FR2459471A1 - METHOD AND APPARATUS FOR MEASURING THE WATER CONTENT OF A MATERIAL, IN PARTICULAR A FOOD PRODUCT - Google Patents

Abstract

THE INVENTION CONCERNS A METHOD FOR MEASURING THE WATER CONTENT OF MATERIALS. WATER IS EXTRACTED FROM THE MATERIAL BY IMMERSING IT IN A HYDROPHILIC AGENT, SUCH AS ETHANOL, CONTAINING A COLORED REAGENT, SUCH AS COBALT CHLORIDE, AND MEASURING THE VARIATIONS IN COLOR APPEARING AS A FUNCTION OF THE QUANTITY OF WATER PRESENT. THE PROCESS MAY BE CARRIED OUT IN A CONVENTIONAL SPECTROPHOTOMETER, OR, MORE ADVANTAGEOUSLY, IN A SPECIALLY DESIGNED COLORIMETER. THIS INCLUDES A HEATING BLOCK 1 MAINTAINED AT A CONSTANT TEMPERATURE AND INCLUDING A COMPARTMENT 2 WHICH RECEIVES A CELL CONTAINING A SAMPLE OF PRODUCT, A COMPARTMENT 5 CONTAINING A RED LIGHT DIODE WHICH EMITS AT 655NM MEASUREMENT OF 6 AND A COMPARTIO N PERCENTAGE OF TRANSMISSION THROUGH THE SAMPLE. THE INVENTION APPLIES IN PARTICULAR TO THE MEASUREMENT OF THE WATER CONTENT OF FOOD PRODUCTS.

Description

245947 1

  The present invention relates to a method for measuring the water or moisture content in materials, particularly food products. A simple colorimeter, which implements

  a preferred form of the process is also described.

  There are a number of techniques, for example

  measuring the dielectric constant or the conductivity, to obtain

  the moisture content of a food product, but these techniques rarely apply to materials other than those for which

they were designed.

  For various reasons, other methods, more

  are unsuitable for the intended purpose, as is the case of the Karl Fischer process, for example. It is generally rejected by food analysis specialists because it requires

  frequent calibrations and involves the use of unpleasant reagents.

  Nevertheless, this is a well-established process, the application of which

  is mainly limited to the determination of the humidity of

dehydrated foods.

  The moisture balance process is another rather limited general application technique, as it is only suitable for the quality control of unprepared foods, mainly because the food sample needs to be hardened and any

  the humidity is not eliminated.

  Another well-known technique involves a measurement of the near-infrared reflectance, which provides more accurate results. However, the instruments necessary for measuring the near-infrared reflectance are expensive and require sophisticated calibrations. Due to the need to maintain absolutely reproducible surface properties on the sample, the calibrations performed are product-specific. Moreover, this technique proves to be difficult

  to be applied to foods with high moisture content.

  A method using low resolution nuclear magnetic resonance (NMR) has also been employed to determine

  the moisture content of food products without fat.

  Such a technique is expensive, and it seems that the results obtained depend on the state of relaxation of the water, this tending to make this technique specific to the product. When a subject

  liquid fat is present in the sample, it appears to be important

  interferences with the useful signal, since the protons

  fat is detected in the same way as water.

  An additional disadvantage common to all these

  prior art is that they are slow to implement.

  The object of the invention is to propose a method

  both a measure of the moisture content of materials, especially

  food products, which is fast, correct and economical.

  feasible, substantially independently of the

  water of materials subject to measurement.

  In summary, according to the invention, the water is extracted from the material, the moisture content of which must be determined, by keeping the material in a suitable hydrophilic organic solvent, and the water content of the extract is determined by colorimetry. in

  a colorimeter or a spectrophotometer.

  The basis of the invention is based on known reactions of inorganic salts, these reactions being, however, undisclosed so far used to determine the water content

in an organic solvent.

  According to the invention, it is appealed, for the purpose

  colorimetric elimination indicated above, to a suitable transition metal salt, the metal ion of which, in a solvent,

  to conceal, in the form of a complex, as is known, the molecules

  and the ions of the solvent, the complex, which contains only solvent molecules, giving rise to a first color and being in equilibrium with complexes that also contain ions and give rise to another color. This balance can be moved in both directions according to variations

  respective concentrations of the solvent or ion concentration.According to the

  the ability of the extracted water to move the equilibrium prevailing in the solution system in a direction to the first color, which leads reduction of the intensity of the second color, which is measured to determine the quantity

extracted water.

