Physical variables of grains

The flotation index was evaluated with the methodology by ^{Wichser, (1961)} using a sodium nitrate solution at room temperature with a specific gravity of 1.275. The thousand grain weight was calculated by weighing 100 grains of each variety with a digital Scout^® Pro scale (Ohaus Corporation, 194 Chapin Road. NJ07058. USA) and multiplying the result by 10. Grain and nixtamal moisture was evaluated using the 934.01 ^{AACC method (2000)}, absolute density with the ^{Kniep and Mason method (1989)}, and test weight with the ^{AACC 55-10 (2000)} technique. Grain size was determined by measuring its length, width and thickness with a digital gauge of 152.4 mm (6”) (Truper^®, Mexico, with approximation to 0.01 mm) taking a random sample of 10 grains of each variety.

Pericarp area

This variable was determined by soaking 10 grains of corn in a glass of water at room temperature for 3 min, then they were removed and excess water was dried with a cloth. Several vertical cuts were made in the grain to separate the pericarp with tweezers without breaking it. The extended pericarp was dried during one day at room temperature and then we scanned it (Epson Perfection V200 Photo, USA) to generate a digital image in TIFF format at a 300 dpi resolution. The image was changed to black and white to highlight the pericarp, and colored the area occupied by the pericarp of each grain area with the color selection function of program GIMP version 2.8. The amount of pixels corresponding to each grain pericarp was counted with the Histogram function. The value in pixels was converted to cm² using the image resolution, ie 300 dpi / 2.54 = 118 ppcm (ppi: pixels per inch, and ppcm: pixels per centimeter); thus, 118.11²= 13,950.02 pixels represent 1 cm². Therefore, the value determined in pixels by the software for each pericarp divided between 13,950.02 was reported as pericarp area per corn grain. The average values of each variety were used with average values of the weight of one thousand grains to calculate the area of pericarp per weight unit using the formula: Area per weight unit = (1000/weight of 1000 grains) x (area per grain ). The value was reported in cm² g^-1. These values allowed to display the full pericarp area existing in a same amount of weight and compare the area available for water and calcium ions absorption, both in the profile of water absorption at room temperature and during nixtamalization.

Water absorption profile

The profile of water absorption was determined at 10, 20, 30 and 40 min, for which 20±1 g corn were weighed and placed in 50 mL of water at room temperature using Pyrex^® 150 mL beakers. Grains were removed and excess water was eliminated with a manual plastic centrifuge for vegetables. Then we weighed grains and calculated the water absorbed per gram of corn performing two replicates each time and for each variety using the formula: Water absorbed = Pwater / Pcorn; where Water absorbed = amount of water absorbed per gram of corn; Pwater = weight of grain after being submerged in water and left there for (g) time minus the dry weight of corn (20±1 g) and Pcorn = 20±1 g.

Corn nixtamalization

The traditional method similar to that described by ^{Fernández-Muñoz et al. (2004)} was used for nixtamalization. Treatments were performed at Colegio de Postgraduados, Córdoba Campus, state of Veracruz, located at 645 masl, 18o51’32.65” N and 96° 51’ 36.96” W. For this purpose, 5 L drinking water and 20 g calcium hydroxide food grade (Oxical^®, Mexico) were powred in a stainless steel bowl without lid (to simulate conditions of commercial tortilla stores) and placed it in a stove made of stainless steel, equipped with an iron burner Mod. QH128 of 12.8 cm in diameter (Industrias Magaña L, Santa Ana, El Salvador) at a minimum flame.

Preheated water temperature was measured with a digital thermometer (Taylor^® Precision Products, USA) up to 80 °C. Then we added 2 kg of corn, and corn samples and cooking water were taken for 45 min at t₀, t₅, t₁₀, t₁₅, t₂₀, t₂₅, t₃₀, t₃₅, t₄₀ and T₄₅ times. Three repetitions of nixtamalization were carried out for each variety.

Cooking time was based on the total time requires for pericarp removal. The detachment of the pericarp was determined subjectively by selecting three grains at random from each sample taken every 5 min for 45 min. The grains were rubbed with the fingers trying to remove the pericarp from the grain and the result was recorded using a binomial scale of Yes (1) when the pericarp was detached, and No (0) when there was no removal of the pericarp.

