WO1996024830A1

WO1996024830A1 - Method for determining the particle size of a material

Info

Publication number: WO1996024830A1
Application number: PCT/DK1996/000067
Authority: WO
Inventors: Holm Schwarze; Hilmer Jensen; Freddy Petersen; Claus Borggaard; Jens Havn Thorup
Original assignee: Wolfking Danmark AS
Current assignee: GEA Food Solutions Denmark AS
Priority date: 1995-02-10
Filing date: 1996-02-09
Publication date: 1996-08-15
Anticipated expiration: 1997-08-10
Also published as: DK9196A; AU4619596A; DK171927B1

Abstract

In a method for determining the particle size of a material, the transmittance or absorbance of a sample of the material being placed in a space or region is measured at a number of wavelengths in the near-infrared range. The particle size is determined by comparing the set of measurement values having been obtained with corresponding data sets for a material with a known particle size or by insertion in an algorithm having been produced on the basis of data sets for a material with a known particle size.

Description

24830 --

1 METHOD FOR DETERMINING THE PARTICLE SIZE OF A MATERIAL

The present invention relates to a method for determining the particle size of a material.

It is known to determine the content of various components of material by measuring its absorbance or transmittance of light in the infra-red range. The expression "near- -infra-red spectroscopy" is used for measuring methods based upon the interaction between matter and electromag¬ netic radiation in the wavelength range from 700 to 2500 nm. The reason for using this expression is that the wavelength range used is that part of the infra-red range lying closest to the visible spectral range of 400 to 700 nm. In the literature, the expression "near-near- infra-red range" is used for electromagnetic radiation with wavelengths from 700 to 1200 nm.

Meat consists mainly of water, protein and fat. Each type of chemical bond, e.g. O-H, C-H, C=0, C-N, N-H, absorbs light at wavelengths being characteristic of the particular molecule part. The cause of the absorption is that two different atoms being bonded to each other function in the manner of an electrical dipole taking energy from the electric and magnetic fields in the radia¬ tion, making the group of atoms vibrate. A C=0 bond in a triglyceride absorbs light at a wavelength, that differs from the wavelength of light absorbed by a C=0 bond in a protein molecule. By measuring how much the light is attenuated by passing through a meat sample at one of these characteristic wavelengths, it is possible to deter¬ mine the percentage of a component of the meat. Similarly, other components of other materials can be determined, when the components differ with regard to characteristic wavelengths.

It has now surprisingly been found that it is also pos- sible to determine the grain size or particle size of a material by means of light in the near-infra-red range. The method of making this determination is extremely simple as compared to methods having been used earlier, and if desired, it may be carried out in connection with determining the content of components of the material. Accordingly, the method according to the invention is characterized in that the transmittance or absorbance of a sample of the material placed in a space or region is measured at a number of wavelengths in the near-infra-red region, and that the particle size is determined by com¬ paring the sets of measurement values having been obtained with corresponding data sets for a material with known particle size or by inserting the set of measurement values having been obtained in an algorithm provided on the basis of data sets for a material with a known par¬ ticle size.

The method according to the invention may e.g. be used for determining the grain size or particle size of a material in a mixing tank or tub or for determining wheth¬ er two raw materials with different grain sizes have been mixed sufficiently thoroughly. The method may also be used for controlling a particle-forming process for obtaining particles of a desired size, e.g. for control- ling the comminution of a material or of a process com¬ prising the build-up or adjustment of particles.

The particle size is i.a. important for finding the repre¬ sentative quantity necessary for use when determining the percentage of a component, e.g. fat, in the material, as the quantity may vary with the particle size. When the particle size is known, it is possible to compute the representative quantity, and on the basis of this value it is possible to determine the size of the sample quan¬ tity to be used, or how many times it is necessary to repeat a sampling and measuring cycle, if only small quantities are examined at a time. In this manner it is possible to ensure that the measuring procedure for deter- mining a component is carried out with the least possible sample quantity or in the shortest possible time with a given accuracy, also in those cases in which the particle size of the material can vary considerably.

