Introduction

Wheat proteins have the unique properties of developing a viscoelastic matrix when wheat flour is mechanically mixed with water. This viscoelastic network enables the dough to hold the gas produced during the fermentation process, leading to an aerated crumb bread structure. Unfortunately, gluten must be kept apart from the diet of celiac patients, who suffer very important intestinal damage when they ingest gluten-containing products. This technological obstacle has been overcome by using complex bread recipes with different starches and cereal flours like corn starch, brown rice, soy and buckwheat flour (Gallagher et al. 2004; Moore et al. 2006), or a composite blend of rice flour with corn and cassava starches obtaining gluten-free bread with a well-structured crumb and pleasant flavour and appearance (Sanchez et al. 2002; Lopez et al. 2004). Generally, in the performance of gluten-free bread, a variety of hydrocolloids or gums have been used for creating a polymer network with similar functionality than the wheat gluten proteins. In fact, gluten-free breads have been successfully developed using several combinations of cellulose derivatives (Gujral and Rosell 2004a, Schober et al. 2007). With the same purpose, crosslinking enzymes (glucose oxidase and transglutaminase) have been used as processing aids for improving rice-based gluten-free bread quality (Gujral and Rosell 2004a, b; Moore et al. 2006). Lately, different proteins have been proposed as alternative for both playing the polymer role and increasing the nutritional value of gluten-free products (Marco and Rosell 2008a, b, c).

It is clear that the common player when gluten-free breads are developed is the presence of a polymer with certain viscoelasticity and ability to entrap the other components of the system; and usually they are incorporated as ingredients of the recipe. Nevertheless, an attractive alternative would be to perform gluten-free breads by using only gluten-free cereals and to generate ‘in situ’ during the breadmaking process the required holding biopolymer.

Broa is Portuguese ethnic bread made with more than 50% of maize mixed with wheat or rye flours, highly consumed in the north and central zones of Portugal (Brites et al. 2007a). Bread making process is mainly empirical and several types of broa are produced depending on maize types and blending flours, although local maize landraces are usually preferred (Vaz Patto et al. 2007). Maize flour for breadmaking was traditionally obtained in stone wheel mills, moved by water or wind, and nowadays frequently by electricity. There are many recipes to prepare broa, but the traditional process (Lino et al. 2007) involves adding maize flour (sieved whole meal flour), hot water, wheat flour, yeast and leavened dough from the late broa (acting as sourdough). After mixing, resting and proofing, the dough is baked in a wood-fired oven. This empirical process leads to an ethnic product highly accepted for its distinctive sensory characteristics. Nevertheless scarce scientific studies on broa breadmaking have been reported, and research have been focused on the partial replacement of wheat flour by maize flour (Martiìnez and el-Dahs 1993) or maize starch (Miyazaki & Morita, 2005) or developing formulations based on maize starch (Sanni et al. 1998; Özboy 2002).

Maize is a gluten-free cereal, thus suitable to produce foods addressed to celiac patients. The acquired knowledge on broa (made from composite maize–rye–wheat flour) is important for facing the challenges in producing gluten-free bread that usually exhibits compact crumb texture and low specific volume (Rosell and Collar 2007; Rosell and Marco 2008). Therefore, a better understanding of this breadmaking process would provide the basis for developing gluten-free bread based on maize flour.

The objective of this study was to assess the impact of different factors as maize variety, type of milling, water mixing temperature on maize dough rheology and to define and optimize the maize breadmaking and to identify their effect on specific volume, texture and sensory quality of the maize bread performed by applying the technology of broa.

Material and Methods

Maize Flours Characteristics

Four maize varieties selected based on genetic background (Moreira 2006) were used in this study (Table 1). Whole meal maize flour was obtained after milling the selected varieties in artisan water-mill and electric-mill (model M-50, Agrovil, Portugal), both having millstones. Whole meal flour was sieved through 0.5-mm screen, and larger particles were discarded to obtain maize flour.

Table 1 Characteristics of maize varieties used in this study
Full size table

Flour protein and ash content were determined in triplicate following ICC standard 105/2:1994 and 104/1:1990 methods. Apparent amylose content was determined following the ISO 6647–2:2007 using 720 nm as wavelength, and a calibration curve previously performed with maize flour samples according to ISO 6647–1:2007.