Cooking water pH determination

Of the samples taken from the cooking water in periods of 5 min, we placed 50 mL in beakers PYREX^® 100 mL and measured the pH with an Orion 3-Star potentiometer (Thermo Scientific, Waltham, Massachusetts, USA). Three replicates per variety were carried out for each time range (t₀-t₄₅).

Grain cutting force and nixtamal

The cutting force of dry and nixtamalized grain was determined by a texture analyzer Shimadzu^® model Ez-5 (Kyoto, Japan), the Trapezium2^® software for Windows^® and equipped with a Warner-Bratzler knife designed to evaluate meat, but it is also used to evaluate products such as pasta and vegetables (^{Bibat et al., 2014}). The grain was placed transversely at a distance of 1.5 cm from the knife to determine the force required to cut the grain. The determination was performed at room temperature in compression mode, at a speed of 1 mm s^-1 in three grains at t₀ and three grains at each sampling time during nixtamalization (t₅-t₄₅).

Nixtamal drying process and flour production

The nixtamal samples were placed in tubes for an Evergreen^® centrifuge and sealed with self-adhesive film with small holes. Then we put them in a Savant Model SC210A Thermo Scientific Model 81 concentrator (Wyman Street Waltham, MA, USA) and dehydrated them at 45° C for 24 h. The grain was ground with a Krups Mod GX410011 mill (Col. Polanco 11560, Mexico, DF) and passed through a 60 mesh sieve to obtain flour.

Dough properties (viscosity analysis)

The dough properties were determined with a Rapid Visco Analyser (RVA series S4A Newport^® Sci. Unit, Australia). The sample was prepared by weighing 4 g of grain flour (t=0), and nixtamalized and adjusted to 14 % moisture on a scale (Dever ^®Instrument, USA); sufficient distilled water was added, up to 28 g and the sample was placed in the equipment. The temperature profile was: initial temperature 50 °C; it remained 1 min at 50 °C; temperature rose from 50 to 92 °C at a rate of 7.5 °C/ min; it remained 5 min at 92 °C; it fell from 92 to 50 °C at a rate of 7.5 °C/min (same heating rate); it remained 1 min at 50 °C, and the test ended at 22 min. With the computer software a viscosity curve over time and the temperature profile was generated. From this profile of viscosities, the peak viscosity and viscosity in centipoise retrogradation (cP) of the sample were obtained. For this test we used flours of times t₀, t₁₀, t₂₀, t₃₅, and t₄₅, of each variety in duplicate.

Thermal properties of nixtamalized corn flour

To determine the initial, maximum and final temperatures, and enthalpy of gelatinization of nixtamalized corn flour we used a differential scanning calorimeter (Mettler brand Mexico DF). The sample was prepared with 3 mg flour, then we placed it in a low pressure aluminum crucible and added 7 μL distilled water. The crucible was sealed and placed together with a reference sample without material in the plate heat transmitter. The heatingratewas10°Cmin^-1 from30°Cto100°C.Forthis test we used raw flour (t₀) and nixtamalized flours for t₂₀ and t₄₅. The test was performed in duplicate for each variety.

Determination of cooking heat

Heat required for nixtamalization

The required heat was calculated for each variety during nixtamalization. Nixtamalization was considered from the time the grain was poured into the alkaline solution heated until the grain pericarp detached (Figure 1). During this period and in periods of 5 min we calculated the heat levels required by corn and water, integrating all sections to obtain the heat required by grain and water during cooking. To the latter heat value the necessary heat was added during the water preheating stage to have the heat needed throughout the process. This was performed in triplicate for each variety. Temperature was held below 100 oC throughout the cooking time to prevent boiling, and only an average temperature of 90.1 °C was reached at the end of nixtamalization. For the energy calculation we discarded the heat required to warm up calcium hydroxide because the amount used was low (1 % of grain weight). To estimate heat we used a corn Cp = 2.27 kJ mol kg^-1 K^-1 which we obtained as average of the Cp values reported for corn of ^{ASAE/ASABE (1999)}. For water we used Cp= 4.18 kJ mol kg^-1 K^-1. Calculation of the sensible heat used in the system was performed using the equation: Q = mCpΔT, where m = mass of the material (kg); Cp = specific heat of the material (kJ mol kg-1 o K-1); ΔT = difference of the final temperature minus the initial temperature (oK).

