WO2000062058A1 - Methods and apparatus for estimating the oxidative stability of an oxidizable material - Google Patents

Abstract

The method comprises determining the oxygen consumption by a sample of the material at a first temperature, determining the oxygen consumption by a sample of the material at a second temperature, and estimating the oxidative stability of the material at a third temperature by extrapolating the temperature-dependent data obtained at the first temperature and the second temperature to the third temperature.

Description

METHODS AND APPARATUS FOR ESTIMATING THE OXIDATIVE STABILITY OF AN OXIDIZABLE MATERIAL

BACKGROUND OF THE INVENTION

The present invention relates generally to methods and apparatus for estimating the oxidative stability of an oxidizable material. More particularly, the present invention is directed to methods and apparatus for estimating the oxidative stability of an oxidizable material comprising a lipid, particularly a fat or an oil. The methods and apparatus of the present invention, for example, can be used to accurately determine oxidative stability using a unique approach that measures the rate of oxygen consumption at various temperatures and accurately extrapolates the measured data to ambient storage conditions.

The stability of various materials to oxidation is a difficult parameter to measure. The methods employed by commercially available equipment to measure oxidative stability range from a simple measurement of the consumption of oxygen as determined by a pressure drop in a sealed container to the measurement of induction periods until the oxidation reaction exceeds a predetermined rate. Many of these tests are conducted at high temperatures to minimize the time required for completion of the test. The data obtained from such tests is difficult, at best, to extrapolate to typical storage temperatures such as those temperatures commonly employed in the storage of high fat ingredients for the manufacture of animal feedstuffs. Oxidative stability measurements conducted at more meaningful, ambient temperatures are too slow to provide the necessary information in a useful time frame. Accuracy and reproducibility are also common problems with the existing methods. Furthermore, many of these methods require the extraction and testing of a fat when solid samples are to be analyzed. For solid samples, the oxidative stability is determined for the isolated fat. This approach introduces numerous opportunities for error and can positively or negatively affect the actual stability of the fat in the original sample.

When left exposed to an oxygen-containing atmosphere, an oxidizable material, particularly a material comprising a lipid such as an oil or fat, undergoes oxidative degradation. For a material comprising a lipid this degradation mainly involves the oxidation of the unsaturated fatty acids (or unsaturated fatty acid derivatives) present in oils and fats. The oxidation of saturated fat is usually significantly slower.

The mechanism of lipid oxidation is generally known. Because the reactions of unsaturated fatty acids and oxygen to produce peroxides require relatively high activation energy, the direct attack of oxygen on unsaturated fatty acids generally does not occur. Instead, lipid auto-oxidation is generally believed to involve a free radical chain mechanism: (1) initiation steps that lead to free radicals (R-);

(2) propagation of the free radicals (R- + O₂ - ROO-, ROO- + RH → ROOH + R-); and

(3) termination steps (R- + R- - R-R, R- + ROO- - ROOR, ROO- + ROO- - O₂ + ROOR (or alcohol and carbonyl compound)).

The oxidation of lipids thus results in peroxides as the primary oxidation products. Typically, an induction period is observed before the onset of significant production of peroxides. During this induction period, however, oxygen is consumed in a zero order process, apparently leading to poorly characterized initial intermediates prior to the formation of the peroxides. The peroxides in turn degrade further to products such as aldehydes, ketones and carboxylic acids. This degradation leads to rancidity, loss of nutritional value and increase in toxicity, which in turn contributes to diseases, and accelerates the aging process.

The degree of unsaturation, the presence of indigenous or added antioxidants and prooxidants, and thermal and illumination conditions of storage all can affect lipid oxidative stability. The same type of material comprising a lipid may even exhibit different stabilities from one sample to the next. The types and degrees of unsaturation present in porcine lard have been shown, for instance, to be diet-dependent. The stability also depends on the exact source of the lipid. For example, the stability of piscine lipids from the skin is different from that of lipids from other tissues.

Evaluation of the oxidative stability of lipids is an old and complex topic. A generally applicable, fully satisfactory method of evaluation has yet to emerge. For oil and fat samples, it is important to know both the current oxidative status and the relative potential to undergo oxidative degradation. A single measurement of the peroxide content (the peroxide value or "PN") is useful as an index of current oxidation status only if the peroxides formed are sufficiently stable so that they do not decompose after formation. Very easily oxidized lipids typically are not sufficiently stable to permit the effective measurement of peroxide content.

While the activation energy for the peroxide formation from unsaturated acids is about 146 to 272 KJ/mol, the activation energy for the decomposition of several lipid peroxides has been reported to be in the range of about 84 to 184.5 KJ/mol, suggesting that the peroxides are less stable than the lipids themselves. In addition, PV and the onset of rancidity are not always well correlated. It is possible for a rancid sample to have a low PV. Ironically, oxidative stability assessment based on a single P V measurement is often used to estimate the shelf life of a product (i.e., the time before the onset of rancidity allowing for an adequate safety margin).

While PV may provide an indication of the current oxidative state of a sample, it provides no information on the relative potential of the sample to oxidize. There are, however, two general approaches to determining potential oxidative stability. The first is to evaluate the induction period observed before significant production of peroxides (or secondary products) begins. The induction period is very dependent on the conditions of the oxidation experiment and the oxidation experiment therefore must be carried out according to a set protocol. A commonly used method is the Active Oxygen Method ("AOM"), promulgated by the American Oil Chemists Society (American Oil Chemist 's Society Official

Method Cd-12-57, 1993, American Oil Chemists Society, Champaign, Illinois). This method has been largely unchanged over the past five decades. Several 20 mL aliquots of fat samples are taken and aerated at 2.33 mL/s at 98 °C and periodically analyzed for PV by an iodometric procedure. The time to reach a PV of 100 meq/Kg is taken to be an index of the oxidative stability. The actual endpoint in industrial practice can be as little as 20 meq/Kg, depending on the type of fat. This method requires a large amount of sample, numerous analyses, and critical control of the air flow rate. A typical AOM determination requires 20 to 40 hours to perform. For samples that form relatively unstable peroxides, a PV of 100 meq/Kg may never be reached and an AOM measurement will have little meaning. The induction period also can be monitored by measuring the consumption of oxygen.

In the Sylvester test, a sample is heated to 100°C in a closed vessel and a single measurement of the pressure decrease, due to the consumption of oxygen, is taken. A higher oxygen concentration and a longer induction period are employed than in the method of the instant application. See, Wewala, A. R, Natural Antioxidants. Chemistry, Health Effects and Applications. Shahidi, F. Ed. AOCS Press, Champaign, IL 1997, pp 331-345. An automated embodiment of this test is the Oxidograph in which the sample is heated in a reaction vessel under 100% oxygen and the induction period is determined from a sudden pressure decrease.

Another official method used to measure the induction period is the Oil Stability Index ("OSI") method (American Oil Chemist's Society Official Method Cd-12b-92, 1993, American Oil Chemists Society, Champaign, Illinois). A stream of purified air is passed through a hot lipid sample, and the effluent air is bubbled through deionized water. The conductivity of the water is continuously monitored. As the final oxidation products, volatile carboxylic acids (mostly formic acid) are formed in the oil or fat sample. These acids are flushed out by the flowing air and are collected in the water, resulting in an increase in the conductivity of the water. The time of the onset of the major conductivity rise of the water is obtained by tangential extrapolation. The OSI values are generally well correlated to the corresponding AOM values if the PV is 100 or greater. The method is automated and thus easier to use than the AOM. It is, however, time-consuming and suffers from the common problem that large errors can result from small variations in the air flow rate. An additional concern is that carboxylic acids are further down the degradation chain and are only one of the possible products of hydroperoxide decomposition. The OSI value is thus based on an event occurring well after the process of interest and may not bear a linear correspondence with that process. Moreover, several volatile acids other than formic acid are known to form in different ratios in different oil/fat samples. Finally, with formic acid itself significant losses from the collector solution may occur if collection temperatures exceed 20°C. The other principal option for assessing potential oxidative stability is to monitor the rate of peroxide formation or oxygen consumption. Versions of the AOM that attempt to measure the oxidative stability over a shorter period involve a single measurement of the PV after either 4 or 20 hours of aeration. In one version of an oxygen consumption method, the sample is sealed under air or oxygen and stored at a constant temperature. The oxygen concentration in the headspace is monitored by periodically withdrawing small samples through multiple sealed septa affixed to the container and analyzing those samples by gas chromatography.

The fundamental problem with all of these methods is that lipid stability is determined at a fixed temperature that is usually far above the ambient because it otherwise takes unacceptably long otherwise to make meaningful measurements. The result must be extrapolated to determine the comparative stabilities of two different types of samples under ambient storage conditions. This extrapolation makes the tenuous assumption that the activation energies for the lipid oxidation of the two samples are the same. It has been shown that the rate of lipid oxidation can be dependent on whether the sample is ground or whole and whether the lipid is isolated from the sample and then examined by itself. The assumption that two altogether different samples have identical activation energies is difficult to justify.

The determination of the oxidative stability of solid samples containing fat and oil, which can range from milk powder to potato chips to animal feed, presents an even greater challenge. There is no standard method for the direct determination of lipid oxidative stability in solid samples. Methods presently used involve multiple steps. Typically, the total amount of fat/oil that is present in a solid state sample is extracted first. A mixed organic solvent (petroleum etheπmethanol = 1 : 1, v/v) may be used to triple-extract the fat existing in solid samples. The extract is then washed with deionized water. The washed extract is dried in a modestly warm water bath (typically 40°C). Finally, the oxidative stability of the extracted fat is determined by a standard method such as the AOM or the OSI method. The extraction step itself requires about 5 hours and a typical AOM or OSI method determination can require 20 to 40 hours.