The method according to the present invention may be used on material from a tank or tub, e.g. a mixing tank or mixing tub, one or a number of samples being taken for examination. The method may also be used on material being conveyed in a conduit. It is e.g. possible to insert a tube segment with measuring equipment in the conduit, so that the material passing the measuring equipment may be examined directly. Another possibility consists in mounting a branch tube or a tube loop with corresponding measuring equipment on the conduit, so that only a part of the material stream in the conduit is examined. The method according to the invention may also be used on material coming from or being present in other equipment, e.g. screening equipment, sorting equipment (e.g. clas¬ sifying equipment) , sedimentation tanks, spray-drying plants, pelletizing equipment and other particle-forming equipment.

An embodiment of the method comprises that the measure¬ ments are repeated on a fully or partly new quantity of the material being placed in said space or region, and that the value having been obtained for the particle size or the deviation of the measurement from a pre-de- termined value is used for continuous control or onitor- ing a particle-producing equipment, e.g. a machine for finely mincing or grinding the material or forming it into particles. In this manner, the method can be used for dynamic control of various processes.

Measurements are preferably carried out at near-infra¬ red transmittance (NIT) . Measurements may be carried out at a number of wavelengths in the near-near-infra-red range between 700 and 1200 nm.

The measurements may be carried out at various wavelengths using a rotatable filter disc placed in the beam path between a light source and a light receiver, said filter disc having filters, each allowing a respective wavelength interval to pass, placed around the shaft at equal radial distances, enabling one filter at a time to be brought into the beam path by means of a motor connected to the shaft of the filter disc.

Especially the measurements may be carried out at various wavelengths by means of a number of narrow-spectrum light sources, each emitting light in a respective wavelength interval.

Highly preferred the measurements are carried out by means of 4-20 monochromatic light sources in the form of laser diodes, each emitting light in a respective wavelength interval within the range 700-1200 nm.

The measurements may be carried out with the material at rest, so that it is also possible to use measuring methods requiring relatively long measuring times.

Material that contains air or gas which is deformable may be compressed before carrying out measurements, pre¬ ferably to a pressure of between 200 and 2000 kPa (2 and 20 bars) . Hereby any bubbles of air or gas in the material will be compressed or dissolved so as normally to improve the measuring accuracy and simplify the measurements.

It is preferred that the measurement values or sets of same having been recorded are also used for determining the content of one or a number of components of the ma¬ terial.

The method according to the present invention is especi¬ ally useful within the food industry, the pharmaceutical industry and the technical-chemical industry for deter¬ mining the grain size of materials occurring in these areas of industry. The method is preferably used on food¬ stuffs, fodder and pharmaceutical material. In addition to determining the particle size, the infra-red measuring equipment may, as mentioned above, also be used for deter¬ mining the content of various components of material; examples of such components are indicated in brackets in the next paragraph.

Amongst the material, the following may be mentioned:

- vegetable foodstuffs, such as wheat, barley, rye, maize, rice, coffee and cocoa in the form of whole grains or ground product (analysis for protein, starch, carbohydrates and/or water) , seeds, e.g. peas and beans, such as soybeans (analysis for protein, fats and/or water) , products mainly consisting of or ex¬ tracted from vegetable raw materials, such as snacks, dough, vegetable mixtures, margarine, eatable oils, fibre products, chocolate, sugar, syrup, lozenges and dried coffee extracts, in the form of powder or granu¬ late,

- animal foodstuffs, such as dairy produce, e.g. milk, yoghurt and other soured milk products, ice cream, cheese (analysis for protein, carbohydrate, lactose, fat and/or water) , meat products, e.g. meat from pork, beef, mutton, poultry and fish in the form of minced or emulgated products (analysis for protein, fat, water and/or salts) and shellfish as well as eggs, which foodstuffs, may be present in a fully or partly frozen condition,

- fodder, e.g. pellets or dry/wet fodder mixtures of vegetable products, fats and protein-containing raw materials, including pet food,

- pharmaceutical products, such as tablets, mixtures, creams and ointments,

- technical substances, e.g. wet and dry mixtures of cement and mortar, plastics, such as plastic granulate, mineral materials, such as solvents and petrochemicals, such as oils, hydrocarbons and asphalt, or suspen- sions/solutions of organic or inorganic substances, e.g. sugar solutions.

When carrying out the method according to the present invention, the conditions for examination being most appropriate for the type of material in hand are pre¬ ferably used.