  It should be noted that, in this context, several transition metals have already been used as qualitative indicators of moisture content. Cobalt chloride paper, gel

  self-indicating silica and the safety stain of trigger pears

  Many cameras use the same property. However, the circumstances in which cobalt or another transition metal is employed, according to the invention, allow a

  quantitative measurement of the water content of the material.

  A first important advantage conferred by the method of the invention is speed. The determination of the moisture content is limited only by the speed with which the extraction of moisture can be accomplished and, possibly, the

  filtration of the solid residues of the extract.

  A second important advantage of the process of the invention

  tion is that accuracy of results is not sacrificed to speed. Experience shows that it is possible to obtain without

  great effort, a relative error of + 1% (with

  95X confidence level /). It is to be hoped that, in a mixture and closed calorimetric measurement system, the error will be even smaller because of the decrease in the number of

  manipulations of the solution and, consequently, of less contamination

  by water. In addition to the above, the method of the invention has the advantage that it can be applied to almost any food or other material from which water can be extracted, such as textiles or gas. Minimal changes are required to cover the moisture content values of most high-moisture food products. By relatively simple changes, for example regarding the concentration of reagent, it is possible to treat

  almost anhydrous food products.

  It is envisaged according to the invention to be able to extract water quantitatively from many food products by leaving them in a suitable hydrophilic liquid. Among the hydrophilic liquids suitable for the process of the invention, the lower alcohols have many desirable characteristics,

  such as low cost, low toxicity and

  These properties recommend them as solvents for the extraction process of the process of the invention. Lower alcohols have already been used in analysis as a means of quantitatively extracting water from food products prior to moisture measurement in Karl Fischer processes and infrared reflectance measurement. It is better to use, as a solvent

  for extraction and the color reaction, ethanol or dna-

  because, in addition to the desirable characteristics

  above, ethanol is readily available but not very inflamed

  mable, can be thrown with a minimum of precautions and presents a

strong water absorption capacity.

  The terms "moisture" or "water" refer to the

  context of this description, moisture or water that can be

  extracts of the food product or other material by a solvent

  hydrophilic organic. It may happen that certain food products

  are resistant to desiccation with a hydro-

  phile, so that it is not possible to make quantitative extracts. The water retained in the form of crystalline hydrates, for example in the lactose monohydrate of dairy products, may

not to be extracted.

  To accelerate the extraction of moisture or water from the material, it would be advantageous to mix it well with the extracting solvent. The use of a laboratory homogenizer or a high speed mixer of the similar type is satisfactory for this purpose. It is best to use a blender that can be

  hermetically sealed to prevent any solvent loss by evapo-

  or any absorption of moisture from the atmosphere by hygroscopic solvents such as ethanol. However, even when using an open top-mounted homogenizer, it appears that atmospheric moisture introduces only a very small error. The amount of solvent required depends on the nature of the material or sample to be extracted

  and the water content present in the sample. Generally speaking

  In addition, an excess of solvent must be added to the sample in proportion to the water content of the sample in order to obtain the highest possible quantitative extraction of water, this excess being most often easily determined by an experienced person from the knowledge she has about different types of materials. If necessary, several extractions can be carried out successively on identical samples, the colorimetric result of the preceding extractions being used to adjust the necessary amount of solvent. An excess of solvent reaching several times the water content of the sample should be preferred, although this is not critical. Of course, if the material to be extracted is a solid, the extract must be filtered to remove the particles.

  which would otherwise intervene in the colorimetric

  cudgel. Such filtration can be carried out in the fastest manner by means of glass fiber filter papers. Such papers also have the advantage of being able to be oven dried and preserved in a desiccator so that they do not bring

of water to the extract.

  The equilibrium described above between the complex-related solvent molecules and the ions bound to the complex is strongly temperature dependent, so that to obtain the correct calorimetric measurements it is necessary to keep the temperature as constant as possible for as long as possible. colorimetric determination. However, the prior evolution of the solvent that undergoes the colorimetric measurement is not important because the color variations, that is to say the shifts in the equilibrium indicated above, which are due to variations in temperature, are instantaneous, reversible and devoid of hysteresis. Therefore, the measurement temperature need not be known

  precisely, the important thing being that it remains constant.

  If the material to be extracted contains certain ions, the extract must be saturated with these ions so that the ions contained in the material do not displace the aforementioned equilibrium of the complex-to-complex relationship. By achieving a saturation in these ions, we obtain one and the same

  influence of these ions on the color reaction in the solution

put to the measurement.

