Transcript
Study on formation of starch–lipid complexes duringextrusion-cooking of almond flour Teresa De Pilli a,* , Kirsi Jouppila b , Jarno Ikonen c , Jarno Kansikas d ,Antonio Derossi a , Carla Severini a a Department of Food Science, Faculty of Agriculture, University of Foggia, Via Napoli 25, 71100 Foggia, Italy b Department of Food Technology, P.O. Box 66 (Agnes Sjo¨ bergin katu 2), University of Helsinki, Helsinki FI-00014, Finland c Finnsoy Oy Talttatie 3, FI-23500 Uusikaupunki, Finland d Department of Chemistry, P.O. Box 55 (A.I.Virtasen aukio 1), University of Helsinki, Helsinki FI-00014, Finland Received 21 September 2007; received in revised form 21 December 2007; accepted 31 December 2007Available online 12 January 2008 Abstract A blend containing almond and wheat flours (about 1:2.5, dry basis) was extruded through a co-rotating twin-screw extruder with ascrew diameter of 25 mm. The effects of barrel temperature (61.72–118.28 C) and feed moisture (21.17–26.83%) on starch–lipid complexformation (X-ray diffraction, melting enthalpy, and complexing index) as well as on fat loss (which occurs during extrusion processing),break strength, deformability and expansion index of extrudates were investigated using response surface methodology (RSM).The only variable that had a significant influence on the formation of starch–lipid complexes was feed moisture (the high values of melting enthalpy and complexing index, i.e. 3.67 J/g and 94.34%, respectively, were obtained with the smallest value of feed moisture, i.e. 21.17%). The highest fat loss and the hardest texture of extrudates were obtained at the highest values of barrel temperature and feedmoisture (118.28 C and 26.83%, respectively). 2008 Published by Elsevier Ltd. Keywords: Starch–lipid complexes; Almond flour; Wheat flour; X-ray; DSC 1. Introduction The extrusion-cooking is principally used to producedirectly expanded products, as snacks, corn-flakes andpet food (Rokey, 1994), that have certain/special textureand flavour characteristics. Desirable properties in theextrudates are obtained by finding the appropriateprocessing condition as well as the proper composition of the raw material. Typical feedstocks include starch, pro-tein, lipids, water and additives. In general, the extrusionof starchy food results in gelatinization of starch, denatur-ation of protein and the formation of complexes betweenstarch and lipids and between proteins and lipids (Mercierand Feillet, 1975; Mercier et al., 1980; Ho and Izzo, 1992).The formation of complexes between starches and lipids isdue to the ability of the amylose to bind lipids such as fattyacids.Complex formation during twin-screw extrusion-cook-ing has been studied by Mercier et al. (1979, 1980), Colon-na and Mercier (1983), Schweizer et al. (1986), Gallowayet al. (1989), Guzman et al. (1992), Ho and Izzo (1992)and Strauss et al. (1992). Mercier et al. (1979, 1980)reported that twin-screw extrusion-cooking of manioc,potato and corn starches in the presence of native lipidsor added saturated and unsaturated fatty acids containing12–20 carbon atoms, glyceryl monostearate and sodiumstearoyl lactylate resulted in the formation of V-amylosecomplex.Amylose–lipid complexes can be prepared by the addi-tion of a complexing agent to a hot aqueous solution of starch. During cooling of the hot solution, the amylase 0260-8774/$ - see front matter 2008 Published by Elsevier Ltd.doi:10.1016/j.jfoodeng.2007.12.028 * Corresponding author. Tel.: +39 0 881 589245; fax: +39 0 881 589308. E-mail address:
[email protected] (T. De Pilli).www.elsevier.com/locate/jfoodeng Available online at www.sciencedirect.com Journal of Food Engineering 87 (2008) 495–504 constituent of the starch separates in a crystalline complex,which can be recovered readily. According to Kugimiyaet al. (1980), complex formation takes place immediatelyafter gelatinization in an exothermic process as observedin DSC experiments. Temperature studies on fattyacid–starch complexing have shown that an increase intemperature results in a decreased swelling power thathas been attributed to an increased complex formation(Gray and Schoch, 1962). Before gelatinization, starchhas limited binding capacity for lipid because most of thelipid in the system is unable to come into contact withthe starch. As the starch is dispersed, the amylose becomesavailable for complexation. However, for a process such asextrusion there is an optimum temperature for complexa-tion which depends on fatty acid chain length (Bhatnagar,1993) and above and below that temperature a decrease inlipid binding occurs (Bhatnagar and Hanna, 1994b). Otherstudies reported the effect of complex formation on physi-cal and chemical properties such as expansion ratio, bulkdensity, water