Measurements in the near-infra-red range may be carried out in two ways, either by transmitting light through the sample (near-infra-red transmission, NIT) or on the basis of the reflection from the surface of the sample (near-infra-red reflection, NIR) . In samples with a high content of water, such as meat, NIT cannot be used for measurements at wavelengths above 1300 nm, because the absorption in the water molecules is far too strong at longer wavelengths. If an Si-detector is used, it is only possible to operate at shorter wavelengths than approximately 1050 nm, as this type of detector is insen- sitive at longer wavelengths.

Reflection measurements have the disadvantage of requiring the measurements to be carried out through a glass window (ordinary glass allows little near-infra-red light to pass) . In this case, fat from the meat having been com¬ minuted will unavoidably adhere to the inside of the glass window, possibly resulting in erroneous measurement. Further, because of the small measuring volume, a reflec¬ tion measurement will not be so representative as an NIT-measurement.

For this reason, optimum measuring conditions for meat and similar products likely to form deposits on the glass windows are achieved by carrying out near-infra-red meas- urements based on transmission with a physical path length in the measuring tube of e.g. 40-60 mm. The sample should preferably remain at rest during the fraction of a second needed to carry out the measurement. As far as possible, the sample should be free from air pockets; this may be achieved by compressing it.

For detecting the near-infra-red radiation, the following materials can be used: - Si: A very sensitive and cheap type of detector useful in the range from 400 to 1100 nm.

- Ge: Hardly as sensitive as Si, but is useful from 800 to 1800 nm. - InGaAs: Only half as sensitive as Si, but reacts very quickly and is useful from 800 to 1760 nm.

- PbS: Low sensitivity, but is cheap and is useful from 650 to 3000 nm. Temperature stabilization is required. - PMT (Photo Multiplier Tube) : This is by far the most sensitive type of detector.

Measurements on natural products have shown that there is no linear correlation between the absorption of light and the percentage of a compound in the sample. The ab- sorbance is not only due to the presence of absorbing compounds in the sample, but is also influenced by the dispersion of light in the sample. Further, account must be taken of the fact that the composition of natural products is so complicated, that absorptions due to dif¬ ferent compounds/functional groups overlap each other in the spectrum. For this reason, it is necessary in this connection to use more complicated mathematical models, e.g. neural networks or classical statistical methods, for determining the content of a compound in the sample.

Tests with minced or comminuted meat raw materials on an NIT-analysis instrument have shown that the three main components fat, water and protein can be determined, even in the situation in which they do not add up to 100% because of the presence of other additives or because of the material having varying particle size.

A special use of the method according to the invention consists in that the measurement values of sets of same having been recorded are used to ascertain whether or when a quantity of material is sufficiently homogeneous with respect to particle size and/or content of compo- nents. Hereby it is e.g. possible to avoid problems of so-called "over-mixing" when producing a homogeneous mixture.

Especially preferred is an embodiment comprising that the transmittance or absorbance of a sample of the mate¬ rial, e.g. a minced or comminuted meat product with an average particle size of between 2 and 30 mm, is measured at a number of wavelengths in the near-infra-red range, preferably between 700 and 1200 mm, and that the meas- urement values or sets of same having been recorded are used to determine the particle size, and further to determine the content of one or a number of com¬ ponents of the material, in a meat product e.g. pref¬ erably its content of fat, protein, collagen, and/or water, and/or

- to determine whether or when a quantity of material is sufficiently homogeneous with respect to particle size and/or content of components.

In this way it is possible to control or determine not less than three different parameters of substantial im¬ portance for a satisfactory mixing process.

The transmittance or absorbance is preferably determined using an examination unit having a tube with an opening for receiving samples of material from a tank, a tub, a conduit or other equipment containing the material, and with a tube segment adapted for measurements on material having been received; and a measuring device placed ad- jacent said tube segment and comprising a light source on one side of the tube segment and a light receiver on the opposite side, the walls of said tube segment in the beam path between the light source and the light receiver being formed of a material being translucent for the radiation at the wavelength range to be examined.

In the present description, the expressions "particles" and "grains" are used for elements having a size of l mm or more, especially 3 mm or more, the material especially being present in its natural form, e.g. as a natural product or as a material only having been subjected to coarse comminution, i.e. not in a very finely divided or homogenized form.