  This factor, along with others, influences measurement

  colorimetric, as well as the variations in volume, but the

  following description of an embodiment using the chloride

  of cobalt as a color reagent for colorimetric determination.

  water content will provide a better understanding of the features and benefits of the invention and

  clarification of the issues outlined above; the description

  is based on the attached drawings, among which:

  - Figures 1 to 4 illustrate the functional relationships

  between optical amplitudes specific to samples

  and the water content of these samples; FIGS. 5A and 5B are respectively a plan view and a side view of a colorimeter according to the invention;

  FIG. 6 illustrates a cell, or bowl, belonging to

  according to the colorimeter of the invention;

  FIG. 7 illustrates a light source of the color

  millimeter of the invention; FIG. 8 illustrates a photometric detector used in the colorimeter of the invention; FIG. 9 is a block diagram of a bridge circuit; FIG. 10 is a block diagram of the detector circuit; and

  FIG. 11 is a diagram indicating the operations

  to be performed according to a preferred form of the process of the invention.

  A particular advantage resulting from the use of cobalt is that the color variations associated with it are intense. This makes it possible to use very dilute indicator solutions, such a characteristic being essential if a

  sufficient amount of cobalt indicator is dissolved in soil.

  substantially anhydrous organic

  Cobalt ions in solution are never isolated but are always associated with other ions and

  solvent. As part of this description, it will be assumed that the

  cobalt can exist in solution in two forms, depending on the

  sition of molecules and ions around cobalt ions. In water, cobalt has an octahedral organization and is entirely surrounded by water molecules. The cobalt salt solutions have a characteristic pale pink color. In organic solvents such as ethanol or other lower alcohols, the cobalt ions are surrounded not by water but by molecules

  of solvent, even if the octahedral organization is preserved. always

  however, some of the cobalt ions in the organic solutions are not well protected by the solvent and surround themselves with ions. The dissolution of cobalt chloride in ethanol gives rise, for example, to situations where a certain number of cobalt ions are surrounded by chloride ions. When this happens, cobalt takes on a tetrahedral organization and strongly absorbs red light. Solutions containing tetrahedral cobalt

  so are not pale pink but intense blue.

  If water is added to a tetrahedral cobalt solution, the water molecules help the organic solvent to protect the cobalt ions. As a result, the blue color is added as water is added and, although this is not visible to the eye, the pale pink color slightly increases in color.

c8té.

  The pink color is such a weak color that it is of little use to indicate the amount of water present. However, the blue color is so deep that it can be used to achieve

  a sensible measure of the amount of water in solution.

  It appeared that the blue color of cobalt tetra-

  hedric was broadband absorption with a very large peak width (approximately 100 nm). When ethanol is used

  as the solvent, the peak of the peak is 655 nm, this difference

  slightly from that which corresponds to other organic solvents, polar, for example acetone. This fact is proving to be of great

  importance, since the wavelength of 655 nm is almost

  that of the maximum emission of light-emitting diodes.

  These diodes are inexpensive sources of a relatively intense monochromatic light whose use is adapted to the color reaction of cobalt. Being able to have such a simple light source makes it possible to design a special colorimeter that can be

  specifically used for the determination of the water content.

  What has just been said above makes it possible to understand why the colorimetric measurement is not conventional. Ordinarily, color enhancement is associated with an increase in concentration in the species under consideration. In the context of the description, the pink color actually increases, but not in a way

  practically usable. Conversely, there is a decrease

  much stronger color, this decrease

  being the measured quantity. The blue color results from the

  tion of chloride ions around cobalt and the intensity of this color depends on the number of tetrahedral cobalt ions present in the solvent. This number, in addition to depending on the nature of the solvent and the total concentration of cobalt ions, depends

  also the total concentration of chloride.

  Many food products contain chloride ions from internal sodium chloride, these ions

  able to pass into solution at the same time as the water during extrac-

  described above. If a cobalt chloride color reagent is used for the colorimetric determination of the water content of chloride-containing food products, the chloride concentration of the extract to which the color reagent has been added varies with the chloride content of the food product. The chloride content of food products differ widely, and it is not possible to define different classes of foods on the basis of their chloride content.

  correct the effect of the chloride thus extracted on the color reaction.

  Since it is not practically feasible to remove the chloride from the samples before extraction, another possible approach is to ensure that all the extracts are saturated with chloride.

sodium.

  The decrease in the blue color as water is added to the ethanolic cobalt chloride does not obey the Beer-Lambert laws. The theoretical reasons are known,

  but they will not be described for the sake of clarity.