This approach is flawed in numerous ways. It is difficult to select an extraction solvent having proper polarity. If the solvent is not sufficiently polar, the peroxides already formed will not be extracted. If the solvent is too polar, the peroxides will be preferentially extracted over the fat and the analytical result will be biased. The sample also must be extracted at room temperature to avoid heating and oxidizing the fat during extraction or decomposing the peroxides already present. This requirement eliminates the use of a Soxhlet extractor. In addition, there is no guarantee that three extractions are sufficient to fully extract both the fat and the peroxides. Once extraction is complete, however, the solvent has to be evaporated with a rotary evaporator at low temperature, again to avoid altering the oxidative stability of the fat. Finally, the solvent has to be fully evaporated to obtain an accurate weight of the fat.

There is also the potential that extracting fat from a solid sample may not provide an accurate picture of the actual stability of the original sample at all. The lipid may be preferentially located within the interior of the sample matrix and thus be less sensitive to oxidation than extraction and subsequent determination may suggest. In addition, a basic problem with the AOM and the OSI method is that lipid stability is determined at a fixed temperature (98 °C). Although this is far above ambient storage temperature, this temperature is used because it takes unacceptably long to make the necessary measurements at ambient temperatures. The extrapolation of such data in making decisions about comparative stability of two different types of samples under ambient storage conditions makes the tenuous assumption that the activation energies for the lipid oxidation in the two samples are the same. It has been shown that the rate of lipid oxidation can be dependent on whether the sample is ground or whole and whether the lipid is isolated from the sample and then examined by itself. The assumption of identical activation energy between two altogether different samples is an over-simplification due to the lack of better analytical methods. SUMMARY OF THE INVENTION

Novel methods and apparatus for the direct determination of the oxidative stability of an oxidizable material, particularly a material comprising a lipid, have been developed.

In one embodiment of the invention, the novel method comprises determining the oxygen consumption by a sample of the oxidizable material at a first temperature; determining the oxygen consumption by a sample of the oxidizable material at a second temperature; and estimating the oxidative stability of the oxidizable material at a third temperature by extrapolating the temperature-dependent data obtained at the first temperature and the second temperature to the third temperature. In another embodiment of the invention, the novel apparatus comprises a reactor comprising a reaction chamber having an inlet and an outlet. The reaction chamber is adapted to receive and hold a sample of an oxidizable material. The apparatus further comprises a means for controlling the temperature of the reaction chamber, a means for sensing the oxygen content of a diluent gas comprising oxygen, a means for controlling the flow of the diluent gas, and a valve means for directing the flow of the diluent gas comprising oxygen. The valve means has a plurality of operating positions. In a first position of the valve means, the inlet of the reaction chamber is in fluid communication with a source of said diluent gas and the outlet of the reaction chamber is in fluid communication with the oxygen sensing means. In a second position of the valve means, the diluent gas passes from the source of diluent gas to the oxygen sensing means without first passing through the reaction chamber and the inlet and outlet of the reaction chamber are substantially sealed.

Other features of the invention will be in part apparent and in part pointed out hereinafter.

BRIEF DESCRIPTION OF THE DRAWINGS

Fig. 1 schematically shows an analysis system for the determination of oxygen consumption by a liquid material comprising a lipid in accordance with one embodiment of the present invention.

Fig. 2(a) schematically shows a gas phase injection analysis system for the determination of oxygen consumption by a liquid material comprising a lipid in accordance with the present invention.

Fig. 2(b) schematically shows an oxidation reactor employed in the gas phase injection system of Fig. 2(a).

Fig. 3 shows a plot of the oxygen sensor signal output as a function of oxidation reactor temperature for a cottonseed oil sample in the gas phase injection analysis system of

Fig. 2(a). Fig. 3 reports triplicate measurements at each of the following oxidation reactor temperatures: (1) 70 °C, (2) 80 °C, (3) 90 °C, (4) 100 °C, (5) 108 °C and (6) 120 °C.

Fig. 4 shows a plot of the oxygen sensor signal output as a function of sample volume for a safflower oil sample in the gas phase injection analysis system of Fig. 2(a).

Fig. 5 shows a plot of the oxygen sensor signal output (both peak height and peak area) as a function of (diluent gas flow rate)^"1 for a safflower oil sample in the gas phase injection analysis system of Fig. 2(a).

Fig. 6 shows a plot of the oxygen sensor signal output as a function of reactor residence time in the gas phase injection analysis system of Fig. 2(a) for the following samples: (1) safflower oil, (2) olive oil, (3) cottonseed oil, (4) com oil, and (5) fish oil. Fig. 7 shows an Arrhenius plot of the log(oxygen sensor signal output) as a function of 1000 x (reactor temperature)^"1 in the gas phase injection analysis system of Fig. 2(a) for the following samples: (1) vegetable oil, (2) safflower oil, (3) cottonseed oil, (4) com oil, (5) fish oil, (6) lard, (7) peanut oil, (8) poultry fat, and (9) olive oil.

Fig. 8 shows an Arrhenius plot of the log(oxygen sensor signal output) as a function of 1000 x (reactor temperature)^"1 in the gas phase injection analysis system of Fig. 2(a) for com oil and fish oil samples over a broader reactor temperature range than reported in Fig. 7.

Fig. 9 shows a plot of iodine absorption number (" IAN") decrease ( left ordinate, closed symbols) and peroxide value ("PV") increase (right ordinate, open symbols) against the calculated relative oxygen consumption in the gas phase injection analysis system of Fig. 2(a) for the following samples: (1) vegetable oil, (2) safflower oil, (3) cottonseed oil, (4) peanut oil, and (5) olive oil.

Fig. 10(a) schematically shows a gas phase injection analysis system for the determination of oxygen consumption by a solid material comprising a lipid in accordance with one embodiment of the present invention. Fig. 10(b) schematically shows an oxidation reactor employed in the gas phase injection system of Fig. 10(a).

Fig. 11 shows an Arrhenius plot of the log(oxygen sensor signal output) as a function of 1000 x (reactor temperature)^"1 in the gas phase injection analysis system of Fig. 10(a) for eight bone meal samples. Fig. 12 shows an Arrhenius plot of the log(oxygen sensor signal output) as a function of 1000 x (reactor temperature)^"1 in the gas phase injection analysis system of Fig. 10(a) for three bone meal samples. Fig. 12 reports two separate measurements for each sample.

Fig. 13 shows a plot of the oxygen sensor signal output as a function of sample weight for a bone meal sample in the gas phase injection analysis system of Fig. 10(a). Fig. 14 shows a plot of the relative oxygen consumption as a function of lipid loading in the gas phase injection analysis system of Fig. 10(a) for four silica gel matrix samples containing cottonseed oil in an amount by weight of: (1) 3.4%, (2) 6.3%, (3) 9.1%, and (4) 12%), respectively.

Fig. 15 shows an Arrhenius plot of the log(oxygen sensor signal output) as a function of 1000 x (reactor temperature)^"1 in the gas phase injection analysis system of Fig. 10(a) for the following four bone meal samples: (1) an original composite mixture of bone meal, (2) a sieved fraction derived from the original mixture and having an average particle size between 0.250 to 0.600 mm, (3) a sieved fraction derived from the original mixture and having an average particle size between 0.600 to 0.850 mm, and (4) a sieved fraction derived from the original mixture and having an average particle size greater than 0.850. Fig. 16 shows a plot of the oxygen sensor signal output as a function of (geometrical mean particle diameter)^"1 for four lipid-coated glass bead samples having a particle diameter of (1) 0.12 to 0.16 mm, (2) 0.25 to 0.30 mm, (3) 0.45 to 0.50 mm, and (4) 1.0 to 1.05 mm, respectively.

Fig. 17 shows a plot of PV value determined by the abbreviated AOM method (4 hours at 98 °C) against the oxygen consumption predicted in accordance with a method of the present invention.

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

In accordance with the present invention, novel methods for estimating the oxidative stability of an oxidizable material, particularly an oxidizable material comprising a lipid, have been devised. The novel methods comprise determining the oxygen consumption by a sample of the oxidizable material at a first temperature; determining the oxygen consumption by a sample of the oxidizable material at a second temperature; and estimating the oxidative stability of the oxidizable material at a third temperature by extrapolating the temperature- dependent data obtained at the first temperature and the second temperature to the third temperature.

The novel methods can be used to predict the oxidative stability of oxidizable materials, such as a material comprising a lipid, at temperatures, such as ambient storage temperatures, where direct experimental measurement of oxidative stability at those temperatures conventionally is difficult and/or time-consuming. The oxidation of lipids necessarily involves the consumption of oxygen. Methods such as AOM use the formation of peroxides, an intermediate product of the oxidation, as an indirect quantitative index of oxidative stability of a material comprising a lipid. In comparison, the novel methods described in this application directly use the extent of oxygen consumption by the material, a more direct and reasonable parameter than formation of peroxides, for the determination of oxidative stability. The novel methods possess one of more of the following advantages over conventional methods employed to determine oxidative stability. The methods are significantly faster than conventional methods. For example, a typical multi-temperature oxidative stability analysis using the gas phase flow injection analysis systems described in the examples of this application requires about 2.5 to 3 hours thereby increasing the throughput of test samples relative to conventional methods. The novel methods provide highly reproducible results. The apparatus needed to practice the novel method are relatively simple instrumentally and require no chemicals other than a relatively small flow of an inexpensive diluent gas. The apparatus easily can be automated to further increase the throughput of test samples.

In addition, the novel methods can be used to measure the oxidative stability of a material comprising a lipid that is a solid at the testing temperature without the need to extract and recover the fat from the sample for further processing as is required by conventional methods used to measure oxidative stability. Solid sample amounts required by the present methods are an order to two orders of magnitude less than the sample requirement of conventional procedures using extraction procedures.