In the following detailed portion of the present descrip¬ tion, the invention will be explained in more detail with reference to the drawings, in which Figures la and lb show an apparatus for automatic taking of samples and measuring their grain size and content of components using NIT,

Figure 2 is a graph showing the transmittance at various discrete wavelengths of meat samples containing much and little fat, respectively, as measured in the apparatus, Figure 3 shows part of a different embodiment of the measuring equipment in the apparatus of Figure 1, Figure 4 shows NIT-spectra of meat with various degrees of comminution and fat content, and Figure 5 is a classification diagram obtained by principal component analysis of average spectra for meat samples with different degrees of comminution and fat content.

The apparatus of Figure la comprises a tube 10 composed of three tube segments 10a, 10b and 10c. The tube 10 is mounted on a component containing or conducting the mate¬ rial to be examined, e.g. a tank or tub, the opening at one end of the tube being adapted to receive material and the opening at the other end to return material having been examined to the tank or tub.

The lower tube segment 10a comprises a cylinder 14 with two pistons, of which the upper one is provided with a short tube 17 adapted to slide within the vertical part of the tube segment 10a, while the lower piston carries a plunger 18, the outside diameter of which corresponds to the inside diameter of the short tube 17, so that the plunger slides within the short tube. The short tube 17 and the plunger 18 can be made to move mutually indepen- dently by means of compressed air admitted to the cylin¬ der.

The upper tube segment 10c comprises a cylinder 27 with a piston. On one side, this piston carries a plunger 29 adapted to slide within the tube segment 10c and to be moved by means of compressed air being admitted to the cylinder.

The segments 10a and 10c serve to convey the material received from the tank or tub into the segment 10b and compress it to enable an NIT-measure ent to be carried out. Figure lb shows the apparatus in the measuring con¬ dition. The functions of the segments 10a and 10c are described in detail in the description belonging to the Danish patent application No. 0155/95.

The apparatus of Figure 1 can also be used without the parts 10a and 10c, if the sample is introduced in the segment 10b by some other means. The segment 10b may e.g. be mounted directly on a tank or a conduit containing the material. Preferably, pumps or similar arrangements are provided to convey the material into the segment 10b.

The tube segment 10b serves as a measuring chamber when measuring the transparency of the material to infra-red light of various wavelengths. To this end, the segment 10b comprises two windows 24 of glass or other transparent material inserted in cut-outs facing each other.

On the tube segment 10b a housing 25 is mounted, said housing containing various devices for measuring the transparency of the material being present between the windows 24. A wide-spectrum light source 32 emits light within the operating range, in the present case the near- infra-red range between 700 and 1200 nm. The light source 32 comprises a tungsten/halogen lamp capable of emitting a large proportion of the electrical energy being supplied in the infra-red range of the spectrum. The power rating is between 20 and 70 W, but may also be larger, e.g. 100 W.

Adjacent the light source 32 there is a preferably ellip¬ tical or parabolic reflector 33, so that the light is mainly directed towards the right. A rotatable filter disc with 6-20, e.g. 12, different filters 35, each al¬ lowing passage of light at a respective wavelength through the windows in the tube segment 10b, is placed between the light source 32 and the windows 24 in the tube segment 10b. The monochromatic light entering through the left- hand window 24 passes through the material in the tube suffering a substantial loss and exits through the right- hand window 24, after which it impinges on a wide-spectrum photo-detector 36, e.g. a plate composed of a number of Si-wafers.

The attenuation of the light in the material is due to the absorption caused by the various components in the material, as well as the dispersion and the reflection of the light caused by phase transitions or particles in the material.

The photo-detector 36 produces signals depending i.a. on the material, its particle size and content of components. The signals are amplified, filtered, digitized and stored in an electronic memory. The windows 24 and the path of the light rays are dimensioned in such a manner that the detector 36 receives light having passed through a volume of material of more than 100 ml. The volume of material corresponds to the volume of the space between the windows 24.

For determining the absorption at various wavelengths, the measuring equipment comprises a motor 37 for rotating the filter disc 34, so that the filters 35 are brought one by one into the beam path between the light source 32 and the detector 36. Each time a new filter has been placed in the measuring position, the signal from the detector 36 is recorded and stored, the strength of this signal depending on the absorption of the material con¬ cerned in the wavelength interval of the particular fil¬ ter. When measurement values have been recorded and stored for all the filters in the disc 34, the measuring process is complete. Measurements may be carried out while the material is at rest or moving.