  Instead of the linear relationship predicted by the Beer-Lambert laws for an additional chloride-free system, there is a semilogarithmic relationship between the water concentration and the absorption coefficient at 655 nm. This relation is illustrated in FIG. 1. The equation of the curve of FIG. 1 is given by M = -1.495 Log. k + 1.607 (1) where M is the mass of water and k is the absorption coefficient. The correlation coefficient r between the logarithm of the coefficient

  absorption and the mass of the water is 0.99996. It expresses a rela-

  virtually perfect functional

  As explained above, the presence of salt in foods requires that the reaction system be

  saturated with sodium chloride to account for salinity

  variable number of samples. As necessitated by the theory of the color reaction, the useful range of the color changes undergoes dilation, but the relation ceases to be perfectly semilogarithmic and the deformation thus introduced is visible on the curve of FIG. possible to use such a curve for calibration, but it would be easier to have another straight line. In particular, it is known that drawing a straight line requires fewer calibration tests than a curve. An appropriate linear relationship between the water concentration and an optical property has been found, and it is the transmission percentage T that has been used instead of the absorption coefficient, the graph then becoming an S-shaped curve. very flat. A property of such graphs is

  that their central parts are approximately straight lines.

  The linear relation between the water concentration and the percentage of transmission of the solution is illustrated on the

figure 3.

  The water concentration shown in Figure 3 is

  expressed in the same units as those previously used.

  A consequence of this is that the upper points do not seem to fit in the straight line. The reason is purely due to the way the concentration is expressed, as will be explained below. As shown in Figure 3, the working range of the reaction is narrow. When using a conventional spectrophotometer, it is axiomatically true that measurements exceeding one unit of absorption factor (ie below 10% of transmission) become less accurate because the detector has a small amount of light compared to the stray light inside the instrument. For typical cobalt solutions placed in a cell 10 mm in length, the useful range of 10 to 80% transmission is

  variation in moisture content by 4% (by weight by volume) only.

  For example, a content of 1 to 5% of water covers the precisely useful range. Such a situation is not really a disadvantage, since

  the water whose content is to be measured has already been extracted from the

  and is necessarily a fairly well diluted solution of water in alcohol. Adding the alcoholic cobalt chloride reagent to the extract further reduces the water concentration. The narrowness of the measurement range associated with the reaction therefore has the effect of compensating for the narrowness of the absolute range of water contents.

  samples obtained by extraction.

  Once extracted, the water increases the volume of the solvent.

  The color reaction is carried out according to the concentration of

  water and is therefore influenced not only by the chemical modifications

  resulting from the presence of a certain body of water, but

  also by the volume variation introduced by the water.

  There are two ways to express the water concentration in the color system. First, it is possible to perform measurements in terms of water mass in a given volume of anhydrous extraction solvent. It is on this basis that FIGS. 1, 2 and 3 have been constructed, with the advantage that the volume of extracting solvent can be precisely determined. However, when using these units, calibration is not linear for high water levels, since the change in volume

  brought by the water begins to be appreciable.

  This is the reason for the curvature presented by the

figure 3.

  The other way to proceed is to express the

  concentration of water measured as a mass in a

  some volume of wet extract (that is, loaded with water). On this basis, volume effects are eliminated and calibration

  remains linear. Difficulties arise when one wants to

  the case of samples with an unknown water content. In order to pass from the expression "mass of water per unit volume of wet extract" to the expression "mass of water in the sample", it is necessary to know the final volume of the wet extract. Well

  the amount of added anhydrous solvent is known, the contribution

  extraction of the extracted water to the volume is not.

  Both unit systems therefore have

  problems in relation to the variations of the volume. These problems have been addressed by using the second form of units to preserve the linearity of the calibration. For application to samples, the final volume of wet extract must be estimated, but an equation could be developed for this purpose. It is m = AV / (I-A) (2), where m is the mass of water extracted from the sample,

  A is the mass of water per unit volume of wet extract (i.e.

  say the result given by the calorimetric measurements) and

  V is the initial volume of anhydrous solvent used for the extraction.

  This equation gives the increase in volume that occurs when the anhydrous solvent dissolves the water in the sample. This correction assumes that the solvent and water volumes are additive. This is not strictly true, the final volume being less than

  sum of the two components due to solvation reactions.

  However, the mistake made by neglecting solvation

is trivial.

  Calibration data corresponding to a solution of 0.02M cobalt chloride (molar mass) are presented in the table below. These data were obtained from known masses of water, to which were added 100 ml of cobalt reagent in ethanol and an excess of anhydrous solid sodium chloride. The resulting mixture was stirred for 2 minutes in a top-macerator. Homogenization having been carried out, the calibration reflects the amplitude of all the errors coming from

  evaporation of the solvent or the acquisition of a moisture of origin

  atmospheric. Optical measurements were performed in a cell of a length of 10 mm maintained at 20 ± 0.1 ° C. in a conventional spectrophotometer. Ethanol was used as a reference solution to establish the optical data corresponding to 100%

of transmission.