Figure 1 Diagram of the temperature and time profile used for calculating heat during the cooking process. 

Statistical analysis

The experimental design was completely randomized for the characterization of corn varieties. For monitoring nixtamalization the experimental design was completely randomized in a factorial arrangement with two factors: Variety with five levels, and Time which depended on the times set in each test. Multiple comparisons between means were carried out with the Tukey test (p≤0.05). These statistical analyses were performed with SAS version 9.2 (^{SAS Institute, Inc. 2002}). In addition, the R package version 3.1.0 Spring Dance with the integrated development environment RStudio version 0.98.932 were used for linear regression to calculate the water absorption rate.

Physical characteristics of grain

The flotation index showed that the variety with the lowest value was Mont41 (19 %), and in other varieties values were close or equal to 100 % (Table 1). Mont41 had the highest test weight (79.74 kg hL^-1) and the Criollo had the lowest value. According to ^{Salinas Moreno et al. (2010)}, a low flotation index and high test weight are typical features of hard grain corn.

The Criollo corn variety had the highest 1000 grain weight and Mont265 had the lowest. This variable is a good indirect indicator to predict the size of grains (^{Lee et al., 2012}). Thus Criollo corn grains were larger and grains of Mont265 were smaller.

Grain moisture (Table 1) showed a range from 10.75 % to 11.48 %. Moisture differences between the varieties studied were not significant, indicating that the grains were homogeneous in this variable at the beginning of the study. Absolute density values (Table 1) showed an inverse relationship with the flotation index, thus confirming that a denser grain will have a lower flotation index (r=-0.908). The relationship between the test weight and the flotation index was inverse. In contrast, the density of corn and grain test weight indicated a direct relationship (r=0.969).

Dimensions of grain and pericarp area size

Table 2 shows the data of actual dimensions of grain for each variety. The Criollo corn had the highest values in grain size and Mont265 the lowest.

These results agree with those obtained in thousand kernel weight and also with the pericarp area of our experiment. The smaller pericarp area was found in Mont360 and is due to the width of this grain variety (7.58 mm). However, in the pericarp area per weight unit the opposite occurred since the grain of the smallest size (Mont265) showed a larger area per weight unit. This is related to the variable weight of 1000 grains, where for the same weight the number of grains is greater when these are smaller.

Water absorption profile by grain at room temperature

The variety that absorbed more water was the Criollo, contrasting with Mont41 and Mont360 that absorbed less (Table 3). However, the variety that absorbed more water at the beginning was Mont265 (Figure 2) which had a larger pericarp area per weight unit (Table 2). Despite this, the grain tended to stop absorbing water after 20 min, which can be due to its small size that led to its rapid saturation with water. The Criollo variety absorbed water for a longer time because, according to the flotation index and thousand kernel weight, the Criollo is a soft grain with greater porosity which allowed absorbing water even up to 40 min. The Mont363 variety had a similar behavior. The varieties Mont41 and Mont360 absorbed less water although their rate of water absorption was intermediate in relation to other varieties. ^{Bressani (2008)} indicated that the grain hardness is a variable linked to water absorption, because in corn varieties with higher content of hard endosperm starch granules are covered in a protein matrix, and the endosperm is more compact, which hinders water absorption.

Figure 2 Effect of soak time on the water absorption by the grain of five varieties of corn at room temperature. 

In summary, taking into account the variables that showed the best correlation with grain hardness (flotation index, test weight and density), grain varieties were arranged in the following sequence from lower to higher hardness: Criollo, Mont265, Mont360, Mont363, Mont41. The Mont360 and Mont363 varieties were statistically not different in the flotation index.

