) was also found for the 10% replacement with WF-200, implying greater predominance of the elastic as opposed to the viscous component. The effect of the fibres on the onset of non-linear response (γ _c ) was only significant for the 10% fibre addition level, which reduced the γ _c values, meaning that the sample became less resistant to the applied strain.

Table 3 Influence of different fibres and flour replacement levels on linear viscoelastic properties during strain sweeps at 25°C. The values obtained were: G′ (storage modulus), G″(loss modulus), |G*| (complex modulus), tan delta, and critical strain γ _c

Sample	G′	G″	|G*|	Tan delta	γ c
Control	223,483a (18,567)	128,615a (12,114)	257,853a (22,117)	0.58d (0.01)	0.054bc (0.01)
5%WF-101	277,455b (22,114)	154,886ab (12,623)	317,773ab (25,253)	0.56cd (0.01)	0.058c (0.00)
5%WF 200	359,857c (24,016)	195,714c (14,302)	409,544cd (27,855)	0.54bc (0.01)	0.051abc (0.01)
5%AF	226,500a (21,079)	129,250a (12,632)	260,871a (24,436)	0.57cd (0.01)	0.059c (0.00)
10%WF-101	365,914c (16,145)	196,430c (7,204)	415,346cd (16,385)	0.54b (0.02)	0.039ab (0.01)
10%WF-200	661,200d (35,968)	321,200d (13,627)	751,351d (83,974)	0.49a (0.02)	0.037a (0.01)
10%AF	317,000bc (32,501)	169,714bc (19,533)	359,661bc (37,443)	0.53a (0.02)	0.045abc (0.01)
Values in parentheses are standard deviations. Means in the same column without a common letter differ (p < 0.05) according to the Tukey test.

Biscuit Properties

Many chemical and physicochemical reactions occur during baking, like protein denaturation, some loss of the starch's granular structure, fat melting, Maillard reactions and browning, dough expansion, water evaporation, and the production and thermal expansion of gases.[26] After baking, the dough will be transformed into a solid structure where each ingredient has different roles: the flour influences water binding and limits the dough's expansion, fat gives aeration and lubrication and interferes with gluten development and sugar influences dough viscosity and gluten development.[23] In the present work, the flour was replaced by fibres with different properties and compositions; in consequence, the dough properties had changed and the properties of the biscuit would conceivably also be affected.

Table 4 Instrumental texture characteristics, composition data, and dimensions of control and fibre-added biscuits

	3-Point break test			Cone penetrometry			Composition data		Dimensions
Sample	Max force (N)	Rupture (mm)	No. of sound peaks	Area (N/s)	No. of peaks at 1 s	Force at 1 s (N)	Moisture (%)	Fibre (%)	Height (cm)	Length (cm)	Width (cm)
Control	13.70ab (4.36)	0.38a (0.2)	3.17abc (1.5)	20.4bc (6.7)	15.40cd (2.41)	13.53a (2.95)	3.75c (0.03)	5.90a (0.55)	0.77a (0.01)	6.75a (0.02)	6.80a (0.02)
5%WF-101	17.42bc (2.33)	0.42a (0.1)	2.77ab (1.8)	25.8cd (5.6)	12.79b (2.15)	18.94bc (3.99)	3.54a (0.02)	8.96 (0.03)	0.78a (0.05)	6.76a (0.03)	6.75a (0.01)
5%WF-200	23.24d (5.32)	0.53a (0.1)	2.91ab (1.4)	21.5bcd (7.7)	13.77bc (2.33)	18.15bc (1.96)	3.65b (0.03)	8.86 (0.14)	0.79a (0.04)	6.66a (0.01)	6.74a (0.07)
5%AF	14.16ab (2.99)	0.44a (0.1)	5.10c (2.0)	13.0a (3.9)	16.89d (2.08)	14.23a (3.72)	3.67bc (0.02)	9.70 (0.71)	0.75a (0.03)	6.74a (0.01)	6.80a (0.01)
10%WF-101	21.03cd (4.37)	0.48a (0.1)	3.63abc (1.4)	27.7de (8.8)	12.35b (3.30)	16.48ab (4.86)	4.03e (0.02)	11.08 (0.86)	0.78a (0.07)	6.74a (0.02)	6.79a (0.01)
10%WF-200	24.18d (3.84)	0.52a (0.2)	1.94a (0.8)	29.8e (8.5)	9.71a (2.35)	21.60c (2.69)	3.90d (0.03)	11.08 (0.86)	0.79a (0.01	6.70a (0.06)	6.78a (0.02)
10%AF	12.39a (2.54)	0.41a (0.2)	4.25bc (1.8)	15.5ab (8.8)	14.86cd (1.68)	16.48ab (5.05)	3.70bc (0.08)	11.67 (0.32)	0.75a (0.02)	6.75a (0.03)	6.75a (0.04)
Values in parentheses are standard deviations. Means in the same column without a common letter differ (p < 0.05) according to the Tukey test.