BOARD

  Calibration for 0.02M cobalt chloride Mass of water, A (in grams per 100 milliliters of final wet extract) 2,508 2,568 2,638 2, 707 2,785 2,880 2,935 2,978 3,124 3,240 3,327 3,411 3,515 3,579 3,667 3, 752 3,846 3,932 4,027 4,108 4,144 4,202 Percentage of transmission 17.6 18.6, 0 22.1 22.6, 1 26.4 26.9, 9 32.6 34.5 36.3 37.8 39.7 41.2 42.8 44.9 47, 2 49.1, 5, 9 52.8 The: figures given in the table allow to establish the following equation corresponding to a curve of best adaptation (least squares method):

A = 0.0484T + 1.6635 (3),

  o A is the mass of water (expressed in grams per 100 milliliters of wet extract), and T is the percentage of transmission. The regression coefficient r is 0.9996 and the normal error of

  the estimation of the mass of water from the percentage of

  is 0.016. The relative error deriving from the measurement is therefore

  0.93% for a 95% confidence interval.

  The above data were used to plot the curve of Figure 4. The curve presented by this figure is shifted with respect to the true origin (the point 0.0), mainly

  because of the difficulty of getting a perfect solvent

  anhydrous. This is theoretically possible when the calibration graph starts from the true origin, but there are practical problems in relation to this way of proceeding. First, it is assumed that total absorption, or zero transmission, can be measured. Secondly, it is very difficult to maintain an anhydrous system in this state because of the great attraction of the solvent to the water. In addition, the lower the water content of the reaction mixture, the more sensitive the color.

to slight traces of water.

  It is therefore more convenient to accept a practical system that is slightly contaminated with water in an unavoidable way.

  This in no way introduces an error in the measurements, but only requires the calibration curve to be defined for each bath of solvent and colored reagent. This is not a difficult manipulation, since the calibration curve is a straight line. The verification of the calibration curve plays two additional useful roles, namely it allows to verify that

  was not accidentally wetting the solvent and correcting

  variations in cobalt salt concentration between

rents lots of reagent.

  There is no single combination of colored reagent and extraction conditions that can cover the entire range of 0 to 1007. of water for a fixed mass of sample while maintaining a relative error of 1 %. To illustrate this problem, the calibration graph of Figure 4 can be used. The linear region of this graph extends beyond what is shown to cover the range from less than 2.5 to more than of 5 g of water per 100 ml of wet extract. If these numbers represent the useful limits for this particular reaction mixture, the water content of the sample should be

  then be limited to 25 to 50 mg / ml of

  final result. Assuming a 1 + 1 dilution of the sample extract with the colored reagent, then the extract should contain

  between 50 and 100 mg of water per milliliter.

  Extraction requires an excess of solvent to ensure the quantitative extraction of water. Assuming that a sample solvent ratio of 5/1 is a

  reasonable value, then the amount of water in the sample must be

  tillon falls in the range of 250 to 500 mg / ml, or 25%. There is no difficulty in increasing the ratio of the extracting solvent to the sample and, therefore, it is possible to use an excess of solvent of the order of 10, that is to say 10 / 1 for the sample solvent ratio) to cover the range of 50 to 1007. of water in the sample. Water contents of 25 to 100% can

  therefore be treated by an appropriate choice of solvent-to-sample ratio

  lon. A basic modification is necessary to treat samples with a water content of less than 25%. Among the possibilities, it is possible to modify the ratio of the extract to the colored reagent or to use a more diluted and therefore more sensitive solution of cobalt. It is also possible to continue to use the same reagents, but by adjusting the colorimeter so as to allow accurate measurement of very low transmission coefficients. This can be done by means of a cell shorter than 10 mm, or by increasing the intensity of the light arriving on the cell, so that a sufficient amount

  radiation can be quantified accurately.

  With laboratory spectrophotometers, the intensity of illumination is virtually fixed, so the only possible option is to reduce the length of the cell. Cuvettes with a length of 1 mm can be easily used for the calorimetric measurement of low moisture food extracts. However, one can also remember the idea of increasing the intensity of the incident light, such an option

being described below.

  All the optical property measurements that were used to plot the curves of FIGS. 1 to 4 were made on a conventional laboratory spectrophotometer working in the wavelengths of visible light. The instrument was equipped with a heated cup holder maintaining the temperature at + 0.10C inside the cell compartment. We used

  for all measurements of cuvettes with a length of 10 mm.