Further, oxygen consumption as determined in accordance with the present methods is well correlated to the concomitant decrease in iodine absorption number (IAN). The temperature dependence of the oxidizable materials tested exhibited Arrhenius behavior, i.e., the log(oxygen consumption) varied linearly with reciprocal absolute temperature, allowing for the extrapolation of the temperature dependent oxygen consumption value to any storage temperature. Because the relative order of oxidative stability for different oxidizable materials can be temperature dependent, the determination of oxidative stability at a fixed elevated temperature as carried out in the AOM or OSI method may not provide meaningful results.

One illustrative method in accordance with the present invention is now described in detail with reference to analytical system 1 shown in Fig. 1. It will be understood by those skilled in the art that the present invention may be used in conjunction with oxidation reactors of various configurations and modes of operation, the following description is intended to be merely illustrative of the type of system in which the present invention may be applied. Analytical system 1 comprises an adjustable temperature reactor 2 in fluid communication with oxygen sensing means 3. Reactor 2 has an internal volume, said reactor further having an inlet port 4 through which a gas may be supplied to the internal volume of reactor 2 and an outlet port 5 from which a gas may be removed from the internal volume of reactor 2 and directed to oxygen sensing means 3. A sample of an oxidizable material having a defined sample volume when placed in reactor 2 has a defined surface area in contact with a gas supplied to the internal volume of reactor 2. The reactor head space is the internal volume of the reactor less the sample volume.

Temperature controller means 6 allows the temperature of reactor 2 to be adjusted to a desired temperature. During the operation of analytical system 1, temperature controller means 6 is operated so as to maintain reactor 2 at a first temperature, T,. The flow of a diluent gas comprising oxygen into reactor 2 through inlet port 4 is initiated and a comparable volume of gas is allowed to exit reactor 2 through outlet port 5. After a time suitable for the temperature of the sample to reach T_j and to purge the internal volume of reactor 1 such that the gas exiting the reactor has substantially the same concentration of oxygen as the diluent gas entering the reactor, the flow of diluent gas into and out of reactor

2 is discontinued and inlet port 4 and outlet port 5 are closed to seal reactor 2 from the external environment. The oxygen present in the gas occupying the headspace of reactor 2 is then allowed to react with the sample for a predetermined period of time, the residence time. As the oxygen present in the gas charged to reactor 2 reacts with the sample, the concentration of free oxygen in the gas decreases. At the end of the predetermined residence time, inlet port 4 and outlet port 5 are opened and the flow of the diluent gas comprising oxygen through reactor port 4 into reactor 5 is reinitiated. The gas exiting reactor 2 through outlet port 5 is then directed to oxygen sensing means 3, which includes but is not limited to sensors such as electrochemical oxygen sensors, which measures the oxygen concentration of this gas.

Temperature controller means 6 is then adjusted so as to maintain reactor 2 at a second temperature, T₂, and the method described above is repeated to measure the amount of oxygen consumed by the sample at the second temperature for the same residence time. This measurement step likewise can be repeated at additional temperatures to obtain more temperature dependent data (i.e., oxygen consumption measurements). While temperature dependent data at two different temperatures is generally satisfactory for estimating oxidative stability, preferably such data is obtained for three or more different temperatures.

The temperature dependent data obtained are then extrapolated to a third temperature, T₃, to estimate the oxygen consumption, and thus the oxidative stability, of the sample at T₃. Preferably, the temperature dependent data measured above are employed to determine a linear equation for estimating the oxidative stability of the sample over a selected temperature range. In one embodiment of the invention, the extrapolation is accomplished using an Arrhenius-type plot wherein log(oxygen consumption) is plotted as a function of the reciprocal of absolute temperature for the temperature dependent data measured. The Arrhenius-type plot yields a linear equation in the form of log(oxygen consumed) = -a/T + b wherein T is absolute temperature and a and b are constant parameters calculated on the basis of the temperature dependent data. Once constants a and b have been determined (for example, using conventional linear regression techniques), the oxygen consumption by a sample at a specified temperature can be estimated to provide an indication of oxidative stability of the sample at that temperature.

As used herein, the term "oxidative stability" means the tendency of a material to resist reaction with oxygen. The greater the oxidative stability, the less likely the material is to react with oxygen. The sample tested is preferably an oxidizable material comprising a lipid. The material, for example, can be a fat or an oil, and can be in either liquid or solid form at the measured temperature. The definitions of the terms "ϋpid", "fat" and "oil" are generally known to those of ordinary skill in the art and should be interpreted broadly. The term "hpid", for example, generally refers to the group of biomolecules characterized by their insolubility in water and their solubility in fat solvents such as alcohol, ether and chloroform. Non- limiting examples of lipids includes fats, lipoids (such as phospho lipids, cerebrosides and waxes), and sterols (such as cholesterol and ergosterol). The term "fat" generally refers, for example, to esters of fatty acids and glycerol. Non-limiting examples of fats include peanut oil, vegetable oil, com oil, olive oil, cottonseed oil, safflower oil, coconut oil, palm oil, fish oil, tallow, poultry fat, bacon grease, butter and lard. The term "oil" generally refers to those fats that are liquids at 20°C.

The amount of the oxidizable material charged to the reactor may vary depending upon parameters including, but not limited to, the size of the reactor employed, the configuration of the reactor, and whether the sample is a solid or liquid at the test temperature. One of the advantages of the present method is that, if a low concentration of oxygen is used in the test, a single sample of the oxidizable material can be used to obtain all of the temperature dependent data measurements needed. Only a small amount of the sample is oxidized during the course of each oxygen consumption measurement and additional measurements at other temperatures can be taken using the same sample until the testing is completed. Preferably, therefore, the amount of the oxidizable material initially charged to the reactor is sufficient to allow for each of the desired oxygen consumption measurements to be taken. Typically, the amount of oxidizable material present in the reactor for each determination step is between about 0.001 g to about 100 g, preferably between about 0J g to about 20 g, more preferably between about 0J g to about 15 g, and still more preferably between about 0J g to about 10 g. As noted above, the gas introduced to reactor 2 through inlet port 4 is a gas comprising oxygen. This gas preferably comprises a diluent gas and oxygen wherein the diluent gas is substantially nonreactive with the oxidizable material tested. The diluent gas preferably is selected from the group consisting of nitrogen, carbon dioxide, helium, neon, argon, krypton and xenon.

A relatively low concentration of oxygen in the gas introduced to reactor 2 is preferred. The concentration of oxygen in the gas preferably is low enough that significant oxidation of the sample does not occur during the various measurement steps. Although suitable results can still be obtained at higher oxidation levels, the best results are obtained when less than 10% by weight, preferably less than 5% by weight, and more preferably less than 3% by weight of the sample is oxidized during the overall testing of the sample. Typically, the gas supplied to reactor 2 has an oxygen concentration less than about

20% by volume and preferably less than about 5% by volume. More specifically, the gas has an oxygen concentration between about 0.00002% by volume to about 20% by volume, more preferably between about 0.0002% by volume to about 5% by volume, still more preferably between about 0.0001% by volume to about 2% by volume, still more preferably between about 0.0001% to about 1% by volume, still more preferably between about

0.0001% to about 0.5% by volume, and still more preferably between about 0.01% to about 0.3% by volume.

While the gas flow rate is not narrowly critical, it will affect the mean sweeping time and gas dispersion in the reactor. The mean sweeping time is the reactor volume divided by the gas flow rate. The gas dispersion refers to the efficiency with which the reactor is swept. In a poorly-designed reactor, the original contained volume is washed out slowly and poorly leading to a broadening of the peak for oxygen consumption. Consequently, this leads to decreased sensitivity.

The desired gas flow rate will depend, in part, upon the internal volume of the reactor. Typically, the ratio of the gas flow rate to the internal volume of the reactor is between about 0.02 mL/cm³ -minute to about 4 mL/cm³ -minute, preferably between about 0.2 mL/cm³ -minute to about 3 mL/cm³ -minute, more preferably between about 0.4 mL/cm³-minute to about 2 mL/cm³ -minute, and still more preferably between about 0.6 mL/cm³ -minute to about 1 mL/cm³-minute. In one embodiment of the invention, for example, the gas flow rate employed in a reactor having an internal volume of about 30 cm³ is between about 1 mL/minute to about 150 mL/minute; preferably, between about 10 mL/minute to about 120 mL/minute, and more preferably between about 20 mL/minute to about 40 mL/minute.

If the residence time of the sample in reactor 2 (i.e., the period during which reactor 2 is sealed to the external environment and the diluent gas comprising oxygen is allowed to react with the sample) is too long, depletion of the oxygen in reactor 2 and depletion of the oxidizable material on the surface of the sample could measurably affect the observed rate of oxygen consumption. Accordingly, relatively short residence times are preferred. Because reaction temperature also affects the extent of oxidation, longer residence times may be employed for lower reaction temperatures. Typically, the residence time is between about 10 seconds to about 60 minutes, preferably between about 30 seconds to about 40 minutes, more preferably between about 1 minute to about 30 minutes, still more preferably between about 2 minutes to about 20 minutes, and still more preferably between about 3 minutes and about 10 minutes.

For each measurement step, the temperature of reactor 2 can be selected from a wide range of temperatures provided that the low end of the temperature range still provides a sufficiently high rate of oxidation so that accurate measurements of the temperature dependent data needed can be obtained within a reasonable time (i.e., within the specified residence time), and the high end of the temperature range still provides a sufficiently low rate of oxidation so that the sample is not extensively oxidized during the determination steps. This temperature range can be determined by one of ordinary skill in the art and will depend upon the oxidizable material selected for testing and the conditions under which that material is normally used. For example, the third temperature selected may be a relatively low temperature where the oxidizable material is generally maintained at an ambient storage temperature, or may be a much greater temperature where the oxidizable material is used as a frying medium.