Viscosity profiles were obtained with a Rapid Viscosity Analyser (RVA, Newport Scientific, Australia), according to Almeida-Dominguez et al. (1997) at 15% solids, using the following time (min): temperature (°C) settings 0:50, 2:50, 6.5:95, 11:95, 15:50, 25:50, the time (min):speed (rpm) programme were 0:960, 10:160. Maximum, minimum (or trough) and final viscosities (cP units) were recorded and the breakdown calculated as maximum viscosity-minimum viscosity.

Maize flour colour was determined on 10–12 g of sample in an opaque recipient by using a Minolta Chromameter CR-2b. Maize flour tristimulus colour parameters included: L*—lightness, a*—red/green index, b*—yellow/blue index and ΔE—colour total variation relative to a white surface reference (L* = 97.5 a* = –0.13 b* = 1.63). Values are the mean of ten replicates.

Dough Rheological Properties

Dough rheological properties were evaluated in a conventional Brabender Farinograph® (Duisburg, Germany) following the method ISO 5530–1:1997 with some modifications. Maize flour (40 g) were mixed in the Farinograph 50-g bowl with 44 mL of distilled water (110% flour basis) during 20 min. Assays were carried out under two different conditions, at 25 °C and by adding boiling water (100 °C). The parameters obtained from the farinogram included development time in minutes ((Td), time to reach the maximum consistency), the dough consistency at the Td (CTd) and the consistency after 20-min mixing (C20), both in BU (Brabender Units).

Mechanical and surface-related properties were determined in the resulting dough, either from 25 °C and 100 °C. Dough machinability was determined by assessing the texture profile analysis (TPA) and dough stickiness in a TA-XT2i texturometer (Stable Micro Systems, Godalming, UK) as described by Armero and Collar (1997) using the Chen & Hoseney cell. Primary textural properties were measured in absence of dough adhesiveness by using a plastic film on the dough surface to avoid the distortion induced by the negative peak of adhesiveness (Collar and Bollaín 2005). The adhesiveness was measured without the plastic film. Three and ten repetitions for the TPA parameters and stickiness were done, respectively. Compression test was performed with a 50 mm of diameter cylindrical aluminium probe, a 60% compression rate followed of 75-s interval. TPA profile recorded the following parameters: hardness (g/force), adhesiveness (g/s), cohesiveness and springiness. For dough stickiness (g/force) determination was used the Chen & Hoseney cell with a cylindrical probe of 25-mm diameter (Armero and Collar 1997).

Breadmaking Process

The traditional broa formulation included 70% of maize flour, 20% of commercial rye flour (Concordia type 70, Portugal), 10% of commercial wheat flour (National type 65, Portugal), 95% (v/w, flour basis) of water, 3.6% (w/w, flour basis) sugar, 2.2% (w/w, flour basis) salt, 0.5% (w/w, flour basis) of improver (S500 Acti-plus, Puratos) and 0.8% (w/w, flour basis) dry yeast (Fermipan, DSM, Holland). Sourdough was prepared using the same recipe of broa and adding enough bacteria suspension (Lactobacillus brevis and plantarum previously isolated) to yield 107 CFU/g mass concentration. Sourdough was kept at 25 °C during 12 h before its use. Traditional broa baking trials were performed with the four maize varieties milled with two different mills, which gave a total of eight different maize flours (n = 8).

Breadmaking process consisted in mixing the maize flour with 77% (v/w, flour basis) boiling water containing 2.2% salt for 5 min in the bowl of Kenwood kitchen machine. Dough was left idle till cooling to 27 °C, then the remaining ingredients (including 18% water containing 2.2% salt and 10% w/w flour basis of sourdough) were added and dough was kneaded again for 8 min and left resting for bulk fermentation at 25 °C for 90 min. After fermentation, the dough was manually moulded in balls of 400 g and baked in the oven (Matador, Werner & Pfleiderer Lebensmitteltechnik GmbH) at 270 °C for 40 min. For each trial, three samples were produced and analysed separately.

An adapted breadmaking process was carried out for obtaining gluten-free maize bread, in which rye and wheat flours were replaced by maize flour and recipe contained 110% (v/w, flour basis) of water, identical proportion of the other traditional broa ingredients (sugar, salt, improver, dry yeast). Gluten-free baking trials were performed with Pigarro and Fandango maize varieties milled in artisan water mills (n = 2).

Bread Analyses

Quality technological parameters of breads were determined the following day to its production. Quality parameters included: weight (g), volume (cm3) using polyethylene spheres displacement method (Esteller and Lannes 2005). Specific volume was then calculated in cm3/g.

