A Standardized Method for the Instrumental Determination of Cooked Spaghetti Firmness
By Sissons, M J Schlichting, L M; Egan, N; Aarts, W A; Harden, S; Marchylo, B A
ABSTRACT A standardized method to determine cooked spaghetti firmness was developed. The effects of process and instrument variables were investigated and optimized to provide reproducible results between laboratories and to enable discrimination among samples with similar firmness characteristics. Commercial spaghetti samples of varying thickness were chosen to artificially create a range in firmness, and used to investigate the effect of a wide range of variables on cooked spaghetti firmness including sample preparation, cooking procedure, postcooking treatment, sample presentation, and instrument settings. Cooked spaghetti firmness determined using a TA-XT2i texture analyzer was significantly affected by optimum cook time, postcook cooling, rest time, and crosshead speed (P
Cereal Chem. 85(3):440-444
Pasta is a popular food because of its sensory appeal, versatility, low cost, ease of preparation, excellent dried storage stability, and strong nutritional image. Sensory appeal is determined by appearance, texture, and flavor (Bourne 2002). Of these three, textural properties of pasta have received more research effort because of its importance to consumer acceptance (Cole 1991; D’Egidio and Nardi 1996). Sensory analysis using highly trained panelists is considered the ultimate tool for measurement of the cooking quality of pasta products (Matsuo 1988). An internationally recognized standard sensory method for pasta, however, is not adhered to, most likely because of the influence of local preferences on sensory parameters. In addition, sensory evaluation is time consuming and impractical when sample size is limited. In response to these constraints, various instruments have been developed to evaluate the texture of pasta and have been reviewed extensively (Cole 1991). One of the more popular and commonly used methods is measuring the degree of compression of cooked spaghetti strands using universal testing machines like the Instron (Approved Method 66-50, AACC International 2000), Lloyd, LFRA, tensipresser, and TA.XT2i’ texture analyzers (Smewing 1997; Bourne 2002). Manufacturers have provided guidance on the application of these instruments, in particular for the TA.XT2/ texture analyzer, to various foods but to our knowledge, a detailed scientific evaluation of the effects of variations in cooking and instrument procedures has not been published. Oh et al (1983) found that noodle firmness was affected by many factors including diameter, degree of overcooking, and postcook rest time before compression and cutting. Although Approved Method 66-50 (AACC International 2000), which is based on the procedure of Oh et al (1983), lists values for some of these parameters, values for some instrument settings and other test parameters are not clearly indicated and it is not apparent whether all parameters had been optimized to provide reproducible firmness results that would facilitate discrimination of textural differences among similar samples. Preliminary comparisons of firmness results obtained by three different laboratories using an interpretation of the protocol described in AACC Approved Method 66-50 showed that the somewhat different cooking and instrument settings used by the laboratories appeared to influence sample rankings (Sissons et al 2004). It was concluded that a standardized method optimized to provide reproducible firmness results was needed to enable discrimination of textural differences among similar samples. Such a method will facilitate quality control programs within the durum trade and pasta manufacturing industry and support international interlaboratory research studies.
Accordingly, the objectives of this study were to 1 ) investigate the effect of process and instrument variables on the firmness test for cooked spaghetti, and 2) optimize these variables and devise a detailed standard operating procedure that provides reproducible firmness results both within and among laboratories and thus enable discrimination of textural differences among similar samples.
MATERIALS AND METHODS
Three commercial spaghetti samples with the same lot number were purchased from supermarkets in large quantities to provide sufficient material for method the optimization experiments. Three different strand diameters were chosen as a simple means of providing samples with a range of cooked spaghetti firmness values. Dry spaghetti diameter was determined by measuring four randomly chosen strands at three places along the strand with digital calipers (model 62379-531, VWR Canlab, Mississauga, ON). A TA.XT2f texture analyzer with accompanying Texture Expert software (Texture Technologies, Scarsdale NY) was used to determine cooked spaghetti diameter by subtracting the distance the blade traveled into the spaghetti minus the set distance (Wood et al 2001). The average dry and cooked spaghetti diameters are shown in Table I.
The optimized method was validated by cooking and testing 29 spaghetti samples with the same nominal diameter, commercially processed by different manufacturers, and offered for sale under different brand names in supermarkets in Australia, Canada, and the United States, which included imported brands from Italy.