solubility and iodine binding capacity of starches during extrusion-cooking (Mercier et al., 1979,1980; Colonna and Mercier, 1983; Schweizer et al., 1986;Galloway et al., 1989; Bhatnagar, 1993; Bhatnagar andHanna, 1994a,b).As a result, most studies on the starch–lipid complexformation during extrusion of model systems like starchand free fatty acids have been carried out, instead very littleis known about starch–lipid complex formation duringextrusion of flour blends containing fatty meal like almondflour.So, the aim of this work was to investigate the formationof starch–lipid complexes during extrusion-cooking of ablend containing almond and wheat flours at differentoperating conditions (barrel temperature and feed mois-ture) by X-ray Diffraction, DSC and Complexing Index.Wheat flour, which was considered as control material,was extruded using the same processing variables.Moreover, the effects of starch–lipid complexes on fatloss, expansion index, breaking strength and deformabilityof extrudates containing almond flour were determined. 2. Materials and methods 2.1. Flours Wheat flour was supplied by Pedone (Corato, Bari,Italy) and almond flour by Straziota (Ceglie del Campo,Bari, Italy). The chemical compositions of both flours arereported in Table 1. 2.2. Chemical analyses of flours The contents of moisture, ash, protein and fat weredetermined according to the AACC methods (2003). 2.3. Experimental design In Table 2, the factorial design of two variables (barreltemperature and feed moisture expressed as the percentageof dry basis) and five levels, obtained by Central CompositeDesign (CCD) (Box et al., 1978), is reported. This methodwas used to evaluate the single influence of the processingvariables as well as their possible interactions. Theapplication of the CCD allows the number of possible com-binationstobereduced toamanageablesizebecause itusedonly a fraction of the total number of factor combinationsfor experimentation. Eleven tests with differentcombinationsofprocessvariablevalues(barreltemperatureand feed moisture) were obtained using the following equa-tion n tot = n 0 + n c + n *= 4 + 3 + 4 = 11, where n 0 = 2 n , inwhich n is the number of variables, n c is the number of cen-tral points and n * is the number of star points (Box et al.,1978). The 11 combinations obtained for Central Com-posite Design and used to extrude wheat flour and a blendof wheat and almond flours are reported in Table 3.The appropriate range of processing parameters toextrude wheat flour and a blend with wheat and almondflours was chosen according to De Pilli et al. (2007). 2.4. Extrudates formula The doughs with wheat and almond flours were pre-pared according to the following formula: 75% wheat flourand 25% almond flour (De Pilli et al., 2000). Table 1Chemical composition of wheat and almond flourFlours Moisture (%) Ash (%) Protein (%) Fat (%)Wheat flour 11.61 ± 0.04 0.52 ± 0.01 9.57 ± 0.07 1.09 ± 0.07Almond flour 4.67 ± 0.03 3.32 ± 0.01 22.24 ± 0.02 58.7 ± 0.28Table 2Central composite designCoded values Barrel temperature ( C) Feed moisture (%) 1.41 61.72 21.17 1 70 220 90 24+1 110 26+1.41 118.28 26.83Table 3Experimental factorial designTests Barrel temperature ( C) Feed moisture (%)1 70 222 70 263 110 224 110 265 61.72 246 118.28 247 90 21.178 90 26.839 90 2410 90 2411 90 24496 T. De Pilli et al./Journal of Food Engineering 87 (2008) 495–504 2.5. Extruder A BC-21 CLEXTRAL (Firminy, France) co-rotatingtwin-screw extruder was used. The screw geometrical fea-tures were the following: diameter 25 mm and length900 mm ( L / D = 36:1) and distance between shafts 21 mm.The screw configuration used (from inlet to the die) isreported in Table 4.During extrusion experiments, the flour feed rate wasmaintainedconstantat3 kg/husingavolumetricgravityfee-der; while the moisture contentof dough wasadjustedusinga water pump. Water was pumped to the first zone of theextruder.Theextruderwasdividedintoninezones,indepen-dent of each other, for temperature control and adjustment.The first five zones were kept at room temperature,whereas the last four zones were adjusted to the tempera-tures reported in experimental factorial design (Table 3).The screw speed was kept constant at 200 rpm.The die used has a spherical shape (a diameter of 60 mm) in which there were two spherical holes with adiameter of 5 mm.According to the following equation (Bhattacharya andChoudhury, 1994), the specific mechanic energy (SME) wascalculated,SME ¼ rpm of screw ð run Þ rpm of screw ð rated Þ % torque ð run Þ 100 motor power ð rated Þ production capacityThis parameter did not show any change and its value was113.57 kJ/kg.At the exit of the die, the extrudates were cut into sticks(20 mm in length) through a cutter (210 rpm).For chemical and physical analyses, the extrudates weredried overnight at 40 C in a vacuum oven (Termaks, TS-8265, Norway) and were finely ground (particles < 300 l m)by mill BUHLER ML 1204 (Germany). All the groundsamples were defatted in a Soxtec fat-extractor with petro-leum ether at 37 C (b.p. 34.6 C) to remove uncomplexedlipids before chemical analyses.All analyses were performed at least in triplicate. 