Nixtamal moisture

Moisture increased during nixtamalization since the action of OHradicals when reacting and degrading part of the grain pericarp facilitated the absorption of water by grain endosperm (^{Sefa-Dedeh et al., 2004}). Average moisture values during nixtamalization had a range of 29 % to 32 %; Mont360 and Mont41 showed the lowest values (28.68 % and 28.55 % respectively) and the Criollo the highest value (31.37 %).

The variable water absorption at room temperature (Table 3) was directly related to the average grain moisture from t₀ to t₄₅ (Table 4) (r=0.628). This suggests that nixtamalized corn moisture can be predicted before grain nixtamalization by using the water absorption test at room temperature. In the Criollo and Mont41 varieties the moisture of nixtamal stabilized after 35 min, whereas in the Mont265, Mont360 and Mont363, humidity increased even after 45 minutes (Figure 3).

Figure 3 Grain humidity of five varieties of corn at different times of nixtamalization. 

In all cases, the nixtamal of the varieties had low moisture (35 % average), compared to those reported by ^{Billeb and Bressani (2001)} 40 %-43 %, ^{Vázquez et al. (2012)} 43.3 %51.3 % and ^{Salinas et al. (2010)} 45.8 %-49.7 %. These differences can be explained by the different times, temperatures and stages of nixtamalization used by these authors to report moisture: ^{Billeb and Bressani (2001)} used 50 min, 96 °C and reported about this variable during the heating period; ^{Salinas et al. (2010)} and ^{Vázquez et al. (2012)} measured nixtamal moisture after soaking, when the grain continued to absorb water.

Nejayote pH

Nejayote pH (Figure 4) in relation to time showed the same trend during the first 5 min. All samples started with a high pH, but as time passed, this value decreased until the end of the process in an average of 0.2 units. The Mont360 and Mont41 varieties absorbed less water during nixtamalization, suggesting the trend of these varieties of absorbing fewer ions both OH^-1 and Ca⁺² (^{Fernández-Muñoz et al., 2004}), which could cause the cooking water to remain more alkaline. In contrast, the cooking water from the Criollo variety had a lower pH than the rest, suggesting it absorbed more OHand Ca⁺², leaving more H⁺ ions available in the solution. Additionally to water absorption by the grain, the decrease of alkalinity may also be attributed to the lixiviation (^{Ruiz-Gutiérrez et al., 2010}) of components of hemicellulose such as uronic and phenolic acids (^{González et al., 2005}) which cause a neutralization reaction of Ca(OH)2.

Figure 4 Cooking water pH (nejayote) of five varieties of corn; each corn variety during nixtamalization. 

Grain and nixtamal cutting force

In time (t₀), the Criollo and Mont265 corn grains had the highest cutting force, but Mont363 required less cutting force (Figure 5). This initial behavior of the grain cutting force was explained by ^{Shandera and Jackson (2002)}, who indicated that grain shape, surface and thickness have an important effect on grain texture testing (^{Blandino et al., 2010}).

Figure 5 Cutting force of corn grain (t=0) and nixtamal at different cooking times of five varieties of corn. 

The cutting force of nixtamal was inversely proportional to the time since the grain was softer due to water absorption (Figure 5), reducing from 235 N to 95.3 N on average in the first minutes. ^{Ibarra-Mendívil et al. (2008)} reported a similar trend in the force to penetrate the grain with a punch from 57.2 N (25 min) to 43.1 N (45 min); then the reduction of cutting force was slower. Table 4 shows a summary of the average behavior of moisture, pH and cutting force in the five varieties.

Rheological properties of untreated corn flour (t ₀ ) and nixtamal flours (t _{10 - 45} )

The average maximum viscosity showed significant differences between varieties (Table 5). Mont265 variety showed the highest value and the Criollo variety the lowest value for all the times evaluated (t₀, t₁₀, t₂₀, t₃₅ and t₄₅). The maximum viscosity decreased at time t₁₀ and kept increasing until time t₄₅. This behavior was similar for all varieties during the nixtamalization, except for Mont265 where the value of this variable increased at times t₁₀ and t₂₀ but decreased at t₃₅ and t₄₅. This behavior could be related to the grain size (small).