  Such apparatus can be used to measure the intensity of color in a normal way, but this equipment is somewhat bulky. In particular, it takes 5 to 10 minutes to stabilize the temperature of the cuvettes after their introduction into the sample compartment. This is due to the fact that the contact

  thermal between the cell and the shirt is mediocre. A modification

  It is preferred to use a continuous flow cell of a sufficiently small volume to permit a faster realization of the temperature equilibrium. This equipment also prevents the operator from manipulating the cuvettes and, therefore, avoids

  the risks of errors due to the contamination of the optical faces.

  As noted above, the temperature of the solution under investigation is not critical as long as it is kept constant. The temperature of 20C mentioned above was chosen

  for convenience and may actually be higher or lower.

  An increase in temperature has the effect of increasing the inten-

  sity of absorption and thus move a calibration graph, such as that of Figure 4, to the right. However, there is,

  obviously, an upper limit to the increase in the

  which is the boiling point of ethanol. Similarly, the temperature of the measurements should not fall much below the extraction temperature, because other difficulties may result from a precipitation of the fats extracted at the same time.

that water.

  We will now describe the solution of

  use other equipment than a laboratory spectrophotometer

  tory. It is indeed possible to use, in place of the latter, a colorimeter specifically designed for the determination of the water content by means of a reagent colored with cobalt salt. This colorimeter has the necessary temperature control devices and is also less expensive than a spectrophotometer because the monochromator is here replaced by a red light-emitting diode light source whose emission peak has the length

655 nm wave.

  To allow application to the analysis of food products, a quantitative extract of the water present in the food product must be made, filtered and mixed in known proportions with the colored reagent before measurement in the colorimeter. The mixing can be carried out by the operator himself by means of an ordinary laboratory apparatus, or else

  by an automatic means of the type of injection or injection

segmentation.

  The functions of the colorimeter are as follows (1) accept a certain flow of ethanol containing water and the colored reagent; (2) adjust the temperature of the current to a constant level; (3) apply 655 nm wavelength peak light over a fixed length of solution; and (4) detect and measure

  transmitted fraction of the incident light.

  Figures 5Met 5B schematically show a colo-

  meter using a commercially available continuous flow cell. A metal block 1 has compartments 2, 3, 4, 5 and 6 for respectively containing a bowl, or cell, 7 (FIG. 6), a heating element (not shown), a thermostat (not shown), a support 8 of light source (Figure 7) and a detector support 9 (Figure 8). The temperature regulation of the sample is carried out via a tube 10

  spiral wound from the sample to cell 7

  A cylindrical extension 11 of the metal block 1. The inlet of the tube is designated by reference numeral 12, and a coating 13

  can be placed around the spiral to thermally isolate it.

  Cell 7 (FIG. 6) is of a commercially available type and can take different sizes depending on the volumes to be contained. The normal capacities are 750, 500 and 80fl, for a length of 10 mm in each case. A mask 14 is integrally associated with the cell, and the only other masks are therefore intended to ensure that the light source and the detector are isolated from the light. The optical face of the mask is designated by reference numeral 15, the cell comprising an inlet 16 for connection to the tube 10 and an outlet 17 for connection to a

  tube 18 to evacuate the solution.

  The light source 19 (FIG. 7) is an electrical diode

  red troluminescent terminated by an optical lens. It is possible to use almost any type of diode, but the one whose use is more especially considered is a relatively large indication diode able to withstand direct currents of 50 MA. The diode is mounted, for example using an epoxy adhesive, in the support 8 with a metal tube. A part of

  screwing 20 of the support 8 provides insulation vis-à-vis the light.

  Reference 21 designates a power supply means. The detector 22 (FIG. 8) is mounted in the same manner as the source

  light and is constituted by a phototransistor or a photodiode.

  Photogenerators are too large in size

  for the application in question, whereas photoresistances are too unstable to constitute quantitative detectors

  reliable. The complexity that is associated with the use of photomultiplied

* Cateurs does not guarantee its use. The support 9 for detecting

  is also a metal tube with a part of

  screwing 23 allowing the attachment to the block 1 in an insulating manner vis-

  with respect to light and a power source 24.

  The tube 10 can be made of plastic, although

  it is possible to use a metal providing better thermal contact

  if it can be shown that there is no reaction

  between the metal tube and the reagent. The necessary length of tube 10 is determined by the chosen flow rate, the yield of the thermal contact between tube and block, and the dimensions chosen.

for the tube.

  Most light-emitting diodes have a voltage drop of about 1.4 V and operate at currents up to 50 m. Light emission is proportional

  direct current. For fixed illumination, the power supply

  It must produce a stable current within the operating characteristics of the diode and an appropriate voltage. To achieve a variable illumination (as it has

  discussed above), it is necessary to have a food

  switchable current circuit whose output currents are

  multiples exact of the lowest current level.