Figure 2 shows the signal from the detector 36 during one revolution of the filter disc 34. The fully drawn curve represents a finely comminuted sample of pork with a fat content of approx. 50%. The lightly drawn curve represents a finely comminuted sample of beef with approx. 5% fat. The samples attenuate the light approx. 4000 times. The peak values represent the transmittance at the 11 different wavelengths. It will be seen that the samples attenuate the light differently at the various wavelengths because of the differences in their fat and water content; this fact may be used for computing these values.

If the samples have different sizes of grains or par¬ ticles, other values for the transmittance will be ob¬ tained, and according to the present invention, this fact is used to determine the particle size of a sample being examined, as the data having been obtained are analysed and the particle size determined by comparison with data from samples with known particle sizes.

In this manner, it is possible to use the apparatus for continuous monitoring of a mixing operation so as to ascertain whether or when the mixture is sufficiently uniform with regard to particle size. Further, it can automatically compute the content of e.g. fat in the material, the apparatus containing a program with the necessary computing routines. When proceeding as described above with samples of 60-400 ml, a single measurement result is not sufficiently certain when measuring on coarsely comminuted meat, for which reason it is necessary to examine a number of samples, e.g. 10 samples, before it is possible to compute a sufficiently reliable number for the fat content on the basis of the sum total of the measurement values or results. The measurements on the particle sizes may also be used for controlling a comminuting process.

All computations and adjustments may be carried out auto- matically by a computing unit on the basis of the measure¬ ment data having been received.

Instead of a wide-spectrum light source with a filter disc in front, it is possible to use narrow-spectrum, discrete light sources, each emitting light at a respec¬ tive wavelength. Such an embodiment is shown in Figure 3 and is based on the use of laser diodes instead of the lamp and the filter disc. The arrangement of Figure 3 possesses the advantage of having no moving parts.

The arrangement comprises a number of (power) laser diodes 40, each emitting light at a certain respective wavelength against the sample of material. Typically, 4-20 diodes are used, all placed on the same chip. Each laser diode emits light at a respective wavelength within the range from 800 to 1050 nm, so that the use of filters is un¬ necessary. A PMT-detector 41 is used for sensing the light having passed through a sample of thickness 5-10 cm. By activating one of the diodes at a time, the detec- tor 41 is used to measure how much light penetrates through the sample at the various wavelengths.

Example

NIT-measurements on pork with varying fat content and degree of comminution

The tests are to clarify whether it is possible to deter¬ mine the particle size in materials with different con¬ tents of fat, and whether it is necessary to comminute the material completely before measuring the fat content.

From a slaughterhouse, 2.5 kg of each of the following products are procured: Foreloin (fat content 9%) , neck (fat content 24%) and shoulder cuts (fat content 45%) . By using a cutter, these samples are comminuted down to 13 mm.

From each sample 400 g are taken and measured on an NIT- analysis instrument for laboratory use (model Infratec 1255 from Tecator) . The NIT-instrument comprises 5 cups with sample material. The sum of the effective sample volumes of the five cups is 25 ml. Then, the material having been taken out is returned to the samples, which are then comminuted to 10 mm. From each sample type, part-samples of 400 g are taken and measured on the NIT- instrument. The procedure is repeated to provide measure¬ ments on meat material of 8, 5 and 3 mm. After this, the samples are analysed for fat content using traditional laboratory analysis.

Figure 4 shows two average spectra for shoulder cuts and neck comminuted down to 13 mm, as well as for two previ¬ ously produced model spectra for meat of 3 mm with cor- responding fat content.

It can be seen from the Figure, that spectra of materials with the same percentage of fat but different particle size have the same appearance; there is, however, a paral- lei displacement along the Y-axis, so that the material having the largest particles also has the highest Y-value. With materials with varying content of components, the curve shape varies. This is the reason why it is necessary to measure the absorbance or transmittance at a number of wavelengths in the near-infra-red range when the par¬ ticle size is to be determined in the materials occurring in actual practice, so that the variation in the component content of the material can be eliminated. If only the content of one single component is to be determined, the absolute Y-values may be ignored, as it is only the rela¬ tive changes at the various wavelengths that are to be used.

A principal component analysis (PCA) of the average spec¬ tra for each of the three meat products and each of the five degrees of comminution is carried out, making the analysis cover a total of 15 spectra. The result of this classification is shown in Figure 5.