Mean Spaghetti Strand Diameter for Method Standardization Samples (n = 12)
Optimum Cooking Time (OCT)
For each sample, 8 g of spaghetti (7-cm lengths) was cooked in 250 mL of boiling water that was filtered by reverse osmosis. The level of boiling water was maintained at [asymptotically =]90% of the original volume and samples were stirred throughout the cooking process. Optimum cooking time, defined as the time required for the center core of the spaghetti strand to disappear plus 15 sec, was determined by squeezing at least three strands between clear plastic plates (13 x 6 cm hinged Perspex plates) at 15-sec intervals. OCT for all samples was determined before firmness testing.
Cooked Spaghetti Firmness
A TA.XT2( texture analyzer equipped with a 5-kg load cell and the TA-47 pasta blade (Texture Technologies) was used to determine cooked spaghetti firmness. Spaghetti strands were undercooked, overcooked, or cooked to optimum (OCT) according to predetermined times, followed by cooling in 250 mL of distilled water. After cooling for a preset time, strands were placed adjoining or adjacent but spaced 1 cm apart on the TA.XT2; base plate and cut crosswise with the blade. Three subsequent cuts of new strands from the same cook followed at 45-sec intervals. This procedure was performed in triplicate giving a total of 12 measurements per sample. Spaghetti weight, strand length, cooling water temperature, rest time, instrument settings, and number of strands cut were varied to determine optimum conditions for measuring firmness. Spaghetti firmness was defined as the height of the peak (in grams), which is analogous to maximum cutting force (Smewing 1997).
In the series of independent method optimization experiments, the objective was to determine the effect of various treatments on the reproducibility and ability to discriminate between samples based on firmness. Firmness was analyzed with linear mixed models using the software package ASReml (Gilmour et al 2002). The fixed terms included in the model were spaghetti samples, treatments, and all interactions. Treatments were included in the models as covariates, except the adjoining versus adjacent but spaced 1 cm apart by treatment factor. Random terms to account for replicate effects and the repeated testing of samples in groups of four were also included in the models. Because the method optimization set of spaghetti samples was chosen to give a wide range in diameter and hence a wide range in firmness values, significance of differences between the spaghetti samples was not of interest and, in general, will not be commented upon. To determine whether some levels of a treatment resulted in a higher variation in firmness, models similar to those above were used, except treatments were included as factors rather than covariates. These models assume equal variances for each level of the treatment factor; however models allowing the variances to differ can also be fitted. The change in log-likelihood (ALL) that occurs by allowing the variances to differ can be compared to a chi^sub 1^^sup 2^ distribution, so that if (2) (ALL) >3.84, the model is improved by having differing variances.
For method validation, the firmness values for 29 samples were analyzed by two laboratories. To compare the agreement between the laboratories, firmness values from one laboratory were regressed against values from the other laboratory and the variance accounted for by the model was determined. The Spearman rank correlation was also calculated. These results were then compared with the same statistics calculated for data initially obtained using nonstandardized methods. RESULTS AND DISCUSSION
Method Standardization and Optimization
To develop and optimize a standard spaghetti firmness method, a series of independent experiments were performed to investigate the effect of a variety of factors on firmness values, and to maximize the ability to discriminate between samples. The relationships between all treatments and firmness were linear (Fig. 1), except crosshead speed, which exhibited a logarithmic relationship with firmness (Fig. 2). Optimum cooking time (OCT) for samples S, M, and L was determined in duplicate as described above, and variations in the procedure were introduced. The first set of experiments focused on variations in the cooking procedure including length of spaghetti strands, ratio of spaghetti to water, and cooking over and under the OCT. Dry spaghetti strand lengths of 2, 5, and 7 cm were chosen based on those reported in the literature (Matsuo and Irvine 1971; Voisey et al 1978; Wyland and D’Appolonia 1982). Strand length had no effect on firmness although there was a sample by strand length interaction (Fig. IA, Table II) indicating that the effect varied with the sample. The interaction term was only just significant at P
Fig. 1. Effect of treatments on cooked spaghetti firmness. A, Strand length; B, dry spaghetti-to-water ratio; C, under-OCT and over-OCT; D, postcook cooling time; E, postcook rest time; F, cooling water temperature; G, difference in firmness/five strands for 10 samples; H, adjoining or adjacent and spaced 1 cm apart. Sample S (A); Sample M (O); Sample L (V).