Table 5 Rheological and thermal characteristics of nixtamalized corn flours. 

Mean ± standard deviation. Means with different letters in a column are statistically different (Tukey, p≤0.05). ΔH = Enthalpy.

The viscosity of retrogradation presented a behavior similar to the maximum viscosity, that is, a decline to t₁₀ nixtamalization followed by an increase to t₄₅ time. The apparent decline between t₀ and t₁₀ may have been caused by the initial changes, such as calcium absorption at time t₁₀. The increase was most noticeable in the Criollo corn that coincided with the characteristics of a soft corn. The average retrogradation viscosities of Mont41 and Criollo varieties were similar, the same as those of Mont360 and Mont363. After cooking starch recrystallizes or anneals to form new structures, which is known as retrogradation (^{Paredes López et al., 2009}. By extending time, viscosity also increased due to the effect of temperature, since as starch granules are heated, they capture more water and swell (^{Corn Refiners Association, 2006}); ^{Rodriguez et al., 2008}; ^{Vaclavik and Christian, 2008}).

Thermal properties of corn flour untreated (t ₀ ) and nixtamal flours (t _{10 - 45} )

The highest average temperature of gelatinization was found in Mont41 and Mont360 varieties, whereas the lowest was in Criollo, and these values are consistent with those reported by ^{Ruiz-Gutiérrez et al. (2010)}. The temperature at which starch is gelatinized depends on the concentration of starch, suspension pH, heating and the specific procedure rate (^{Corn Refiners Association, 2006}). According to ^{Pineda-Gómez et al. (2011)}, the gelatinization temperature tends to increase with the longer cooking time due to the increase of calcium content during nixtamalization. This explains the increase in temperature of gelatinization found after 45 minutes in all varieties. Besides, these values are consistent with those reported by ^{Arámbula-Villa et al. (2001)}.

With respect to the enthalpy of gelatinization (ΔH), which is the energy required by grains of each variety for the process of gelatinization (^{Pineda-Gómez et al., 2010}), the variety that needed more energy was Mont363, followed by Criollo and Mont265; Mont41 and Mont360 had the lowest values (Table 5).

Enthalpy showed an opposite trend regarding the gelatinization temperature. Those varieties with higher energy requirements for nixtamalization had lower values of gelatinization temperatures. This trend was similar in all varieties except Criollo, which showed no significant change in enthalpy over time. This behavior can be related to the composition of the starch granules. The amylose-amylopectin proportion in starch of serrated type corn ranges from 25 % to 75 % (^{Salinas-Moreno et al., 2003}). According to ^{Rojas-Molina et al. (2007)}, the corn starch granules of hard endosperm have a higher proportion of amylopectin, whereas those of soft endosperm have a higher proportion of amylose, which affects their thermal characteristics.

Heat required compared to that provided by fuel during cooking

During nixtamalization there were differences in the variables related to fuel consumption (Table 6). Gas consumption, the heat provided by the latter and heat required for each variety to reach the removal of the pericarp depended on the duration of heating. The varieties evaluated required on average 1323 kJ (Qreq; Table 6) to reach cooking, and the heat provided by the fuel (approximately 121 g) was on average 5603.6 kJ (Qgas; Table 6). This represented an average loss of 4280.6 kJ (76.3 %) (Losses; Table 6) of the total heat supplied by the fuel. Although this is related to the system cooking configuration used in our study for nixtamalization (traditional nixtamalization), we found that a longer cooking time implies a higher fuel loss. The variety requiring a longer cooking time was Mont41, followed by Mont360, Mont265 and Mont363 which had a range of time from 26.67 to 31.67 min (Table 6). Criollo required less cooking time and had a difference of 5 min with Mont41 (the variety that required the longest time). The cooking time difference between the last two varieties resulted in a loss of 321.95 kJ. According to ^{Rodríguez et al. (2008)}, quality characteristics, physical, chemical and technological properties can influence the grain processing and are related to factors such as flotation index, test weight and water absorption. On this basis the behavior of the Mont41 variety is explained as it required a longer cooking time, compared to Criollo, which was softer.