  As shown in Figures 9 and 10, the cir-

  The detector is designed according to this example in the form of a bridge circuit, which comprises a power source applying between the points A, A, two useful resistances.

  and 51, two compensation resistors 199 and 201 and a resistor

  200. The balance of the bridge is measured between points B, B. The two so-called "useful" resistors 50 and 51

  are respectively the detector associated with the bowl and a detection

  reference. The detector 50 associated with the sample is illuminated by the light of a diode 61 fed by a source 71, the light being transmitted through the cuvette. The reference detector 51 is illuminated only by a second diode 61, powered

  by a source 71 independent of the source 70. The source of food

  71 of the reference light-emitting diode is a

  completely variable current source, so that the incoming light

  on the reference detector can be set between darkness and full intensity. The direct current applied to the reference diode is measured by a diode current control device 81 which supplies its signal to a reading device 91.

  assembly is completed by a means 150 for detecting the

  balance of the bridge. This bridge balance detector 150 may be a microamp meter, but it will preferably be an amplifier connected to the power source 71 of the reference diode,

  way to maintain the balance of the bridge.

  In order to avoid errors relating to temperature differences, it is preferable to mount the diode 61 and the reference detector 51 in the metal block which houses them.

  sample measuring devices. It is for reasons of simplicity

  plification that such an assembly has not been shown in the figures. To use the colorimeter, one begins by stabilizing the temperature of the metal block. By pumping, we make arrive

  in the cell a solvent that has no reagent, that is,

  to say perfectly colorless. We then adjust the resistance 200 of equi-

  liberation to obtain equal currents in the diode 60 of the sample.

  tillon and the reference diode 61. Such a situation corresponds to a 100% optical transmission. When the water-ethanol-colored reagent mixture is introduced into the cuvette, there is a decrease in the amplitude of light reaching the detector of the sample, if

  although the bridge circuit is out of balance. To bring back the equi-

  free of the bridge, it is necessary to reduce the current delivered by the source 71 of the reference diode. When the bridge has returned to equilibrium, the new direct current of the reference diode is measured. The decrease of the current of the reference diode is proportional to the percentage of transmission through the sample. The latter quantity is evenly proportional to the water content

  existing in the solution under test.

  The assembly described above assumes a fixed current for the diode associated with the sample. In other words, a constant light intensity arrives on the bowl. For it to be

  possible to measure on extremely absorbent solutions.

  so that the range of water contents to be determined can be extended, it may be useful to increase the intensity of the light source associated with the sample. In this way, a larger amount of light arrives on the cell, so that it transmits a quantity of light that is easier

to measure.

  Unlike the case of conventional spectrophotometers, it is easy to vary the incident light by modifying the direct current of the diode used to illuminate the cell containing the sample. It may be useful to be able to increase the intensity of the source by degrees. The reading scale of the apparatus 91 is then multiplied by a factor accounting for the new intensity of the source, so that, thus,

  we can accurately measure weak transmissions.

  The advantages of the assembly described above are that it always operates in equilibrium state of the bridge and contains a reference element which is identical to that used for measuring the sample. The circuit therefore compensates for itself the non-linearities relating to the source or the detector. In addition, the diode relative to the sample is kept at a constant temperature and, since it is a semiconductor, it does not age in the same way as the incandescent light sources. For these reasons, the intensity of the source does not derive, as far as

  o The power supply is stable enough.

  According to another embodiment, it can be done in that the source relative to the sample also illuminates the reference detector. Such a double-beam scheme is intended to provide compensation for any variation-in the intensity of the

  source, but it eliminates any possibility of varying

  the amount of light arriving on the cell of the sample

  lon. This is because the system can not distinguish between accidental variations and desired variations in

the light emission of the source.

  The description given above relates to

  operation of a diode light source excited by a direct current. However, it is possible to use, for the supply of the source, a pulsed or modulated current, or to cut the light beam. Adequate arrangements for ordering the reference source in a similar way and

  to maintain synchronization between the two sources. The main

  The advantage of such a system is that it is then possible to use.

  AC amplifiers for processing signals from the detectors or the bridge. When these amplifiers are designed to respond specifically to the modulation frequency,

  the equipment is then protected against the effects of light.

  parasite on the detectors. In addition, the use of a pulsed electrical source has the effect of lowering the duty cycle of the diodes. This then makes it possible to excite the diodes at higher direct currents, so that the intensity of the current is increased.

emitted light.

  With regard to the processes described, a program for determining the water content of a food product containing or suspected of containing salt (chloride

  of sodium) as follows, in connection with Figure 11.