3-Point break test

All the biscuits fractured under tension and the fracture took place relatively close to their central zone (in the lower surface), where the maximum stress occurred. This implied that the condition regarding the distance between supports was satisfied. The curves obtained from the 3-point bending test (not shown) were similar to others previously reported for materials with brittle fracture patterns:[27] they were characterized by an initial elastic response followed by a small fracture strain. Significant differences were found in the values for maximum force (hardness) and displacement on breaking into two pieces, being higher difference the biscuits containing WF-200, followed by WF-101 from the control and 5%AF biscuits (Table 4). The samples with AF presented similar maximum force values to those of the control sample. Consequently, a clear relationship was found between fibre size and hardness: the greater the fibre size, the higher force required to break the biscuit. This was observed for both levels of fibres replacing part of the flour. The number of sound peaks, an indication of crispness, showed that the wheat fibre reduced this attribute (at both replacement levels), which means that there was less microcracking and probably denotes a more compact biscuit matrix, whereas the apple fibre retained or increased biscuit crispness. This result is of great technological importance when deciding which fibre to use to enrich a biscuit, as it shows that a smaller particle size will give a crisper, crumblier texture.

Displacement at rupture presented no significant differences between fibres (Table 4), indicating that all biscuits presented the same capacity for elastic deformation. Saleem et al.[27] reported a relationship between the moisture content and the curves obtained by 3-point bending, pointing out that for less moisture the load (N) was higher; in the present case, no relationship between moisture content and force was found with fibre enrichment; this discrepancy is probably because those authors studied changes in moisture higher than 1%, whereas in the present case the maximum difference between the moisture contents of the samples was 0.49%. Previous authors[4,9] have observed increased breaking strength when flour is replaced by fibre. Brennan and Samyue[15] used different soluble and insoluble fibres for flour replacement in biscuits, observing a slight increase in the breaking strength probe values for those containing potato peel (insoluble) and no increase for inulin and beta-glucans (soluble). However, none of these studies discussed the particle size of the fibres added.

Cone penetrometry test

The area under the curve could be an indication of the sample's resistance to cone penetration as well as of the toughness of the sample; 10%WF-200 was the toughest biscuit (Table 4). All the force deformation curves showed numerous peaks that could be understood as the breaking events that occurred when the probe passes though the layers within the product structure. A more compact structure will be reflected by fewer peaks and an airy or layered structure by a high number of peaks. In order to study the biscuit matrix properly, as well as the mean resistance of each biscuit, the number of peaks and the force at one second were also measured (Table 4). Fewer peaks and a higher force value were observed for the 10%WF-200 biscuit, while the AF fibre provided the biscuit with a high number of peaks, related to an airy structure and lower force value as the cone penetrated, which could be related to its having a brittle structure and being the easiest to fracture. The differences between WF-200 and WF-101 could be attributed to their different fibre size, as the long WF-200 could form chain entanglements, creating an internal network that provides strength and more resistance at the breaking point. Despite this, the WF (WF-101 and WF-200) samples were very similar to each other.

Biscuit C-Cell photographs

Baltsavias et al.[28] considered the biscuit a cellular solid, a model structure of connected beams or plates. In the present study, each peak found by the cone penetrometry was related to a microfracture in the biscuit matrix, so it could be said that the control biscuit and AF biscuit contained more of Baltsavias’ plates, created by air inside the biscuit matrix. In order to confirm this theory, the structure shown by the C-Cell images was studied. The photographs of 10%WF-200 proved it to have a highly compact structure, while the control had more air pockets and holes in the biscuit matrix (Fig. 2). Blaszczak et al.[29] explained that biscuit matrix properties are mainly determined by air spaces and fat globules. These observations are in accordance with the texture results obtained in the present study: more microcracking during cone penetrometry occurred in the control samples, favoured by air pockets, whereas the wheat fibres created a more compact structure. Images of the AF biscuits could not be obtained due to the brittleness of the samples, which made them impossible to cut perpendicularly without breaking into several pieces.