A broad range of reactor sizes can be employed. Because large samples are not required for accurate measurements, relatively small reactor sizes can be used. Smaller reactor sizes likely will be desirable to minimize the amount of sample needed for testing, to allow for laboratory scale testing, and to minimize the capital expense associated with larger reactors. Preferably, therefore, the reactor will have an internal volume between about 10 cm³ to about 100 cm³, more preferably between about 10 cm³ to about 80 cm³, and still more preferably between about 10 cm³ to about 60 cm³.

The reactor configuration likewise is not narrowly critical although the reactor employed preferably possesses a configuration conducive to maintaining the relatively low level of oxidation of the sample during each oxygen consumption measurement step. A preferred reactor configuration is a surface reactor wherein only the exposed surface of the sample is contacted with the oxygen present in the internal volume of the reactor. More preferably, the surface reactor is compatible with, and integral to, a gas flow injection analysis system such as the system described in Example 1 below. For samples that are liquids at the measured temperature, the exposed surface area of the sample conesponds to the cross-sectional area of the oxidation reactor defined by a plane through the reactor that is coplanar with the surface of the liquid sample. In one embodiment of the invention, for example, the configuration of a reactor having a head space of 30 cm³ preferably provides a liquid sample surface area that is about 3.3 cm² to about 10 cm², preferably about 3.75 cm² to about 7.5 cm², and more preferably about 5 cm². For samples that are solids at the measured temperature, the exposed surface is different than for a liquid. If the sample depth becomes too large or if the sample is loaded in a tight and compacted fashion, the accessibility of the gas to the sample surface will vary from run to run and will affect the results. This situation is exacerbated when the particle size is small and the sample is prone to being packed tightly. When the sample amount is small, the sample occupies a relatively thin layer in the reactor and the access of the gas to the sample surface is facile. Accordingly, limited access of the oxygen to the entirety of the sample generally is not a significant problem under such conditions. In general, the oxygen consumption by a solid material is related linearly to the amount of sample used and inversely related to the particle diameter of the material. By way of comparison, the amount of a liquid sample loaded into a reactor having a constant cross-sectional area (and thus a constant sample surface area) and adequate headspace after the sample is loaded does not affect, within experimental error, the rate of oxygen consumption because the available surface for interaction is only the top surface of the sample. A typical layer depth for a reactor having an internal volume of about 30 cm³ is less than about 10 mm and preferably less than about 8 mm.

The amount of a solid sample required for testing may be lower than the amount of a liquid sample required for testing. Solid sample amounts in the range of about 1 g, for example, often provide suitable results. Typical solid samples are between about 0J gram to about 5 grams in weight, preferably between about 0J gram to about 3 grams in weight. In accordance with the method of the present invention, the amount of time required to estimate the oxidative stability of the material is typically less than four hours from the commencement of the method, preferably within about three hours from the commencement of the method, and more preferably within about two hours from the commencement of the method. The present invention is further directed to an apparatus for estimating the oxidative stability of an oxidizable material. The apparatus generally comprises a reactor comprising a reaction chamber having an inlet and an outlet. The reaction chamber is adapted to receive and hold a sample of an oxidizable material. The apparatus further comprises a means for controlling the temperature of the reaction chamber, a means for sensing the oxygen content of a diluent gas comprising oxygen, a means for controlling the flow of the diluent gas, and a valve means for directing the flow of the diluent gas comprising oxygen. The valve means has a plurality of operating positions. In a first position of the valve means, the inlet of the reaction chamber is in fluid communication with a source of said diluent gas and the outlet of the reaction chamber is in fluid communication with the oxygen sensing means. In a second position of the valve means, the diluent gas passes from the source of diluent gas to the oxygen sensing means without first passing through the reaction chamber and the inlet and outlet of the reaction chamber are substantially sealed. In one embodiment, the apparatus is manually operated. In another embodiment, the apparatus is automated.

The present invention is illustrated by the following examples which are merely for the purpose of illustration and are not to be regarded as limiting the scope of the invention or the manner in which it is practiced.

Example 1 : Determination of Oxidative Stability of Liquid Materials Comprising a Lipid

The oxidative stability of samples of peanut oil, vegetable oil, com oil, olive oil, cottonseed oil, safflower oil, lard, poultry fat and fish oil was evaluated as set forth below. Peanut oil, vegetable oil, com oil, olive oil, cottonseed oil, safflower oil and lard are commercially available materials. The samples used in this Example were obtained from local supermarkets. The poultry fat and fish oil samples were supplied by Novus International (St. Louis, MO).

Unless otherwise stated, a diluent gas containing OJ % O₂ and 99.9% N₂ (TRIGAS Industrial Gases, Lubbock, Texas) was used. Where air was used instead, the air was first purified by passing it through sequential columns packed with activated carbon, silica gel, and soda- lime. Diluent gas having any other concentration of oxygen was generated by in-line mixing of purified air with cylinder nitrogen. Mass flow controllers (model FC-280, Tylan General, Los Angeles, CA) were used for all flow measurements. All other chemicals and reagents used were reagent grade or better and used without further purification. Peroxide values ("PV") were determined according to American Oil Chemists Society Official Method Cd-8-53, American Oil Chemists Society, Champaign, Illinois. The iodine absorption number ("IAN") for all samples was determined according to American Oil Chemists Society Official Method Cd-1-25, American Oil Chemists Society, Champaign, Illinois

Experimental Arrangement

The gas phase flow injection analysis system shown in Fig. 2(a) was used to determine the oxidative stability of the samples. In this gas phase flow injection analysis system, diluent gas flowed through a mass flow controller ("MFC") and through an electropneumatically operated six-port stainless steel HPLC-type valve V (model 7000, Rheodyne, Cotati, CA).

Alternatively, two three-way valves or other comparable arrangements could be used. Further detail of the oxidation reactor ("OR") incorporated in the loop of valve V is shown in Fig. 2(b). The oxidation reactor was composed of a glass tube (150x16 mm outside diameter; 144x12 mm inside diameter) modified with air inlet, outlet, and a sample injection tube (each 45x4 mm outside diameter). Liquid oil/fat samples were inserted into the reactor through a poly(tetrafluoroethylene) ("PTFE") tube connected to the injection port, equipped with a female Luer fitting at the other end. The air inlet and outlet tubes were cemented to 1.5 mm outside diameter, 1 mm inside diameter nickel tubing with high temperature epoxy (RP 4026, Ciba-Geigy). A heating tape (128x50 mm, 50 W at 110 V, Thermolyne, IA) was secured around the oxidation reactor with glass wool thread. A platinum resistance thermometric detector ("RTD") was placed in-between the heating tape and the glass exterior of the oxidation reactor. The reactor temperature was programmed with a temperature controller (CN8500, Omega Engineering, Stamford, CT).

In the gas phase flow injection analysis system of Fig. 2(a), a cold trap was inserted between the valve V and the oxygen sensor (model 2550, Systech Instruments, IL). In most cases, the trap consisted of a 600 mL beaker filled with ice- water in which a coil of nickel tubing (1.6 mm outside diameter, 1.0 mm inside diameter, 400 mm long, coiled into a 32 mm diameter coil) was immersed, followed by a "buffer" glass tube (8.6 mm inside diameter, 10.2 mm outside diameter, 180 mm long). The cold trap allowed any oil vapor in the air stream to condense. Condensation of the oil in the sensor itself can greatly reduce useful sensor lifetime. The sensor output was recorded by a personal computer (Gateway 2000, P-75) equipped with a DAS- 1601 data acquisition board (Keithley-Metrabyte, Taunton, MA). All tubing used in the gas flow system was 1 mm inside diameter nickel tubes. A homebuilt digital timer was used to control the switching of valve V and hence the residence time of the oxygen bearing diluent gas in the oxidation reactor. Except as otherwise provided below, in performing each stability determination 12 mL of the sample was loaded into the reactor, the reactor was flushed with the diluent gas for two to three minutes, heated to the intended temperature and then switched to the isolated loop position. The loop was switched back in-line to measure the amount of oxygen consumed after a desired interval of time, as programmed by the valve timer. After a desired number of replicates (typically two to three measurements at a given temperature), the temperature was increased to the next desired value and the method was repeated. The entire method was repeated for as many temperatures as desired.

System Output A typical readout for the present system is shown in Fig. 3 for a cottonseed oil sample tested at six different temperatures. The signal value in mV represents the output of the oxygen sensor for each measurement. The oxygen sensor produces a signal that is linearly related to the amount of oxygen consumed. Negative peaks 1, 2, 3, 4, 5 and 6 correspond to the oxygen sensor signal output for tests conducted at oxidation reactor temperatures 70 °C, 80°C, 90°C, 100°C, 108°C and 120°C, respectively. Three measurements were taken at each temperature and are reported in Fig. 3. The residence time of the oil sample in the oxidation reactor for each measurement was five minutes.

As the sample was heated to higher temperatures, greater amounts of oxygen were consumed and the magnitude of the negative peaks increased. The reproducibility of measurements at each temperature was very good. Because oxygen intmsion from the outside environment through the polymeric connecting conduits and fittings of the apparatus can affect the results if the reactor residence time is long enough, silicone oil (which is not oxidized under the present experimental temperatures) was used as a reference sample and oxygen consumption computed relative to this control where necessary.