A wide variety of water to pasta ratios have been reported in the literature ranging from 5:1 (Walsh 1971), 10:1 (Matsuo and Irvine 1969; Voisey et al 1978; D’Egidio et al 1982, 1990), and 31:1 (Grzybowski and Donnelly 1979) to 50:1 (Oh et al 1983). Most researchers have used 25-100 g of dried spaghetti for instrumental testing, however this is impractical in a plant breeding program because quantity of material is limited. Three spaghetti to water ratios were evaluated using 4, 8, and 16 g of dried spaghetti, while keeping a constant water volume of 250 mL. Although the effect of the water to spaghetti ratio is statistically significant, the 1-g decrease in firmness for each 1-g increase in sample weight (Table II, Fig. IB) is insignificant in practical terms. Sufficient water, however, is required to allow strands to move freely during cooking and to rapidly return to boil after addition of strands. Variances for the three ratios were similar, but an 8-g sample size was chosen to account for potentially limited sample availability.
Although spaghetti may be cooked to a standard time, cooking to optimum is the preferred procedure for research purposes (Voisey and Larmond 1973; D’Egidio and Nardi 1996). To determine the effect of cooking under and over the predetermined OCT on firmness, samples were cooked to optimum +- 30 and 60 sec. Both undercooking and overcooking had a significant effect on firmness (Table II) and, as expected, all samples were more firm when undercooked and less firm when overcooked (Fig. 1C), but the magnitude of the change was dependant on the sample. When cooked past OCT, the firmness of sample S (smallest diameter) decreased by 12 +- 3 g/min, while sample L (largest diameter) exhibited a decrease of 40 +- 3 g/min. While there was no difference in firmness at OCT +- 30 sec for samples S and M there was a significant difference for sample L. These results suggest that not only is it important to determine OCT accurately, but that strand diameter also must be specified.
Fig. 2. Effect of crosshead speed on cooked spaghetti firmness. Sample S (Delta); Sample M (*); Sample L ([backward difference]).
Effect of Various Treatments on Cooked Spaghetti Firmness
Several researchers have noted the importance of time lapse between cooking and testing on spaghetti texture (Voisey and Larmond 1973; Dexter et al 1983) and resting times of up to 30 min have been used before testing as noted by Oh et al (1983). For these reasons, the effects of postcooking treatment of samples, including cooling and resting times, and temperature of water used in the cooling step were investigated. Upon reaching OCT, spaghetti must be cooled to arrest cooking. Cooling time had a significant effect on firmness (Table II, Fig. 1D), with a 6 +- l g decrease in firmness for each additional l min of cooling time. Resting time also had a significant effect on firmness but the effect varied by sample (Table II). The largest decrease in firmness with increased resting time was observed for sample L (largest strand diameter), and the smallest decrease was observed for sample S (smallest strand diameter) (Fig. IE). For each l min increase in resting time, estimated firmness decreased by 1,6, and 6 g (+- 2.8) for samples S, M, and L, respectively. A cooling time of 2 min and rest time of 1 min were chosen for the standard procedure to minimize the total time to complete sample testing.
To evaluate the effect of cooling water temperature, samples were cooled in water held at 18, 25, and 30[degrees]C following cooking to OCT. Cooling water temperature did not affect peak firmness (Table II). However, lower variability was associated with a cooling temperature of 25[degrees]C, so this temperature was chosen for standardization purposes (Fig. 1F).
A third set of experiments was conducted to investigate variations associated with sample presentation which included the number of cooked strands to cut and cooked strand position. The number of strands cut has been reported in the literature to range from one (Walsh 1971) to 10 strands (Canadian Grain Commission, Quality of Western Canadian Wheat 1999). Ten commercial spaghetti samples, including samples S, M, and L with dry strand diameters of 1.43-2.15 mm were used to evaluate the effect of the number of strands cut on firmness. Samples were cooked to optimum, and either 5 or 10 strands were cut. As expected, firmness values were approximately twice as high when cutting 10 strands due to the greater force required to cut through more strands. To make comparison easier, the firmness values when testing 10 strands were divided by two so that firmness is expressed as firmness per 5 strands. Firmness for the 10 samples when testing 5 strands was slightly but significantly higher (5 +- 2 g) compared with testing 10 strands (Table II; Fig. IG). Differences for each of the 10 samples are shown in Fig. IG and ranged from 15 g higher to 7 g lower. However, this variation between samples was not significant (Table II). The use of 5 strands was chosen for standardization because for most samples there were an insufficient number of strands to obtain four successive cuts with 10 strands.
Fig. 3. Effect of compression depth (CD) on cooked spaghetti firmness.