The tristimulus colour parameters (L*, a*, and b*) of crumbs were determined using Minolta Chromameter Model CR-2b colorimeter.

Bread slices (25-mm thickness) were used for crumb firmness determination through compression test in a texture analyser (TA-Hdi, Stable Micro Systems, Godalming, UK) using adapted American Institute of Baking–AIB Standard Procedure (2007): 12.5-mm cylindrical probe, 2.0-mm/s test speed, 10-g trigger force, 6.2-cm compression distance after detecting resistance (crumb surface) and final speed test of 10 mm/s. Firmness in g-force was automatically recorded by the data processing software.

Sensory analysis (ISO 8587, 1988) was conducted with a panel of 12 trained judges that quantify the influence of different maize varieties on overall differentiation of broa and maize bread. Triangular assays (AACC 33–50A, 1999) were carried out for each maize variety subjected to the two types of milling. Paired comparison tests (ISO 5495, 1983) were conducted to compare different varieties (within each colour group—white or yellow), panelists were asked to rank the samples based on overall texture, taste and aroma. Traditional broa was compared with gluten-free maize bread.

Statistical Analysis

The effect of different flour (maize variety, type of milling) and dough (water mixing temperature) variables on respectively flour chemical composition, colour, viscosity profile, dough rheological and bread technological quality parameters were analysed by analysis of variance. Means comparisons were performed by Duncan’s test also used for compared traditional broa with gluten-free maize bread. Significant correlations between flour composition, viscosities and dough rheological parameters were determined with Pearson correlations analysis. All statistical analyses were conducted at a significant level of P ≤ 0.05 with Statistical Analysis System (SAS Institute, Cary, NC, 1999).

Results and discussion

Effects of Maize Varieties and Milling Types on Flour Composition, Colour and Viscosities

The viscosity profile of a wide germplasm collection of pure lines, hybrids and local maize populations was previously characterized by Santos (2006) and four varieties (Pigarro, Fandango, Yellow Hybrid and White Hybrid) (Table 1) were selected for bread production with and without composite rye–wheat flours. The effect of milling type on the flour and dough characteristics was also studied to assess the possible influence of the new practices (electrical mill) compared to the traditional ones (stone mill).

Significant differences between maize varieties were detected for protein and amylose contents (P < 0.05) (Table 2). Regional varieties (Fandango, Pigarro) exhibited significant higher protein content and lower amylose content than hybrids. Pigarro flour had the highest ash content probably due to its endosperm of flint type. Type of milling influenced significantly (P < 0.05) the ash content that affects pH profile during fermentation and, in turn, will influence bread quality. The type of grinding did not have any influence on protein and amylose content.

Table 2 Effects of maize variety (n = 6; 2 milling types × 3 replicates) and milling type (n = 12; 4 varieties × 3 replicates) on protein, ash and amylose contents of maize flours
Full size table

As expected, type of variety showed greater significance (P < 0.05) on flour colour parameters than milling type (Table 3). Despite of testing two yellow maize (Fandango, Yellow Hybrid) and two white maize (Pigarro, White Hybrid), there were significant differences (P < 0.05) concerning a* and b* parameters between the yellow varieties and only significant differences in a* values between the white varieties. The type of milling affected significantly (P < 0.05) the lightness (L) of the maize flours (data not shown).

Table 3 Effect of maize variety (n = 20; 2 milling types × 10 replicates) on colour parameters
Full size table

Viscosity profile of four maize flours during a heating–cooling cycle was recorded by using the rapid viscoanalyzer (Fig. 1). Compared with commercial wheat flour (results not shown) maize flour exhibited lower pasting temperature, lower maximum viscosity and higher final viscosity, therefore higher setback was obtained. Similar results were obtained by Martínez and el-Dahs (1993), who detected a reduction of the maximum viscosity and an increase in the final viscosities of the wheat flour when adding instant maize flour (up to 25%). When compared the viscosity profile of the different maize varieties, maximum and final viscosities values from hybrid varieties were significantly (P < 0.05) higher than those of the regional ones (Fandango, Pigarro) (Table 4). Flint maize varieties have harder endosperm than the dent varieties and their flours have distinct viscosity profile (Brites et al. 2007a, b). Fandango maize flour variety (regional dent type) presented superior values than Pigarro (regional flint type), agreeing to previous data (Santos 2006; Brites 2006). Previous findings reported that the flint maize shows lower maximum viscosity and lower setback than dent varieties (Almeida-Domingues et al. 1997; Sandhu et al. 2007; Brites et al. 2007b).