The data reported in Fig. 3 for the triplicate determinations at six different temperatures (a total of 18 measurements) was obtained using the same sample that was loaded only once into the oxidation reactor. Each sample was analyzed, the reactor was flushed with new diluent gas for two to three minutes and the sample re-analyzed, for a total of three replicates. The temperature was then increased to the next desired value and the method was repeated. Because so little of the sample was actually oxidized during each individual measurement, it was not necessary to introduce fresh sample aliquots for subsequent measurements.

Dependence on Adjustable Parameters

1. Sample Volume

The oxidation reactor used in this Example is basically a surface reactor where only the exposed surface of the sample is available for oxidation. The exposed surface area of the sample corresponds to the cross-sectional area of the oxidation reactor defined by a plane through the reactor that is coplanar with the surface of the liquid sample. As the oxidation reactor of the present Example is filled with a liquid sample, the exposed surface area of the sample becomes nearly constant for the reactor, and hence the oxygen consumption signal should remain essentially constant. This was confirmed by measuring oxygen consumption for different volumes of a safflower oil sample loaded into the oxidation reactor. In each case, the reactor temperature was maintained at 100 °C, the diluent gas flow was maintained at 40 mL/min, and the sample residence time was maintained at 9 minutes. The signal response reached a maximum value and remained essentially constant when the sample volume exceeded about 11 mL. The results are plotted in Fig. 4.

There are several advantages to using a surface reactor arrangement instead of a bubbled reactor or stirred reactor. These advantages include the ability to reuse the sample due to the small extent of oxidation and the relative independence from sample size. Further, the availability of oxygen during the course of the reaction does not become so limited that it affects the observed oxidation rate. In addition, a stirred reactor requires some means of agitation and exact reproduction of the reaction conditions to obtain meaningful results.

2. Diluent Gas Flow Rate

The effect of the diluent gas flow rate on the signal peak height and peak area was evaluated using a 10 mL sample of safflower oil with the reactor temperature maintained at 100 °C and the sample residence time maintained at 9 minutes. The experimental results for diluent gas flow rates from 10 mL/min to 130 mL/min are shown in Fig. 5.

A first effect noted is the increased purging time of the reactor caused by decreasing the diluent gas flow rate. This increase in purging time, however, is relatively small compared to the fixed reaction time of 9 minutes. A second effect noted is the greater peak width that results from increasing the diluent gas flow rate. While the peak areas (in units of mV-s) increase linearly with the reciprocal of the flow rate and have a negligible y- intercept (in units of V-mL and calculated by multiplying the peak area by the gas flow rate), which is directly proportional to the actual mass of oxygen consumed, the area counts remain essentially constant. The net effect is that a lower diluent gas flow rate results in higher peak heights but the peaks span a greater width in time. To achieve a compromise between analysis time and sensitivity, 12 mL samples and a 30 mL/min diluent gas flow rate were used for the remaining steps of this Example except as otherwise provided.

3. Determination of Oxygen Consumed at Different Residence Time

The effect of residence time on oxygen consumption was evaluated using 12 mL samples of (1) safflower oil, (2) olive oil, (3) cottonseed oil, (4) com oil, and (5) fish oil. Reactor temperature was maintained at 125 °C for each sample. The experimental results for residence times up to almost 40 minutes are shown in Fig. 6. The results illustrate that the ease with which the five samples consume oxygen is quite different, with fish oil being the most easily oxidized. Liquid phase diffusion is slow, and oxidation substantially takes place only on the surface of the sample. If oxidation of the sample is allowed to continue over a long enough period of time, depletion of both the oxygen in the reactor and depletion of the oxidizable material on the surface of the sample (which is only replaced by unoxidized material through diffusion) can have a measurable effect on the observed rate of oxygen consumption. Depletion of oxygen and oxidizable material do not adversely affect the measurement when the extent of oxidation is limited, for example, because of short reaction times or lower reaction temperatures. The data in Fig. 6 show that for reaction times of about 10 minutes or less under these reaction conditions, the observed oxygen consumption is essentially linear with time except for the fish oil sample.

4. Determination of Oxygen Consumed at Different Temperatures

The experimental results of oxidative stability measurements for samples of peanut oil, vegetable oil, com oil, olive oil, cottonseed oil, safflower oil, lard, poultry fat and fish oil at different temperatures are shown in Fig. 7 in the form of an Arrhenius plot. All of the samples exhibit a linear relationship between log(oxygen consumption) and reciprocal temperature, known generally as Arrhenius behavior. The corresponding linear regression parameters and the calculated activation energies are presented in Table 1 below. The intersecting lines in Fig. 7 illustrate the problems inherent in determining the oxidation rates at any given temperature (e.g., 100 °C) and relying on such results to predict relative stabilities at a different temperature (e.g., a storage temperature of 30 °C). The present experimental arrangement, however, permits ready measurement of oxygen consumption at different temperatures and the conespondence to Arrhenius behavior makes it possible to extrapolate such data to another temperature. Direct measurement of oxidation rates at typical storage temperatures would be prohibitively slow for many samples.

As noted previous y, t e most meaning u ata s obtained by the method o the present invention when the overall extent of oxidation is low. Additional testing over a wide temperature range was carried out using samples of com oil and fish oil. These samples were tested under the conditions previously specified but with a reactor residence time of 9 minutes. The results are shown in Fig. 8. For com oil, the data generated corresponds to Arrhenius behavior over the entire temperature range studied, up to 175°C. For the very rapidly oxidized fish oil the linear range extends only to about 95°C. The nonlinearity observed at high temperatures, however, is a measurement artifact (due primarily to the depletion of the oxygen). Linearity, in fact, would extend to higher temperatures if the reactor residence time was sufficiently reduced. 5. Activation Energies Determined Are Independent Of Experimental Variables Activation energies (E) were then calculated for each sample based on the experimental data using the relationship ln(K) = ln(A) - (E/RT). The slope of the log(oxygen consumption) versus reciprocal temperature plot is linearly proportional to the activation energy. As expected, the activation energy for the oxygen consumption reaction that was determined for each sample appears to be independent of the adjustable experimental parameters such as reactor residence times, oxygen concentration, and the like. For example, the activation energies calculated for cottonseed oil based on measurements at three different reactor residence times appear, within experimental error, to be the same: 36.4 ±1.0 kJ/mol (reactor residence time = 5 minutes);

36.4 ±1.3 kJ/mol (reactor residence time = 7 minutes); and 36.4 ±0.9 kJ/mol (reactor residence time = 9 minutes); Similar results were obtained for poultry fat samples (mean activation energy of 36.2 ±OJ KJ/mol calculated from measurements at three different residence times). The activation energies calculated for the samples based on the data obtained from the surface reactor of this

Example compared favorably with the activation energies determined for the samples based on the data obtained from experiments in which the diluent gas was bubbled through the hot sample in a continuous flow-through mode in a similar reactor arrangement where the inlet tube was immersed in the liquid. Olive oil, safflower oil and poultry fat samples were also tested at two different oxygen concentrations of the diluent gas, 0.1% and 5% by volume. Within experimental error, the activation energies calculated were identical despite the 50-fold change in oxygen concentration: olive oil: 45.6 ± 1.9 KJ/mol (0.1% oxygen); 45.9 ± 1.1 KJ/mol (5% oxygen); safflower oil: 43.3 ± 0.9 KJ/mol (0.1% oxygen); 43.3 ± 0.9 KJ/mol (5% oxygen); poultry fat: 36.0 ± 1.1 KJ/mol (0.1% oxygen); 35.4 ± 0.8 KJ/mol (5% oxygen). For easily oxidizable fish oil, however, because the sample oxidizes even more rapidly with 5% oxygen, the activation energy results calculated from measurements at higher oxygen concentration are not the same if other experimental parameters such as reaction temperature range and reactor residence time are not adjusted to reduce the extent of oxidation.

6. Preference for Low Oxygen Concentration in the Diluent Gas

One of the attractive aspects of the present instrument is that multiple temperature runs with replicate measurements at each temperature can be conducted with a sample loaded into the reactor only once. Such single sample testing generally requires that the sample be altered as little as possible during the entire process for the most meaningful results. Lack of sample deterioration can be verified by measuring the PV of the sample before and after the measurement sequence.

A typical experimental mn confirmed this lack of deterioration. The initial PV of a 12 mL cottonseed oil sample differed from the final PV of the same sample by only about 2% when the sample was subjected to a method of the present invention under conditions where the reactor temperature ranged from 75°C to 115°C, residence time was 9 minutes, and OJ % O₂ was used in a diluent gas. This experimental n included triplicate measurements, at 10°C intervals, for a total of 15 measurements. The initial PV value for the above cottonseed oil sample was 3.88 meq/kg and the final PV value was 3.96 meq/kg. Unsaturated linkages are oxidized to peroxides and destroyed during the oxidation process. These can be measured by the Iodine Absorption Number ("IAN"). The IAN value was measured to be 81 both before and after the experiment (an IAN value of 81 corresponds to over 6 moles of unsaturation/Kg). Given that a 2% change in PV corresponds to an insignificant consumption of the unsaturated linkages (less than 0.01%), it is understandable that the IAN value does not change.

Lipid samples can be relatively viscous. If the surface of the sample is fully oxidized, either dispersion of oxygen into the bulk of the sample or replenishment of the surface material from the bulk liquid will become rate limiting. Accordingly, a relatively low oxygen concentration in the diluent gas is therefore preferred in the present surface reactor arrangement. Furthermore, using a low oxygen concentration in the system improves detector linearity and relative noise levels and the oxygen sensor lifetime. The oxygen sensor used in this Example is basically an oxygen- fueled galvanic cell, the lifetime of which decreases as the concentration of oxygen being measured increases.