The position of the strands in relation to each other when aligned on the base plate might be expected to influence the force versus time curve because of the interaction of forces when strands are touching. Placement of five optimally cooked strands either adjoining or adjacent but spaced 1 cm apart had a significant effect on firmness results (Table II). Firmness values were higher by 16 +- 5 g when strands were adjacent (Fig. 1H). Presumably, when strands are touching, the interplay of forces between strands decreases during compression and reduces the peak force required to cut through the strands. The highly significant effect indicates that strand position must be specified.
Instrument settings such as crosshead speed are known to influence determination of cooked spaghetti firmness (Voisey and Larmond 1973). Therefore, two important instrument settings, crosshead speed and compression depth, were investigated to assess their effect on the firmness measurement and ability to discriminate among samples. Crosshead speed was varied over 0.2-9.5 mm/sec with the resulting changes in firmness shown in Fig. 2. As expected, firmness increased with increasing crosshead speed (Table II). However, the magnitude of the change was highest at lower speeds. The effect of increasing crosshead speed also varied with sample. Sample S (smallest diameter) showed the least increase in firmness and sample L (largest diameter) showed the largest increase. The greatest difference in firmness values among samples was achieved at the highest crosshead speed. However, a somewhat slower speed of 2.0 mm/sec was selected because it was more suitable for sequencing of multiple sample analyses as part of a routine testing regime. It also was the point where changes in the force with increasing test speed began to plateau and thus provided only marginally decreased differences in firmness among samples.
The effect of compression depth on cooked spaghetti firmness was evaluated as a percentage of cooked strand diameters. During shearing, the maximum firmness would be expected at the core of the cooked strand, i.e., 50% compression depth. Therefore, compression depths of 75, 90, and 95% were chosen to ensure that the blade had passed through the center of the cooked strand. To achieve the same compression depth, the distance the blade traveled into each sample differed greatly because of the large range in cooked diameters of the spaghetti samples. No differences in firmness values or variances were observed for each of the three samples tested over compression depths of 75-95% (Fig. 3). A fixed distance of 4.5 mm (crosshead height calibrated to 5 mm) was chosen for simplicity and to ensure the blade would pass through the core of a wide range of strand diameters. TABLE III
Instrument and Sample Preparation Conditions for Standardized Operating Procedures
In summary, examination of the variables associated with sample preparation (precooking and postcooking), presentation, and instrument settings to obtain firmness values with good repeatability (CV
Validation of Standard Method
Work performed in a preliminary study (Sissons et al 2004) found that when three laboratories used similar but not identical methods to measure firmness, a good correlation was obtained (r = 0.85) among laboratories. For comparison, the same 29 samples used in that study were reevaluated in two of those laboratories, using the standardized method and optimized instrument settings (Table III) developed in this study. Linear regression and Spearman rank correlation were used to assess agreement between laboratories before and after use of the standard method. Regression of Laboratory One data against Laboratory Two data gave a correlation of r = 0.96. The residual standard error fell from 64.7 to 26.5 or 45.4 to 27.3 for regression of Laboratory Two data against Laboratory One data. These results are shown in Fig. 4A and B, prestandardization and poststandardization, respectively, by the decrease in scatter about the regression lines after standardization. Before standardization, the firmness values for Laboratory One were on average [asymptotically =]386 g higher than values obtained by Laboratory Two, making direct comparisons between laboratories more problematic. After standardization, Laboratory One values were on average only 12 g higher than those of Laboratory Two. The effect of improvement in agreement between the two laboratories when using the standardized method was illustrated clearly when ranking samples because the Spearman rank correlation increased from r^sub s^ = 0.81 to rs – 0.95.
Fig. 4. Comparison of cooked spaghetti firmness determined in two laboratories before (A) and after (B) standardization. Dashed lines show equal firmness values for both laboratories.
Optimizing process and instrument parameters (TA.XT2i texture analyzer) to determine cooked spaghetti firmness by using Approved Method 66-50 (AACC International 2000) results in significantly improved repeatability and facilitates improved discrimination of textural differences among similar samples. Incorporation of these parameters into a detailed standard operating procedure also will minimize variation between laboratories and consequently will improve interlaboratory comparison of cooked spaghetti firmness for both research and commercial applications.
M. S. thanks the Grains Research and Development Corporation for financial assistance with travel to Canada which allowed the collaboration to begin.
AACC International. 2000. Approved Methods of the American Association of Cereal Chemists, 10th Ed. Method 66-50. The Association: St. Paul, MN.
Bourne, M. C. 2002. Sensory methods of texture and viscosity measurement. Pages 257-292 in: Food Texture and Viscosity. M. C. Bourne, ed. Academic Press: New York.