Fig. 1
figure 1

Viscosities prof of maize flours obtained from electric mill determined by RVA (Rapid Visco Analyser)

Full size image
Table 4 Effects of maize variety (n = 6; 2 milling types × 3 replicates) and milling process (n = 12; 4 varieties × 3 replicates) on RVA parameters: maximum, minimum and final viscosities and breakdown of flour
Full size table

The milling type variation influenced maximum viscosity (Table 4) and also breakdown, being the average values of the flour from water mill higher and significantly different than the results of flour from the electric mill. The electric milling process yielded flour with lower viscosity profile likely associated to the negative impact of damage starch on the ability to absorb water.

Effects of Maize Varieties, Milling Type and Water Temperature on Dough Farinograph and Texture Parameters

The behaviour of the dough during mixing and handling was analysed by using the Farinograph and texturometer respectively, considering the effect of the variety and milling type as well as the mixing water temperature. No influence of variety and milling type was detected (Table 5) on those parameters, with exception of the significant (P < 0.05) effects of the milling process on dough hardness.

Table 5 Farinograph and texturometer dough result parameters from four maize varieties, two types of milling and two mixing temperatures
Full size table

The temperature of the added water for making the dough was the major factor of variability (Table 6). Water temperature significantly (P < 0.05) affected the consistency of the dough and the stability of the dough (related to the consistency after 20-min mixing). Concerning dough machinability, the temperature of the dough did not significantly affect hardness, but resulted in a significant (P < 0.05) effect on adhesiveness, gumminess and stickiness of the dough. When boiling water was used for dough mixing, maize dough showed significantly (P < 0.05) higher consistencies with minor development times compared to doughs mixed at 25 °C dough. Associated with the increase of the water temperature was the increase of mechanical and surface-related parameters adhesiveness, elasticity, and stickiness and subsequent reduction of cohesiveness. These results were not unexpected since previous studies reported that dough rheological parameters were particularly affected by starch gelatinisation (Miyazaki and Morita 2005). The addition of boiling water to the maize flour promoted the partial gelatinisation of the starch, increasing the viscosity of the dough, consequently, leading to higher dough consistency. The gelatinisation occurs as the temperature rises, which increases mechanical strength of dough. This is an important factor to consider when maize flours are destined to gluten-free breadmaking obtaining a viscous system that holds the components of the system. In fact, Rosell and Marco (2007) observed a decrease in the peak of maximum viscosity after heating rice flour dough prepared by using heated water during mixing, due to the previous partial gelatinisation when warm water was added. Similar effects have been observed when the pasting characteristics of native and heat-moisture-treated maize starches were compared (Hoover and Manuel 1996). As a consequence of the initial starch gelatinisation, dough consistency increases, improving the mechanical and handling properties of the rice flour dough compared to those of the dough mixed with water at 25 °C (Marco and Rosell 2008a)

Table 6 Effect of water temperature (n = 8; 4 varieties × 2 milling types) on dough Farinograph and texturometer parameters
Full size table

Therefore, an alternative for improving gluten-free dough consistency is to promote the partial starch gelatinisation through the addition of boiling water when mixing. Relationships between flour composition and viscosity and dough rheological parameters were particularly significant for dough textural parameters vs flour parameters. Significant correlations (P < 0.05) between amylose and cohesiveness were detected (r = 0.72), whereas springiness and stickiness parameters were associated to gelatinisation and retrogradation phenomena (r > 0.71), as were previously found for wheat doughs (Collar and Bollaín 2005; Collar et al. 2007).

Effect of Maize Varieties and Milling Types on Bread Specific Volume, Colour, Firmness and Sensory Assessment

A preliminary breadmaking study was performed varying the temperature (25 °C or 100 °C) of the water added to the maize flour during mixing, broa obtained by adding water at 100 °C showed superior crumb texture quality than the ones obtained at 25 °C water temperature (results not showed). Further breadmaking trials were made following the traditional broa making procedure, using boiled water for mixing maize flour.

Traditional ethnic bread, Broa, was made for defining breadmaking conditions prior to the performance of gluten-free maize bread. The specific volume of the broa ranged from 1.40 to 1.57 cm3/g (Table 7), which could be considered low values if compared with wheat bread loaves. Traditionally, Broa is a type of bread with high density and closed crumb cells, thus high specific volume is not desirable. Besides breads made or containing high amounts of gluten-free cereals show low specific volume compared to the ones obtained with wheat flour (Marco and Rosell 2008a).