7. Reproducibility

The reproducibility of the results obtained using the surface reactor arrangement of this Example is generally excellent. A single cottonseed oil sample was studied over a period of three days. Measurements were made 24 hours apart at temperatures of 70°C, 80°C, 90°C, 105°C, and 120°C. Based on the data obtained from 45 specific measurements, the slope and intercept of the log (oxygen consumption) vs. 1/T linear plot exhibited uncertainties of 2.1% and 2.0% for the slope and the intercept, respectively, with a linear r² value of 0.9929. This reproducibility is typical for the surface reactor anangement of the present invention.

8. Oxygen Consumption. IAN And PV Because the oxidation process in oils and fats essentially involves the preferential oxidation of unsaturated linkages resulting in the formation of peroxides, it theoretically should be possible to conelate the extent of oxygen consumption with the decrease in IAN and the increase in PV upon oxidation of the sample. Traditional methods such as the AOM attempt to conelate the change in P V with the ease and extent of oxidation. One problem with this approach, however, is that peroxides themselves are unstable and enoneous conclusions can easily be reached by misinterpreting the results obtained from measuring PV before and after aeration at an elevated temperature. In such methods, although the PV initially increases, the peroxides produced also decompose in an autocatalytic manner. As a result, the PV value reaches a maximum, remains at a steady state for a period, and then declines as the decomposition processes become dominant over the formation processes. The AOM experimental protocol attempts to measure data points on the ascending portion of the curve where the PV is increasing with time. For samples producing very unstable peroxides, these measurements may be extremely difficult or even impossible to obtain. In addition, even accurately measured ascending and descending profiles of the peroxide content with time cannot provide the type of information that can be obtained from the present work. A direct conelation between such PV-related methods and the methods of the present invention is not possible. The following data are illustrative. Samples (22 mL) of peanut oil, vegetable oil, olive oil, cottonseed oil, safflower oil, and fish oil were oxidized by bubbling air (75 mL/min) for 45 minutes through each of the oil sample while maintaining the sample at a temperature of 160°C. The initial and final PV and IAN values of each sample were measured and the results are shown in Fig. 9. The relative oxygen consumption by each of the samples was then calculated using the linear equations in Table 1 (which were derived using the data measured from the flow injection analysis system previouly described above for this Example) and these calculated results also are shown in Fig. 9. The decrease in the IAN value is well conelated (linear r² = 0.9692) with the calculated oxygen consumption. Whether measured from the calculated oxygen consumption or from the decrease in IAN, the sample stability is in the following order: peanut oil > vegetable oil and cottonseed oil > olive oil and safflower oil » fish oil.

In contrast, the increase in PV shows no such conelation with the measurement of oxygen consumption. For example, the PV for the fish oil sample actually decreases after the oxidation. Similar experiments were conducted with these oil samples for shorter oxidation periods of 4, 13 and 25 minutes and similar results were obtained in all cases. The PV increase for peanut oil was the maximum and that for fish oil was the minimum. These results thus suggest that the measured PV increase under these conditions is a reflection on the stability of the specific type of hydroperoxide formed as well as on the stability of the oil. This conclusion is supported by a comparison of the profiles of PV as a function of time measured for different oils and fats that are aerated at an elevated temperature. For example, while the general pattern of ascending and descending PV resembles a Gaussian curve, the maximum of the PV curve obtained with a fish oil sample reaches only a few meq/kg whereas the maximum for a poultry fat sample under the same conditions can reach 250 to 300 meq/kg. The presently proposed method eliminates these problematic issues in the interpretation of PV analyses.

Example 2: Determination of Oxidative Stability of Solid Materials Comprising a Lipid

The oxidative stability of solid materials comprising a lipid was evaluated as set forth below.

Experimental Anangement

The gas phase flow injection analysis system used in this Example is shown in Fig. 10(a) and is substantially similar, but not identical, to the gas phase flow injection analysis system employed in Example 1. The diluent gas flows through a mass flow controller ("MFC") and through an electropneumatically operated six-port stainless steel HPLC-type valve, V (model 7000, Rheodyne, Cotati, CA). Alternatively, two three-way valves or other comparable anangements could be used. The loop of valve V holds the oxidation reactor ("OR") which is shown in Fig. 10(b). The oxidation reactor comprises a stainless steel hollow cylinder (exterior dimensions: 155 mm high x 33 mm diameter, interior dimensions: 55 mm deep x 20 mm diameter), an outer cap (outside diameter, 39 mm; height: 18 mm), and an inner cover insert that ensures a good seal and reduces the reactor volume. An O-ring seal is used to seal the system. The reactor is provided with 1 mm inside diameter nickel inlet and outlet tubes. Two heating tapes (128 x 50 mm, 50 W at 110 V, Thermolyne, IA) are connected in parallel and secured around the oxidation reactor with glass wool thread. A platinum resistance thermometric detector is placed in between the heating tape and the metal exterior of the oxidation reactor. The reactor temperature was programmed with a temperature controller (CN8500, Omega Engineering, Stamford, CT). The solid sample can be readily placed into the reactor from the top of the stainless steel hollow cylinder.

In Fig. 10(a), a cold trap is imposed between the valve V and the oxygen sensor (model 2550, Illinois Instruments, Napierville, IL). The trap consists of a 600 mL beaker filled with water in which a coil of nickel tubing (1.6 mm outside diameter, 1.0 mm inside diameter, 400 mm long, coiled into a 32 mm diameter coil) is immersed, followed by a "buffer" glass tube (8.6 mm inside diameter, 10.2 mm outside diameter, 180 mm long). The cold trap allows the oil vapor to condense and prevents premature poisoning of the oxygen sensor. The trap is periodically cleaned with acetone to remove oil vapor and thoroughly dried before use. The sensor output was recorded by a personal computer (Gateway 2000, P-75) equipped with a DAS- 1601 data acquisition board (Keithley-Metrabyte, Taunton, MA). All tubing used in the gas flow system are 1 mm inside diameter nickel tubes. A homebuilt digital timer was used to control the switching of valve V and hence the residence time of the oxygen-bearing diluent gas in the oxidation reactor.

Preparation Of Solid Samples

Synthetic solid samples of varying lipid content were prepared by using silica gel as the carrier matrix (100 to 200 mesh, ACS Reagent grade, Fisher). Because many silica gel samples contain metallic impurities that appear to greatly accelerate the lipid oxidation rates, each silica gel sample was washed with sodium hydroxide and then copiously washed with water. This washing procedure appears to be effective in deactivating any catalytic sites in the gel. Each sample of washed silica gel was then placed in a conical flask and doped uniformly with a known amount of a solution of cottonseed oil in petroleum ether. The ether was then allowed to evaporate at room temperature under a gently flowing air stream. In addition, solid samples of varying particle size containing olive oil were prepared by using glass beads (ACE Scientific Supply Co., Linden, NJ) as the carrier matrix and doping the beads uniformly with a known amount of a solution of olive oil in petroleum ether similarly to the procedure set forth above. The use of glass beads instead of silica gel avoids the catalytic site problem discussed above. Finally, the bone meal samples tested were supplied by Novus International (St. Louis,

MO). One of the bone meal samples was sieved to produce different particle sizes (ά~: 0.25 to 0.60, 0.60 to 0.85, and >0.85 mm, hereinafter refened to as small, medium, and large particles, respectively).

Procedure

To determine oxidative stability, 1 to 2 g of the solid sample is loaded into the reactor from the top. The reactor is tapped or shaken slightly to smooth and level the sample surface and the reactor cover secured. The system is mn with the following parameters: the diluent gas is 0.1% O₂, 99.9% N₂ (purchased from TRIGAS Industrial Gases, Lubbock, TX); the diluent gas flow rate is 35 cm³/min; the reactor flush time to introduce fresh gas is a minimum of 1 minute; and the reactor residence time is 5 minutes. The system is run for about 20 to 30 minutes at the first temperature point before any data are recorded.

Performance and Reproducibility The oxidative stability of a set of eight bone meal samples was determined as set forth above. Fig. 11 is an Anhenious plot reporting the data obtained and illustrating the relationship between log(oxygen consumption) and the reciprocal of the absolute temperature for the eight samples. The linear r² values and the conesponding activation energy values are also shown in Fig. 11. For the eight samples studied, the oxygen consumption exhibited Anhenius behavior as shown. The results indicate that the samples fall in two different families: a first group of three samples exhibiting larger oxygen consumption values and larger

ΔH values, and a second group consisting of the remaining samples. The larger values exhibited by the first group suggest that these samples have already turned rancid.

Fig. 12 reports experimental data obtained according to the method of this Example from two measurements on different days for three samples (sample weight for #1793 and #1893: 1.0 g; for # 2560: 2.00g). The reproducibility of the results of this oxidative stability determination appears to be very good.

Under the experimental conditions of this Example, the sample placed into the oxidation reactor forms a relatively shallow layer in the reactor (typically less than or equal to about 7.5 mm in depth), such that the gas can readily contact the sample surface. Consequently, the extent of oxygen consumption is observed to be linearly dependent on the amount of the sample with the linear relationship having a statistically zero intercept. Fig. 13, for example, illustrates this linear dependency for a sample subjected to a reactor temperature of 135 °C and a reactor residence time of 5 minutes. For the oxidation reactor of the present Example (which has a cross sectional area of 3.14 cm² and a total volume for sample plus gas of 16 cm³), the reproducibility of the results decreases markedly for sample amounts of about

2 g or greater. This decrease in reproducibility likely is related to the access of the gas in the reactor to the entire sample surface. The reaction between the oxygen in the diluent gas and the lipid in the solid sample is a heterogeneous two-phase reaction. If the sample depth becomes too large or if the sample is loaded in a tight and compacted fashion, the accessibility of the gas to the sample surface will vary from run to ran and will affect the results. This situation is exacerbated when the particle size is small and the sample is prone to being packed tightly. When the sample amount is small, the sample occupies a relatively thin layer in the reactor and the access of the gas to the sample surface is facile. Accordingly, limited access of the oxygen to the entirety of the sample generally is not a significant problem under such conditions. By way of comparison, the amount of a liquid sample loaded into a reactor having a constant cross-sectional area (and thus a constant sample surface area) and adequate headspace after the sample is loaded does not affect, within experimental enor, the rate of oxygen consumption because the available surface for interaction is only the top surface of the sample.

