|
Perhaps an example of how other instrument companies may not offer the best help for food texture testing is the article "Using Instrumental Texture Analysis to Ensure Product Quality" written by Richard McManuis of Instron Corp in the Nov/Dec 2001 issue of Cereal Foods World. The article had numerous examples of how Instron Corp does not understand the cereal industry's existing test methods and how its materials science approach is misguided in foods. The consequences of the article may be food scientists conducting tests with poor results, and who may then wrongly conclude that texture analysis cannot help solve their problems. (2) McManuis suggests measuring bread to 25% deformation with a flat-ended 10 cm2 probe at a speed between .17 and .83 mm/second. He goes on to caution against 'inadvertently stacking' slices of bread on top of another, and his table down plays the necessity of a trigger force. The AACC (74-09) Standard Method for bread freshness (available in our software's help menu) calls for using two 12.5 mm thick slices of bread (or one 25 mm thick slice) with a 36mm diameter cylinder (10.2 cm2 ) probe with a radius edge (so as to not cut the bread) at a speed of 1.67 mm/second (more than twice as fast as McManuis' upper speed range). While the standard method calls for the force to be captured at 25% strain, the method actually calls for the probe to travel 40% strain. The original method was actually written for InstronTM machines, and the method had to overcome the difficulty they had in triggering at the surface of the product and the strip chart recorders which captured Instron Corp data then (1986). With modern texture analyzers the trigger force allows operators control over an important variable which cannot be controlled with Instrons, perhaps why he down plays its importance. Even mechanically sliced bread is not perfectly flat and the captured force at the 25% strain distance should be calculated from a repeatable determination of the true surface of the bread. Test results which were obtained following McManuis's recommended test method (primarily the wrong speed) would not be comparable with results which followed the actual AACC method. (3) McManuis then suggests that brittleness is the force at which a sample snaps in a three-point bend rig. By that logic a hard granola bar would be more brittle than a thin potato chip. His 20 years of material science logic tells him that brittle materials are also hard materials, and his article suggests inexperience with measuring brittleness as it applies to food products. Brittleness in foods can be measured by the distance or the modulus (slope) to where a product cracks or breaks. A good general rule is that if a product cracks at a smaller deformation it is more brittle. (4) With regard to snack products, McManuis is correct to state that they are often judged by their crunchiness and are irregular in shape and size. He then, however, presents a test which would provide meaningless results for most snack products. The test he describes involves compressing a certain number of pieces between two plates and measuring force at first peak (a), average force over a range of deformation (b), and force at final deformation (c). ____If the products are irregular in size, then the first peak (a) will simply be where the largest snack product happened to crack first. It would not be representative of the selected population. Even crispness (as opposed to crunchiness) would not be quantified with the force at first peak. ____The average force measure (b) might provide some interesting data, but it would not be able to differentiate between curves which had upward sloping graphs (products which compact then compress) or ones which had downward sloping graphs (hard brittle products which break early). ____With regard to final deformation force (c), if these were the crunchy snack products Instron used in the example then they should all have broken or cracked during the deformation and the final force would either be close to zero (in the case of a compressive deformation which breaks the product but is not sufficient to crush the residue), or very high (as the broken residue is even further crushed). In neither case would anything instructive be learned from final force (c) about crunchiness. The real crunchiness measures from the same test would be to quantify the degree of jaggedness of the graph. You could quantify (i) the number and size of the force peaks which occurred as the population of snack products was compressed, (ii) the average gradient of the force peaks, or (iii) even some form of fractal analysis of the plot line (perhaps Kolmogorov plots or even simply quantifying the linear length of the line). The initial slope to say 15% or 25% of the absolute peak force would be a good measure of crispness. The area of work, absolute peak force, or even McManuis' average force might tell you only how tough the material was, not anything about its crunchiness. (5) On the second page of the article McManuis states that for individual snacks like tortilla chips "The force at fracture is used as a measure of crispness." Once again crispness and brittleness are events which are best measured by gradients or distances. (6) We have over one hundred installations worldwide into pasta companies and the texture of cooked pasta is virtually never quantified using a tensile test as described by Instron Corp in the article. Instron Corp appears unawares that the AACC has a Standard Method for determining pasta firmness (Method 16-50 proposed 1983, approved 1989) which calls for shearing five strands of spaghetti, noodles, and other pasta shapes having uniform, solid cross section with a knife blade with a flat 1 mm end to a fixed 0.5 mm distance above the base plate at a speed of 0.17 mm/second. Firmness is defined in this method as the work in grams-centimeter required to shear one piece of pasta (e.g. one strand of spaghetti). For short goods the pasta industry typically uses Kramer Shear cells, Ottawa cells and various back extrusion or compression techniques. A tensile extension test is often used for noodle quality, but usually for second order effects such as to examine the gluten quality of different durum crops, or to evaluate the quality of the extrusion dies. Noodles are still most typically tested in accordance with the AACC standard method 16-50, or some slight variation of it. (7) McManuis also demonstrates a lack of understanding of the importance of dough stickiness when he describes the issue as regarding layered pie doughs leaking their filling. The efficiency of bakeries often relies on controlling stickiness which can be caused by overworking dough, too much water, flour extraction, amount of water-soluble pentosans, differences in protein composition, enzyme activity and even other genetic compositions of wheat. Once bakeries (or their vendors) know a dough's stickiness, then can then make the necesary operating adjustments. Dough stickiness is also extremely difficult to asses properly, since doughs need to be handled consistently, must be allowed to rest, and dough cohesiveness and extensibility must be separated from dough stickiness. Drs. Hoseney and Chen solved these problems with their patented Chen-Hoseney Dough Stickiness Fixture, which was licensed to SMS and Texture Technologies Corp. (8) In his table 'Guide to selecting the best instrument...' McManuis mentioned "modern systems do not require a trigger to complete the test". Well, he is wrong. Few food products have consistent heights, surfaces or geometries, and test results are often extremely sensitive to very small test distance changes. Without a trigger mechanism a test's initial starting position is a function of an operator's skill and judgment, which are not always uniform, leading to variable results. A user-configurable trigger mechanism creates an algorithm to always initiate the test based on an objective determination of the product's surface. This issue is absolutely critical for constently measuring the physical characteristics of products such as soft doughs, fillings, gels, creams, soft cheeses, etc. (9) In his table McManuis also mentioned "test analysis methods should not require using a macro language or programming". He is wrong here as well. Today's graphs can be either very simple or extremely complicated. Macros can be as simple as a few memorized keystrokes, or can execute complex user-defined actions. Twenty-first century users demand flexible easy-to-use texture analysis software which can simply perform complex operations. Perhaps Instron Corp does not have macro capability, but our customers conduct tens, hundreds and thousands of tests and they wish to automate as much of the analysis and data presentation as possible. Macros allow the consistent analysis of thousands of tests with a single mouse click. And easy-to-use macros are what the people on the lab bench scream out for. It is no wonder then that Instron Corp's clients regularly call us for food texture analysis advice.  |