Figure 2 Photographs of biscuits' perpendicular cuts using C-Cell. (a) Control biscuit; (b) 10% WF-200 biscuit.

Scanning electro microscopy

No evident differences due to replacing flour with fibres were observed in the SEM photographs of the biscuits’ matrices (images not shown). The photographs showed two kinds of fat globules in the biscuit matrix, as described by other authors;[29] the matrix observed was mainly protein with embedded fat globules and starch granules.

Dimensions

No significant differences were found in the biscuits’ height, width or length due to the addition of fibres (Table 4). Some reduction in the biscuits’ dimensions due to fibre addition has been observed in previous works. Brennan and Samyue[15] suggested that the fibres may act as biscuit dough mixture stabilizers at up to 10% of replacement, enabling the reformulated biscuit dough to retain its diameter during baking. Higher replacement of flour (60%) with resistant starch has been related to a lower gluten content.[30]

Moisture and fibre content

The analysis showed significant differences in the final moisture content of the biscuits (Table 4). The doughs with the highest water loss were those containing 5% of all fibres, all biscuits having less moisture content than the control; at 10%, fibres demonstrated to have a water retention effect except the AF biscuits, which presented the same moisture content values at both fibre content levels that were not significantly different from the control sample. It would be attributed to its soluble fibre content, pectin, being unable to retain as much water as the solely cellulosic material of the wheat fibre. The fibre content (Table 4) showed the expected differences due to the percentages of fibre added.

Sensory Analysis

The wheat fibres were white powders with a neutral flavour and odour; the apple fibre had a slightly fruity odour and a beige-brown colour. Figure 3 shows the scores obtained, displaying only the attributes which presented significant differences compared to the control. The trained panel scored the 5%WF of both lengths (101 and 200) as being less crunchy. WF-200, the longer fibre, produced biscuits that presented the more different profile compared with the control; they were lighter in colour, harder (manual and in-mouth) and with a more doughy mouthfeel (Fig. 3a). For the biscuits with 10% of the flour replaced by the two wheat fibres (WF-200 and WF-101), the sensory profiles were even more different from the control biscuit (Fig. 2b); at this level of replacement, the profile of the WF-101 biscuits moved further away from that of the control sample. Comparing the wheat fibre samples at the same percentage of replacement, the longer fibre (WF-200) always produced biscuits that differed more from the control than the shorter one (WF-100). These results were in agreement with the instrumental texture results. The sensory textural profile of the AF biscuits presented no significant differences compared to the control, as was expected in view of the instrumental measurements. However, the colour and taste attributes did differ (Fig. 3c): the AF samples were darker, with a toast odour which masked the floury and butter taste and with a certain apple taste.

Figure 3 Mean descriptive sensory scores for biscuits: (a) control, 5%WF-200, and 5%WF-101; (b) control, 10%WF-200, and 10%WF-101; (c) control, 5%AF and 10%AF. (Colour figure available online.)

CONCLUSION

This study has shown the influence of the morphology of the different fibres on the dough and biscuit matrix properties, making proper selection of fibre size and shape essential when formulating a fibre-enriched biscuit. The apple fibre needed more water to reach the correct water level for dough handling, but both the AF fibre-enriched biscuits and control had similar final water content. The texture properties of the AF biscuits were more similar to the control than those with wheat fibre, so using of AF fibre, which has a high proportion of soluble fraction, would be recommended. On the other hand, the apple fibre gave the biscuits a fruity taste. The biscuits with wheat fibre were neutral in flavour but harder in texture; this was attributed to the high water binding capacity of these elongated fibrils, which created a compact biscuit matrix structure. The medium length wheat fibre (WF-101) biscuits were not as hard as those with the longer wheat fibre (WF-200), so the medium length would be recommended over the longer one.

ACKNOWLEDGMENTS

The authors are grateful to the Spanish Ministry of Science and Innovation for financial support (AGL2009-12785-C02-01); to J. Rettenmaier & Söhne and IRS Ibérica for kindly supplying the fibre samples; and to the Consejo Superior de Investigaciones Científicas (CSIC) for the grant awarded to the author Laura Laguna. The authors would also like to thank Mary Georgina Hardinge for assistance with the English manuscript.