Choice of Oxygen Concentration As previously discussed herein, the absolute amount of oxygen consumed depends in a first-order fashion on the oxygen concentration. Accordingly, the quality or the nature of the kinetic or thermodynamic data obtained is not dependent on the precise value of the oxygen concentration. Provided that either the oxygen concentration used is accurately known, or all experiments are conducted at the same concentration, data for different samples can be compared on a common basis. Like the flow injection analysis system of Example 1, the advantages of using a low oxygen concentration in the system of the present Example include (a) a longer lifetime for the galvanic sensor, (b) better sensor response linearity, and (c) minimum change of sample composition due to oxidation, allowing reproducible multi- temperature runs on the same sample aliquot.

Effect of the Sample Lipid Content

Four silica gel matrix samples containing 3.4%, 6.3%, 9.1% and 12% of cottonseed oil by weight, respectively, were studied. These lipid content values represent a typical range of the minimum lipid content for many materials comprising a lipid that are of commercial interest. The oxygen consumption rates for these four samples were measured at a reactor temperature of 130°C and a residence time of 12 minutes. Fig. 14 reports the experimental results and illustrates that, in the range studied and within reasonable limits, the lipid content of the samples has no effect on the oxygen consumption. If the silica gel is not doped with the oil at all as in the control, however, there is no oxygen consumption.

Similar results were also found using bone meal samples. To avoid the influence of other confounding factors such as varying particle size or composition, the experiments were carried out using the above-described synthetic samples so that a constant particle size could be maintained while varying the lipid content. The results obtained from both the bone meal samples and the silica gel matrix samples suggest that given a minimum lipid content sufficient to coat the surface, the rate of oxygen consumption is limited by other factors such as accessible surface area.

Effect of the Particle Size of Samples The chemical reaction between oxygen and the lipid occurs on the surface of the sample particles. The overall reaction rate is directly related to the total surface area of the sample in the reactor. For equivalent sample masses, the particle size of the sample governs the available surface area. Accordingly, the particle size can be related to the oxidative stability of a sample. Fig. 15 shows the experimental results obtained from testing the following four bone meal samples: (1) an original composite bone meal mixture, (2) a sieved fraction derived from the original mixture and having an average particle size between 0.250 to 0.600 mm, (3) a sieved fraction derived from the original mixture and having an average particle size between 0.600 to 0.850 mm, and (4) the sieved fraction derived from the original mixture and having an average particle size greater than 0.850 mm. The oxygen consumption rates by the sieved fractions decreased with increasing particle size with a relative order of oxygen consumption according to particle size as follows: 0.250 to 0.600 mm > 0.600 to 0.850 mm > 0.850 mm. The behavior of the sample of the original, unfractionated composite mixture was virtually the same as the behavior of the sieved sample of intermediate particle size.

In a related set of experiments, the dependence of lipid oxidation rate and stability on the sample surface area and, therefore, on particle size was evaluated for the four bone meal samples by extracting the lipid from each of the samples and determining the stability of the extracted lipid. The results for the extracted samples were comparable thus indicating that lipid oxidation rate and stability are influenced by particle size. Typically, it is preferable that samples be analyzed in the form they are stored, without further grinding or processing. If it is of interest to compare the intrinsic stabilities of two sample types, however, the analysis should employ samples having similar particle size to obtain the most meaningful comparison.

In yet another set of experiments, the observation that oxygen consumption rates decreased with increasing particle size was further tested using synthetic samples of uniform particle size glass beads. The four samples (2.0 g sample amount) employed lipid-coated glass beads having a particle diameter of 0J2 to 0J6 mm, 0.25 to 0.30 mm, 0.45 to 0.50 mm, and

1.0 to 1.05 mm, respectively. The lipid content of all the samples was held constant at 2.0% by weight olive oil and the reactor temperature was maintained at 120°C. The results are reported in Fig. 16. The following relative order of oxygen consumption according to particle size was determined : 0.12 to 0J6 mm > 0.25 to 0.30 mm > 0.45 to 0.50 mm > 1.0 to 1.05 mm particle diameter. This relative order is consistent with the relative order determined above for the bone meal samples. If the mean diameter in each size class is defined as the geometric mean of the range extremities, the oxygen consumption rate is observed to be linearly related to the reciprocal of the mean diameter of the sample particles for both the bone meal (linear r² value of 0.9567 using a reactor temperature of 86°C) and glass bead samples (linear r² value of 0.9650). For equivalent sample masses, the oxygen consumption rate is proportional to the total surface area of the sample.

Oxygen Consumption and Four-Hour AOM.

In practice, the oxidative stability of samples is often determined using an abbreviated version of the AOM in which the peroxide value is determined after 4 hours of oxidation under the conditions of the traditional AOM. Because the method of the present invention produces a different type of information than the traditional or abbreviated AOM methods, the results of these methods cannot be directly compared. Because the intent of both methods, however, is to provide some measure of oxidative stability, one indirect basis for comparison involves the comparison of the 4 hour PV data obtained by the abbreviated AOM with the data calculated by the method of the present invention, the latter data being further reduced to one parameter: the oxygen consumption at 98 °C. For four bone meal samples, the 4 hour PV was determined by a certified commercial laboratory. These results are shown in Fig. 17 together with the oxygen consumption values predicted for the same samples at 98 °C in accordance with the method of the present invention. Although the conespondence is not exact, the general trend of the data is similar. This conespondence suggests that the data from the new method can be related to historical information obtained with the older techniques. It also suggests that the new method will be useful in the comparison of relative efficacies of the various antioxidants available to the industry.

In view of the above, it will be seen that the several objects of the invention are achieved. As various changes could be made in the above methods and apparatus without departing from the scope of the invention, it is intended that all matter contained in the above description be interpreted as illustrative and not in a limiting sense. All documents mentioned in this application are expressly incorporated by reference as if fully set forth at length.

When introducing elements of the present invention or the prefened embodiment(s) thereof, the articles "a", "an", "the" and "said" are intended to mean that there are one or more of the elements. The terms "comprising", "including" and "having" are intended to be inclusive and mean that there may be additional elements other than the listed elements.

REFERENCES 1. Brodnitz, M. H., Autoxidation of Saturated Fatty Acids, J. Agr. Food Chem.,

16: 994-999 (1968).

2. Bloomfield, G. F.; Sundralingham, A.; Stutton, D. A. Trans. Faraday Soc, 1942, 38, 348-355.

3. Bolland, J. Quart. Rev.(London), 1949, 3, 1-21. 4. Farmer, E. H., Bateman, L. Quart. Rev. (London), 1954, 8, 147-167.

5. Swem, D. in "Autoxidation and Antioxidants," Vol. 1, Lundberg, W. O., Ed., Interscience, New York, 1961; pp. 1-54.

6. Hawkins, E. G. E. "Organic Peroxides " Van Nostrand, New York, 1961, pp. 355- 409. 7. Labuza, T. P., Crit. Rev. Food Technol, 1971, 2, 365-368.

8. Labuza, T.P.; Bergquist, S. J. Food Sci. 1983, 48, 712-715. 9. Privett, O. S. and Blank, M. L. J. Am. Oil. Chem. Soc, 1962, 39, 465-469.

10. Shahidi, F. in Natural Antioxidants. Chemistry, Health Effects and Applications, Shahidi, F. Ed. AOCS Press, Champaign, IL 1997. pp 1-11.

11. Wewala, A. R. In Natural Antioxidants. Chemistry, Health Effects and Applications, Shahidi, F. Ed. AOCS Press, Champaign, IL 1997, pp 1 J 7, 331 -345.

12. Rackis, J. J., Sessa, D. J., and Honjg, D. H., Flavor Problems of Vegetable Food Proteins, /. Am. Oil Chem. Soc, 56:262-211 (1979).

13. Kaneda, T. and Ishii, S., Nutritive Value or Toxicity of Highly Unsaturated Fatty Acids, Biochem. (Tokyo), 7:327-335 (1954). 14. Moll, C, Biermann, U., and Grosch, W., Occunence and Formation of Bitter-

Tasting Trihydroxy Fatty Acids in Soybeans, J. Agric. Food Chem., 27: 239-243 (1979). 15. Steinberg, D., Parthasarathy, S., Carew, T. E., Choo, J. C, and Witztum, j. L., Beyond Cholesterol. Modifications of Low-Density Lipoprotein that Increase its Atherogenicity, New Engl. J. Med., 320: 915-924 (1989). 16. Harman, D. __ Free Radicals in Biology, Vol. 5, W. A. Pry or, Ed., Academic

Press, New York, 1982, pp. 255-275.

17. Stine, W. R; Faut, O. D.; Rees, M.; Stockham, E. B. In Applied Chemistry, 2nd ed. Allyn and Bacon, Inc., Boston, 1981, pp. 380-383.

18. Yamada, J. Bull. TokaiReg. Res. Lab. 1979, 99, 23-29. Yamaguchi, K.; Toyomizu, M.; Nakamura, T. Bull Jpn. Soc. Sci. Fish. 1984, 50, 1245-1249.

19. Swem, D. In Organic peroxides, Vol. 1, Wiley-Interscience, New York, 1970, pp 115-128.

20. Helrich, K. Ed. In Official Methods of Analysis of the Association of Official Analytical Chemists, 15th Ed. The Official Methods of Analysis of Official Analytical Chemists, Arlington VA, 1990. pp 955-956.