  A given weight of sample is prepared, to which a fixed volume of solvent (ethanol or denatured alcohol) is added, followed by excess anhydrous sodium chloride, after which the mixture is mixed at high speed in order to homogenize. Using a filter paper based on glass fibers, the previously obtained dough, which contains solids in a continuous liquid phase of water in solution in ethanol, saturated with salt, is filtered, after which adds to the transparent liquid extract obtained a fixed proportion of colored reagent. Then, the percentage of transmission in the constant temperature colorimeter is measured and the calculation of

  the water content of the extract on the basis of the standard equation

  swimming, ie equation (2) or (3).

  As alternatives, it is noted that other suitable reagents can be used for quantitative colorimetric determination of the moisture content of materials. These include CuCl2 (copper dichloride) in anhydrous acetone, for which the measurement of the absorption coefficient is

  made at 473 nm, and VC13 (vanadium trichloride) in ethanol.

  anhydrous carbon, the measurement of the absorption coefficient being then

  at 432 nm. There is an antilogarithmic relation between the coefficient

  absorbency and the "percentage" (in mass by volume) of moisture

  at 473 nm for a percentage (by mass per volume) of 0.08

  CuCl2 in the range of 0 to 2% moisture. And there is a relationship

  logarithmic relationship between the absorption coefficient and the

  humidity (mass by volume) at 432 nm for a percentage (by mass per volume) of 0.1 of VCl 3 in the range of 0 to 10

(mass by volume) of moisture.

  Of course, those skilled in the art will be able to

  giner, from the process and the device whose description comes

  to be given as merely illustrative and non-limiting, various other variants and modifications not out of the box

of the invention.

Claims (9)

R E V E N D I C A T I 0 N S   1 - Method for measuring the moisture or water content of materials, in particular foodstuffs, according to which   the material is subjected to the action of a hydrophilic water extracting agent   organic phase in liquid form and the water content of the extract is determined, the process being characterized by adding to the extract a color reagent consisting of a transition metal salt including the metal ion has the aptitude, vis-à-vis the anions   of the extract, to typically absorb a first half   wavelength of electromagnetic radiation and, vis-à-vis   of a solvent, such as the extractant and water, bound to a complex, to typically absorb a second wavelength of the electromagnetic radiation, an increase in the concentration of solvent relative to the concentration of anions in the solvent causing a reduction in the absorption of the first wavelength, in that it illuminates in transmission by means of the first wavelength of the electromagnetic radiation, while maintaining it at a predetermined temperature substantially a sample of the extract to which color reagent has been added, and measuring an optical amplitude of the sample related to the absorption of the first wavelength by the sample and   this optical amplitude is linked to predetermined data   a correlation between different water contents of the agent   extraction and said optical amplitude.   2 - Process according to claim 1, characterized in that, as optical amplitude, the absorption coefficient, which takes place in a function which links the logarithm of the coefficient, is chosen as the optical amplitude.   absorption and the water content of the extract.   3 - Process according to claim 1 or 2, wherein the material contains anions which, with water, can be extracted from the material when it is subjected to the action of the agent   extraction, the method being characterized in that, after the   extraction, the extract is saturated by means of said anions and in that one chooses, for said optical amplitude, the percentage of transmission.   4 - Process according to claim 1, 2 or 3, characterized   in that a salt of cobalt is selected as the transition metal salt and in that 655 nm is selected as the first wavelength. A process according to claim 3 or 4 for measuring the moisture or water content of a food product which contains or is suspected to contain chloride, the process   being characterized in that one selects, as a transition metal salt,   cobalt chloride, the extract being saturated with chloride ions,   for example by adding sodium chloride to the extract.   6 - Process according to any one of the claims   1 to 5, characterized in that the mass of water extracted from the material is calculated from the equation: AV m 1 -A 'om is the mass of water extracted from the material, A is the mass of water per extract volume unit, and V is the extractant volume of anhydrous water added to the material.   7 - Process according to any one of the claims   1 to 6, characterized in that an excess of extractive agent is used.   preferably, ethanol, several times amount of water in the material.   8 - Apparatus for implementing the   method defined in claims 3 to 7, characterized in that   includes a heated block (1) provided with means for maintaining   at a predetermined temperature a compartment (2) designed to receive   see a cell (7) containing a sample, an electronic diode   red luminescent light (19) which illuminates the sample by means of 655 nm wavelength light, and an electrical detection means (22) designed to measure the percentage of transmission through   the sample at said wavelength.   9 - Apparatus according to claim 8, characterized in that it comprises means for adjusting the direct current applied to the diode.   Apparatus according to claim 8 or 9, characterized   in that a means is provided by which the light sources produce a modulated light, the detection circuit being designed to respond only to a correspondingly modulated light.