2.6. X-ray diffraction Complex formation between starch and lipids duringextrusion was studied using powder X-ray diffraction.Ground extrudates made of wheat flour or a blendcontaining wheat and almond flours were analyzed with aPhilips X-ray powder diffractometer (PW 1830 generator,PW 1710 diffractometer control, PW 1820 verticalgoniometer) equipped with graphite reflected-beam mono-chromator, and PC-APD software for Automatic PowderDiffraction Version 3.0 (Philips Analytical, Eindhovenand Almelo, The Netherlands). The X-ray source was CuK a radiation with a wavelength of 0.15418 nm and theX-ray diffractometer was operated in reflection mode at40 kV and 30 mA. Data were collected from 10 to 30 2 h (2 h being the angle of diffraction) with a step width of 0.02 and step time of 2.5 s. Value of 2 h for each identifi-able peak in the diffractograms was estimated, and crystal d -spacing was calculated using Bragg’s law. 2.7. Differential scanning calorimetry (DSC) The thermal properties of the starch–lipid complexes of extrudates were analyzed using differential scanning calo-rimetry (DSC). The thermograms were obtained with aPyris 6 Diamond DSC (Perkin–Elmer instruments, USA)equipped with cooling system Intracooler (230 V, 50/60 Hz) N5374099. The instrument was calibrated withindium using an empty pan as reference. The samplepreparationincludedweighing2–3 mgofgroundextrudatesin a calibrated DSC pan (inner volume 60 l l, large volumecapsule 0319-0218). This type of pan allows elimination of the interfering effects of the heat vaporization by suppress-ing the vaporization of water. Distilled water was addedwith a microsyringe to obtain solid content of 20% in themixture. DSC pans were hermetically sealed with a Univer-sal Crimper Press equipped with Die Kit (B050-5340). Thesamples were heated from room temperature to 135 C at arateof10 C/mininthepresenceofnitrogenwithapressureof 7 psi. A 25 l l quantity of distilled water was used as thereferencetocounterbalancethemassofwateronthesampleside.ThismakestheDSCmoresensitivetothepropertiesof the ‘solute’ or starch in the starch water sample slurry.Calculated values for onset temperature ( T i ), peak tem-perature ( T p ) and enthalpy ( D H ) were recorded for starch– lipid complex melting endotherms. 2.8. Complexing index Complexingindex(CI)wasdeterminedusingthemethoddescribed by Guraya et al. (1997). The iodine solution usedfor analysis was prepared by dissolving 2 g of potassiumiodideand1.3 gofI 2 in50 mlofdistilledwaterandallowingit to dissolve overnight. Then, the final volume was made to100 ml using distilled water. A 5 g sample (triplicate) wasmixed with 25 ml of distilled water in a test tube. The testtube was mixed for 2 min and centrifuged for 15 min at3000 rpm. The supernatant (500 l l) and distilled water(15 ml) were added to the iodine solution (2 ml). The tubewas turned over several times and absorbance was mea-sured at 690 nm through a UV–vis spectrophotometer Table 4Screw profile used in extrusion experimentsType of screw element Screw element details(pitch/length)Total length (mm)Forward pitch 50/50 15050/33 100Kneading block 50/25 20050/33 100Decreased pitch 50/25 10050/16 250 T. De Pilli et al./Journal of Food Engineering 87 (2008) 495–504 497 (Beckman DU 640, CA). CI was calculated using the fol-lowing equation:CI ð % Þ¼ ab c ab s ab c 100where ab c is the absorbance of control and ab s is the absor-bance of sample. Wheat flour based extrudates were con-sidered as control sample. 2.9. Fat loss % Percentages of fat loss were calculated as the differencebetween fat content in raw mixtures and fat content inextrudates, and expressed as g of fat loss/100 g of product. 