Canadian Grain Commission. 1999. Quality of western Canadian wheat. Grain Research Laboratory, Canadian Grain Comm.: Winnipeg, MB.
Cole, M. E. 1991. Review: Prediction and measurement of pasta quality. Int. J. Food Sci. Technol. 26:133-151.
D’Egidio, M. G., and Nardi, S. 1996. Textural measurement of cooked spaghetti. Pages 133-156 in: Pasta and Noodle Technology. J. E. Kruger, R. B. Matsuo, and J. W. Dick, eds. AACC International: St. Paul, MN.
D’Egidio, M. G., De Stefanie. E., Fortini, S., Galterio, G., Nardi, S., Sgrulletta, D., and Bozzini, A. 1982. Standardization of cooking quality analysis in macaroni and pasta products. Cereal Foods World 27:367-368.
Dexter, J. E., Kilborn, R. H., Morgan, B. C., and Matsuo, R. R. 1983. Grain Research Laboratory compression tester: Instrumental measurement of cooked spaghetti stickiness. Cereal Chem. 60:139- 142.
Gilmour, A. R., Gogel, B. J., Cullis, B. R., Welham, S. J., and Thompson, R. 2002. ASReml user guide release 1.0. VSN International Ltd.: Hemel Hempstead, UK.
Grzybowski, R. A., and Donnelly, B. J. 1979. Cooking properties of spaghetti: Factors affecting cooking quality. J. Agric. Food Chem. 27:380-384.
Matsuo, R. R. 1988. Evaluation of durum wheat, semolina and pasta in Canada. Pages 249-262 in: Durum Wheat: Chemistry and Technology. G. Fabriani and C. Lintas, eds. AACC International: St. Paul, MN.
Matsuo, R. R., and Irvine, G. N. 1969. Spaghetti tenderness apparatus. Cereal Chem. 46:1-6.
Matsuo, R. R., and Irvine, G. N. 1971. Note on an improved apparatus for testing spaghetti tenderness. Cereal Chem. 48:554- 558.
Oh, N. H., Seib, P. A., Deyoe, C. W., and Ward, A. B. 1983. Noodles. I. Measuring the textural characteristics of cooked noodles. Cereal Chem. 60:433-438.
Sissons, M. J., Schlichting, L., Egan, N., Ames, N., Rhymer, C., Marchylo, B., and Batey, I. L. 2004. A comparison of instrumental techniques used to discriminate the cooking quality of spaghetti. CD- ROM, session 12, 12th ICC Cereal and Bread Congress, Harrogate UK.
Smewing, J. 1997. Analyzing the texture of pasta for quality control. Cereal Foods World 42:8-12.
Voisey, P. W., and Larmond, E. 1973. Exploratory evaluation of instrumental techniques for measuring some textural characteristics of cooked spaghetti. Cereal Sci. Today 18:126-143.
Voisey, P. W., Larmond, E., and Wasik, R. J. 1978. Measuring the texture of cooked spaghetti. 1. Sensory and instrumental evaluation of firmness. J. Inst. Can. Sci. Technol. Aliment. 11:142-148.
Walsh, D. E. 1971. Measuring spaghetti firmness. Cereal Sei. Today 16:202-205.
Wood, J. A., Batey, I. L., Hare, R. A., and Sissons, M. J. 2001. A comparison of Australian and imported spaghetti. Food Australia 53:349-354.
Wyland, A. R., and D’Appolonia, B. L. 1982. Influence of drying temperature and farina blending on spaghetti quality. Cereal Chem. 59:199-201.
[Received September 4, 2007. Accepted December 5, 2007.]
M. J. Sissons,1,2 L. M. Schlichting,3 N. Egan,1 W. A. Aarts,3 S. Harden,1 and B. A. Marchylo3
1 NSW Department of Primary Industries, Tamworth Agricultural Institute 4 Marsden Park Road, Tamworth, NSW 2340, Australia.
2 Corresponding author. E-mail: email@example.com
3 Grain Research Laboratory, Canadian Grain Commission, 1404-303 Main Street, Winnipeg, Manitoba R3C 3G8 Canada.
This article is in the public domain and not copyrightable. It may be freely reprinted with customary crediting of the source. AACC International, Inc., 2008.
Copyright American Association of Cereal Chemists May/Jun 2008
(c) 2008 Cereal Chemistry. Provided by ProQuest Information and Learning. All rights Reserved.