21. Coppen, P. P. In Rancidity in Foods, Allen, j. C. and Hamilton, R. j., Eds., 2nd ed., Elsevier Applied Science, London, 1989, pp 83-104.

22. American Oil Chemist's Society Official Method Cd-12-57, Fat Stability, Active Oxygen Method, 1993. American Oil Chemists Society, Champaign, IL. 23. Swem. D. Ed. Bailey 's Industrial Oil and Fat Products, Vol. I, Wiley, New York,

1996.

24. American Oil Chemist's Society Official Method Cd-12b-92, Oil Stability Index, 1993. American Oil Chemists Society, Champaign, IL.

25. Laubli, M. W.; Bruttel, P. A. J. Am. Oil Chem. Soc. 1986, 63, 792-795; deMan, j. M.; Tie, F.; deMan, L. J. Am. oil Chem. Soc. 1987, 64, 993-996.

26. Hill, S. E.; Perkins, E. G. J. Am. Oil Chem. Soc. 1995, 72, 741-743. 27. Bunick, F. J. In Lipid Autoxidation in Human Blood Cell Membrane, Ph. D. Dissertation, University of Massachusetts at Amherst, Massachuset, 1984.

29. Chen, Z. Y., Nawar, W. W. J. Am. Oil Chem. Soc. 1991, 68, 47-50.

30. Ohshima, T.; Fujita, Y.; Koizumi, C. J. Am. Oil Chem. Soc. 1993, 70, 269-276.

31. Chen. Z. Y., Ratnayake,W. M. N., and Cunnane, S. C, Oxidative Stability of Flaxseed Lipids During Baking, J. Am. Oil Chem. Soc,, 77: 629-632 (1994).

32. American Oil Chemist's Society Official Method Cd-8-53, 1996. American Oil Chemists Society, Champaign, IL.

33. Yoshikawa, T., Yamaguchi, K., Kondo, M., Mizakawa, N., Ohta, T., and Hirakawa, K., Thiobarbituric Acid Reactive Substances and Vitamin E in Serum and Cerebrospinal Fluid from Patients with Cerebral Apoplexy, Arch. Gerontol Geriatr., 1: 209- 218 (1982).

Claims

WHAT WE CLAIM IS:

1. A method for estimating the oxidative stability of an oxidizable material, the method comprising: determining the oxygen consumption by a sample of the material at a first temperature; determining the oxygen consumption by a sample of the material at a second temperature; and estimating the oxidative stability of the material at a third temperature by extrapolating the temperature-dependent data obtained at the first temperature and the second temperature to the third temperature.

2. A method for estimating the oxidative stability of an oxidizable material comprising a lipid, the method comprising: determining the oxygen consumption by a sample of the material at a first temperature; determining the oxygen consumption by a sample of the material at a second temperature; estimating the oxidative stability of the material at a third temperature by extrapolating the temperature-dependent data obtained at the first temperature and the second temperature to the third temperature.

3. The method of claim 2 wherein the material is a fat.

4. The method of claim 2 wherein the material is an oil.

5. The method of claim 2 wherein the material is a liquid at both the first temperature and second temperature.

6. The method of claim 2 wherein the material is a solid at one or both of the first temperature and second temperature.

7. The method of claim 2 wherein extrapolating the temperature-dependent data comprises evaluating a plot of log(oxygen consumption) against (absolute temperature)^"1 prepared from said temperature dependent data.

8. The method of claim 2 wherein extrapolating the temperature-dependent data comprises determining a linear equation for estimating oxidative stability using said temperature-dependent data.

9. The method of claim 8 wherein the linear equation is in the form of log(oxygen consumed) = -a/T + b, wherein a and b are constant parameters calculated on the basis of the temperature dependent data and T is absolute temperature.

10. The method of claim 2 wherein the material is placed in an adjustable temperature reactor, said reactor comprises a reaction chamber having an intemal volume, an inlet through which a gas may be supplied to the internal volume of the reactor, and an outlet from which a gas may be removed from the internal volume of the reactor, and wherein a sample of the material having a defined sample volume when placed in the reactor has a defined surface area in contact with a gas supplied to the internal volume of the reactor, and wherein the reactor head space is the internal volume of the reactor less the sample volume.

11. The method of claim 2 wherein the amount of the sample used in each determination step of the method is between about 0.001 gram to about 100 grams.

12. The method of claim 2 wherein the amount of the sample used in each determination step of the method is between about 0J gram to about 20 grams.

13. The method of claim 2 wherein the amount of the sample used in each determination step of the method is between about 0J gram to about 15 grams.

14. The method of claim 2 wherein the amount of the sample used in each determination step of the method is between about 0J gram to about 10 grams.

15. The method of claim 2 wherein the sample used in each determination step of the method is a solid at the temperature of said step, and said sample is between about 0J gram to about 5 grams in weight.

16. The method of claim 2 wherein the sample used in each determination step of the method is a solid at the temperature of said step, and said sample is between about OJ gram to about 3 grams in weight.

17. The method of claim 10 wherein the gas supplied to the internal volume of the reactor comprises oxygen and a diluent gas.

18. The method of claim 17 wherein the diluent gas is an inert gas selected from the group consisting of helium, neon, argon, krypton and xenon.

19. The method of claim 17 wherein the diluent gas is nitrogen.

20. The method of claim 17 wherein the diluent gas is carbon dioxide.

21. The method of claim 17 wherein the gas supplied to the reactor has an oxygen concentration of less than about 20% by volume.

22. The method of claim 17 wherein the gas supplied to the reactor has an oxygen concentration of less than about 5% by volume.

23. The method of claim 17 wherein the gas supplied to the reactor has an oxygen concentration of between about 0.00002% to about 20% by volume.

24. The method of claim 17 wherein the gas supplied to the reactor has an oxygen concentration of between about 0.002% to about 5% by volume.

25. The method of claim 17 wherein the gas supplied to the reactor has an oxygen concentration of between about 0.0001% to about 2% by volume.

26. The method of claim 17 wherein the gas supplied to the reactor has an oxygen concentration of between about 0.0001% to about 1% by volume.

27. The method of claim 17 wherein the gas supplied to the reactor has an oxygen concentration of between about 0.0001% to about 0.5% by volume.

28. The method of claim 17 wherein the gas supplied to the reactor has an oxygen concentration of between about 0.01% to about 0.3% by volume.

29. The method of claim 2 wherein the oxygen consumption by the material at each temperature is determined over a period of between about 10 seconds to about 60 minutes.

30. The method of claim 2 wherein the oxygen consumption by the material at each temperature is determined over a period of between about 30 seconds to about 40 minutes.

31. The method of claim 2 wherein the oxygen consumption by the material at each temperature is determined over a period between about 1 minute to about 30 minutes.

32. The method of claim 2 wherein the oxygen consumption by the material at each temperature is determined over a period between about 2 minutes to about 20 minutes.

33. The method of claim 2 wherein the oxygen consumption by the material at each temperature is determined over a period between about 3 minutes to about 10 minutes.

34. The method of claim 10 wherein the ratio of the reactor head space to the sample surface area is between about 3 cm³/cm² to about 9 cm³/cm².

35. The method of claim 10 wherein the ratio of the reactor head space to the sample surface area is between about 4 cm³/cm² to about 8 cnrVcm².

36. The method of claim 10 wherein the ratio of the reactor head space to the sample surface area is about 6 cm³/cm².

37. The method of claim 10 wherein the internal volume of the reactor is between about 10 cm³ to about 100 cm³.

38. The method of claim 10 wherein the reactor is a surface reactor.

39. The method of claim 10 wherein the surface area of the material exposed to the oxygen is less than about 10 cm².

40. The method of any claim 10 wherein the surface area of the material exposed to the oxygen is less than about 8 cm².

41. The method of claim 10 wherein the surface area of the material exposed to the oxygen is less than about 6 cm².

42. The method of claim 2 wherein less than 10% by weight of the material is oxidized during the oxygen consumption steps.

43. The method of claim 2 wherein less than 5% by weight of the material is oxidized during the oxygen consumption steps.

44. The method of claim 2 wherein less than 3% by weight of the material is oxidized during the oxygen consumption steps.

45. The method of claim 10 wherein the material is a solid at one or both of the first temperature and second temperature, and said solid material forms a substantially uniform layer in the reactor with a layer depth less than about 10 mm.

46. The method of claim 10 wherein the material is a solid at one or both of the first temperature and second temperature, and said solid material forms a substantially uniform layer in the reactor with a layer depth less than about 8 mm.

47. The method of claim 2 wherein the amount of time required to estimate the oxidative stability of the material is less than about 4 hours from the commencement of the method.

48. The method of claim 2 wherein the amount of time required to estimate the oxidative stability of the material is less than about 3 hours from the commencement of the method.

49. An apparatus for estimating the oxidative stability of an oxidizable material, the apparatus comprising: a reactor comprising a reaction chamber having an inlet and an outlet, said reaction chamber adapted to receive and hold a sample of an oxidizable material; a means for controlling the temperature of the reaction chamber; a means for sensing the oxygen content of a diluent gas comprising oxygen; a means for controlling the flow of the diluent gas; and a valve means for directing the flow of the diluent gas comprising oxygen, said valve means having a plurality of operating positions, wherein the inlet of the reaction chamber is in fluid communication with a source of said diluent gas and the outlet of the reaction chamber is in fluid communication with the oxygen sensing means when the valve means is in a first position, and the diluent gas passes from the source of diluent gas to the oxygen sensing means without first passing through the reaction chamber and the inlet and outlet of the reaction chamber are substantially sealed when the valve means is in a second position.