Table 7 Effect of maize variety and milling type on specific volume, colour parameters and crumb firmness of traditional broa
Full size table

Regarding the effect of maize varieties and milling type on the specific volume of broa, no significant differences were detected (Table 7). Significant differences (P < 0.05) were induced by maize varieties in the colour parameters and firmness, by contrast no significant differences (exception to blue/yellow parameter—b*) were obtained between flours obtained from water and electric mills. Maize varieties had a significant effect on the firmness of the bread crumb, being the crumbs from Pigarro maize variety significantly harder than the ones from Fandango.

Sensory triangular assays of broa showed no significant differences ascribed to the type of mill (data not shown). Sensory rank sums and paired comparison test of regional and hybrid maize varieties within each colour (white or yellow type) showed the preference of regional maize varieties in detriment of hybrids (22.0 vs 14.0 in the case of yellow types and 20.0 vs 16.0 in the case of white types). The judges defined Fandango variety broa with better characteristics of mouth feel flavour and texture, even though broas produced with the hybrid varieties had higher specific volume.

From the above results, Fandango and Pigarro varieties were selected for performing gluten-free maize bread, since they were the preferred varieties by the judges. The study was restricted to the maize flours from water mill, because milling type did not induce significant differences on the broa quality. The specific volume of gluten-free maize bread ranged from 1.02 to 1.12 cm3/g. As expected the gluten-free breads presented from 20% to 30% less specific volume than their counterparts produced with the traditional recipe (obtained from composite maize–rye–wheat flour). Sanni et al. (1998) obtained maize bread containing egg proteins with 0.95 cm3/g specific volume. Similar bread specific volume had been reported in rice-based breads, which were improved by using crosslinking enzymes (Gujral and Rosell 2004a, b), hydrocolloids (Marco et al. 2007) or proteins (Marco and Rosell 2008a).

Gluten-free maize bread displayed smaller volume with slightly more compact structure than the traditional broa, which showed defined gas cells in the crumb (Fig. 2). Gluten-free breads due to the absence of a protein network cannot retain the carbon dioxide produced during the fermentation, leading to a product with low specific volume and compact crumb (Rosell and Marco 2008), which has a close appearance resemblance to the Portuguese ethnic bread.

Fig. 2
figure 2

Crumbs of broa produced with traditional and gluten-free formulation. Upper pictures Fandango maize variety, and lower pictures Pigarro maize variety

Full size image

A comparison was made between the quality parameters of the broa and the gluten-free bread. Significant differences were detected in the colour and texture parameters between gluten free and traditional broa. As was expected, crumb firmness of gluten-free bread was significantly higher (+50%) than the one obtained in the traditional broa, which agree with the specific volume results obtained. Reduced loaf volume and firmer crumb texture of gluten-free bread when compared with traditional broa was attributed to maize gluten absence, as has been previously observed in other gluten-free bread recipes (Rosell and Marco 2008).

Sensory ordinance test showed contradictory results depending on the maize variety used for breadmaking. Significant differences (P < 0.05) between gluten free and traditional broa obtained from Pigarro maize flour were obtained in the sensory paired preference test, the sensory panel preferred traditional broa (22.0 vs 14.0). Conversely, in the case of Fandango (yellow variety), no significant differences (P > 0.05) were observed between the scores that received the traditional broa and the gluten-free maize bread. Fandango variety was sweeter than Pigarro and it seems to perform better in breadmaking process that includes sourdoughs.

Conclusions

Breads were obtained from maize and composite maize–rye–wheat flour, studying the effect of maize varieties, milling process and processing variables on the dough characteristics and bread quality. Significant differences between regional and hybrid maize were detected regarding protein, amylose and RVA viscosity profiles. Concerning the effect of milling process, the grinding in a water mill occurs at slower rate than in the electrical, obtaining flour with lower ash content and higher viscosities. Nevertheless the influence of milling type on flour parameters, no significant differences were detected in broa sensory triangular tests and ordinance tests had neglected hybrid maize in relation to the regional ones.

Baking assays demonstrated that broa breadmaking technology could be satisfactorily applied to produce gluten-free broa. An important point in the breadmaking process was the blanching that resulted in doughs with higher consistency, because in the absence or reduced amount of gluten the dough rheological properties are provided by the starch gelatinisation. Maize-based gluten-free bread were obtained following broa breadmaking process, obtaining bread with satisfactory sensory characteristics and similar appearance than the traditional broa.