It will be seen that the spectra are grouped in two direc¬ tions: one direction indicating the fat content of the sample, and another direction at right angles to the first, indicating the degree of comminution. Thus, this Figure shows that it is possible to determine the degree of comminution, and that a complete comminution is not necessary for determing the content of substances in the material by means of NIT-analysis. It is e.g. impor¬ tant to know the particle size when the sample volume being representative for the determination of components of the material is to be found. A computer simulation shows that the RSD in the determination of fat in meat comminuted down to cubes of 10 x 10 x 10 mm in a sample of 5 kg is just below the error occurring when determining the fat in the laboratory. When comminuting to cubes of 20 x 20 x 20 mm it is necessary to take a sample of sub¬ stantially more than 10 kg to make the error fall below the laboratory error. If the measurements are carried out on 125 ml of the material at a time, it will on these conditions be neces¬ sary to repeat the measurements on 40 samples of cubes 10 x 10 x 10 mm to obtain a representative measurement of the fat content.

Claims

C L A I M S:

1. Method for determining the particle size of a mate¬ rial, c h a r a c t e r i z e d in that the trans it- tance or absorbance of a sample of the material placed in a space or region is measured at a number of wave¬ lengths in the near-infra-red region, and that the par¬ ticle size is determined by comparing the sets of measu¬ rement values having been obtained with corresponding data sets for a material with known particle size or by inserting the set of measurement values having been ob¬ tained in an algorithm provided on the basis of data sets for a material with a known particle size.

2. Method according to claim 1, c h a r a c t e r¬ i z e d in that the measurements are repeated on a fully or partly new quantity of the material being placed in said space or region, and that the value having been obtained for the particle size or the deviation of the measurement from a pre-determined value is used for con¬ tinuous control or monitoring a particle-producing equip¬ ment.

3. Method according to claim 1, c h a r a c t e r- i z e d in that measurements are carried out at various wavelengths using a rotatable filter disc placed in the beam path between a light source and a light receiver, said filter disc having filters, each allowing a respec¬ tive wavelength interval to pass, placed around the shaft at equal radial distances, enabling one filter at a time to be brought into the beam path by means of a motor connected to the shaft of the filter disc.

4. Method according to claim 1, c h a r a c t e r- i z e d in that measurements are carried out at various wavelengths by means of a number of narrow-spectrum light sources, each emitting light in a respective wavelength interval.

5. Method according to claim 4, c h a r a c t e r¬ i z e d in that measurements are carried out by means of 4-20 monochromatic light sources in the form of laser diodes, each emitting light in a respective wavelength interval within the range 700-1200 nm.

6. Method according to claim 1, c h a r a c t e ¬ i z e in that material that contains air or/and gas which is deformable is compressed before carrying out measurements, preferably to a pressure of between 200 and 2000 kPa (2 and 20 bars) .

7. Method according to claim 1, c h a r a c t e r¬ i z e d in that the measurement values or sets of same having been recorded are also used for determining the content of one or a number of components of the material.

8. Method according to claim 1, c h a r a c t e r¬ i z e d in that the measurement values of sets of same having been recorded are used to ascertain whether or when a quantity of material is sufficiently homogeneous with respect to particle size and/or content of compo¬ nents.

9. Method according to claim 1, c h a r a c t e r¬ i z e d in that the transmittance or absorbance of a sample of the material, e.g. a minced or comminuted meat product with an average particle size of between 2 and 30 mm, is measured at a number of wavelengths in the near-infra-red range, preferably between 700 and 1200 nm, and that the measurement values or sets of same having been recorded are used to determine the particle size, and further - to determine the content of one or a number of com¬ ponents of the material, in a meat product e.g. prefer¬ ably its content of fat, protein, collagen, and/or water, and/or to determine whether or when a quantity of material is sufficiently homogeneous with respect to particle size and/or content of components.

10. Method according to claim 1, c h a r a c t e r¬ i z e d in that the.transmittance or absorbance is deter- mined using an examination unit having a tube with an opening for receiving samples of material from a tank, a tub, a conduit or other equipment containing the material, and with a tube segment adapted for measurements on ma¬ terial having been received; and a measuring device placed adjacent said tube segment and comprising a light source on one side of the tube segment and a light receiver on the opposite side, the walls of said tube segment in the beam path between the light source and the light receiver being formed of a material being translucent for the radiation at the wavelength range to be examined.