2.10. Break strength and deformability A dynamometer stable Micro System TA-HDi TextureAnalyser (ENCO s.r.l., Venezia, Italy) with a small wedgewas used. Extrudates were placed over two supports,1.5 cm apart, and broken in the middle by a plunger thathad a shape of a cone frustum (the thickness of contact sur-face with extrudate was 1 mm 2 and the speed was constantand equal to 0.5 mm/s).Results were expressed as breaking strength (N/mm 2 ), i.e. the strength needed to break the extrudate, and defor-mability (mm), i.e. the measure of extrudate deformabilitybefore rupture, which was calculated as the ratio betweenthe area of force–deformation curve (N mm) and breakingstrength (N/mm 2 ).The first index is related to microstructure of samplesand it simulates the incisors impact at biting (Hutchinsonet al., 1987; Hayter and Smith, 1988; Georget et al.,1995; Li et al., 1998; Van Hecke et al., 1998). For each sam-ple, at least 10 repetitions were carried out. 2.11. Expansion index The expansion ratio of extrudates was determinedaccording to Kumagai et al. (1987) and it is defined asEI ð % Þ¼ D p D d 100where D p is the diameter of cylindrical product and D d isthe diameter of the die (5 mm).The mean diameter of extrudate cross-section was mea-sured using the OPTIMAS ver. 6.1 software (Bioscan,USA). The data were obtained from 10 extrudates ran-domly selected from each sample. 2.12. Statistical analysis Data were submitted to statistical analysis using Stat-soft, vers. 5.1 (Statsoft, Tulsa, USA) software. The analysiswas carried out in two steps. The first involved a stepwiseregression analysis to identify the relevant variables, andthe second used a multiple regression analysis (StandardLeast Square Fitting) to fit a second order mathematicalmodel, according to the following polynomial equation: y ¼ B 0 þ X B i v i þ X B ii v 2 ii þ X B ij v i v j where y is the dependent variable (melting enthalpy, com-plexing index, fat loss, break strength, deformability andexpansion index), B 0 is a constant value, v i and v j are theindependent variables (barrel temperature and feed mois-ture) in coded values and B i , B ii and B ij are the regressioncoefficients of the model. This model allowed the effectsof the linear ( v i ), quadratic ð v 2 i Þ and combined ( v i v j ) termsof the independent variables to be assessed on the depen-dent variable.Variables with a significance lower than 95% ( p > 0.05)were left out of the equation. To describe both individualand interactive effects of the independent variables of theextrusion-cooking process on D H of melting, complexingindex, fat loss, break strength, deformability and expansionindex, iso-response surfaces were developed. 3. Results and discussion The X-ray diffraction study was performed to obtainqualitative evidence of complex formation. The powderX-ray diffractograms of samples extruded with a blend of wheat and almond flours and wheat flour are shown inFig. 1a, 1b and 1c. For samples extruded with a blend of almond and wheat flours at barrel temperature and feedmoisture less than 90 C and 26%, respectively, peaks wereobserved at 2 h values of 15.18 , 17.13 , 18.03 and 22.86 (Fig. 1a and 1b). The powder X-ray diffractograms of extrudates with only wheat flour did not show a clear dif-ference between peak at 2 h values of 17.13 and 18.03 .Then, extrudates with almond flour, unlike those with onlywheat flour, show a pattern that closely matching A-typewheat starch (Zobel, 1964; Zobel et al., 1988), resultingfrom the formation of double helical structures of the amy-lopectin fraction (Jenkins et al., 1993). The preservation of A-type crystallinity in samples with almond flour could bedue to the lubrication of flour particles by the lipids (pres-ent in almond flour) that tend to protect the starch granulesfrom dispersal (Guy, 1994).Upon extrusion of blend wheat and almond flours atbarrel temperature more than 90 C, the X-ray diffractionpattern changed to V-hydrate form, which was character-ized by peaks at 13.0 and 20.0 2 h (Fig. 1c). The starchmelting that occurs at these operating conditions involvesthe destruction of the double amylopectin helices, whereaspart of the free lipids can form a helical inclusion complexwith the amylose molecules (Hoover and Hadziyev, 1981;Zobel, 1988). Diffraction patterns of samples extruded withonly wheat flour showed a better-defined V-pattern com-pared to the samples with blend containing wheat andalmond flour (Fig. 1c). Crystallization behaviours of amy-lose–lipid complexes are affected by crystallization condi-tion (such as moisture content and temperature). The 498 T. De Pilli et al./Journal of Food Engineering 87 (2008) 495–504 complexes can exist in various states of aggregation. Bili-aderis and Galloway (1989) showed two distinct forms of starch–lipid complexes: I (low T m ) and II (high T m )depending on the crystallization temperature.Form I, which was formed through rapid nucleationwith random distribution of the basic complex structuralelements, lacked a well-defined V-pattern although itstill had a high endothermic enthalpy in DSC analysis. Fig. 1a. X-ray diffraction patterns of wheat (B) and a blend of almond and wheat flour (A) extruded at 61.72 and 70 C and 22%, 24% and 26%.Fig. 1b. X-ray diffraction patterns of wheat (B) and a blend of almond and wheat flour (A) extruded at 90 C and 21.17%, 24% and 26.83%. T. De Pilli et al./Journal of Food Engineering 87 (2008) 495–504 499
