PHYSICOCHEMICAL MEASUREMENTS
THE ROLE OF STANDARD SPECIMENS FOR SUBSTANCES AND MATERIALS IN UNIFYING COMPOSITION MEASUREMENTS UDC 006.065:53.089.68
N. G. Semenko
A central problem in metrology is providing unified measurements, and this affects all other forms of metrological activity, including major areas such as reproducing and transmitting units, carrying out state tests, and checking means of measurement during use. In the existing terminology, unification in measurements is a state in which measurement results are expressed in legal units and the errors of measurement are known with a given accuracy [i]. This brief definition conceals a vast and complicated meaning, whose essence is that all the results from the set of measurements should be accompanied by reproducible dimensions for the physical quantities and by a statement of the accuracy. There are two major aspects to unifying measurements: firstly, providing unified means of measurement, and secondly, estimating the errors in measurement results. The first aspect is covered by the extensive and numerous activities of metrological bodies concerned with reproducing units, state tests on means of measurement, test operations in accordance with test schemes, and metrological certification of unstandardized means of measurement. However, although it is necessary to use tested means of measurement, in most cases it is not a sufficient condition for providing unified measurements, since only in certain very simple cases can one identify the errors in the results with the errors in the means of measurement used. Examples are measuring mass by means of a balance, measuring a voltage in a power circuit with a voltmeter, etc. In most cases, however, one needs special methods of processing and evaluating measurements together with some practical techniques for checking reliability. This applies particularly to measurements in the technical category under the classification of [2, 3], which constitute the main volume of measurements in all areas of practical activity and which are characterized by a wide range of measured quantities. For a long time, we did not have the necessary methods and standards to meet the second condition. According to [4], the technical standardization measures in providing unified measurements include standards for measurement accuracy, methods of making measurements, methods of expressing results, forms of their representation, and accuracy parameters. However, the corresponding national standards began to be devised only quite recently; the first of them were confirmed only in 1972 [5, 6]. We also note that there is a deficiency of specifications and concepts dealing with the practical evaluation and checking for the unity of measurements not only in systems for providing unity in measurements in accordance with [4] but also in existing scientific and technical documentation. These aspects are of primary importance in measurements termination.
concerned with composition de-
In this area, one uses particularly complicated indirect methods, in which the predominant part is not played by the errors of measurement in direct methods for unspecific quantities (mass, volume, optical density, etc.) but instead they only play a secondary part in the overall error. Therefore, unifying the means of measurement is traditionally not considered as the most important factor in providing unified measurements.
Translated from Izmeritelrnaya Tekhnika, No. 3, pp. 48-50, March 1985.
272
0543-1972/85/2803-0272509.50
9 1985 Plenum Publishing Corporation
TABLE 1 2omponent
Manganese Copper Silicon Chromium Titanium Nickel Aluminum Molybdent~n Vanadium Carbon Niobium
Standard deviation in % and estimate of mass fraction in % 0~,03--0,1
0,1--0,3
0,3
1
1--3
0,003
0,006
0,01
0.,005
0,008
0,01
0,03
0,004
0,007
0,01
0,02
0.,003
0,006
0,01
0,02
0,005
0,01
0,02
0,03
0,0{)4
0,007
0,0.1
0,02
0,004
0,007
0,01
0,02
0,008
0,01
0,03
0,04
0.003
0,0I
0,02
0,03
0.003
0,00,5
0,01
0,02
0,006
O,Ot
0,02
0,04
0,02
This feature may have played a part in causing the area of physicochemical measurements to be less well covered by test schemes particularly intended to provide unified means of measurement. For example, there are only two national test schemes together with standard [7, 8] out of 136 total [9] having a direct bearing on composition measurement. It should be emphasized that in composition determination, the main efforts of metrologists should be directed to devising a methodology of accuracy evaluation for technical measurements along with appropriate facilities and practical techniques. Such a methodology has now been set up for composition determination, in which the key role is assigned to standard specimens SS. The basic components are the following: standardizing specifications for accuracy in technical measurements (analyses), standardizing and certifying methods of carrying out measurements, and the general use of standard specimens having certified values for the quantities characterizing the composition as standard means of measurement, and checking the accuracy of the results not only immediately within the laboratory but also between one laboratory and another, which can be carried out by departmental services. We cannot consider all these elements, detail.
but we discuss the role and functions of SS in
We merely note that establishing accuracy standards has already been partly considered for many years by including the corresponding specifications in standards for grades of material and analysis methods. At the stage of drafting over 3500 sectional and national standards, metrological evaluation has been applied, in which the corresponding specifications were checked out or included in the standards [i0]. To these standards we need to add over i000 sets of scientific documentation that have passed metrological evaluation in the Siberian Branch of the All-Union Metrology Research Institute and the Institute of Standard Specimens at the Central Ferrous Metallurgy Research Institute by the end of 1982. The availability of these norms provides a solid basis for certifying methods and evaluating measurement states. Determination of the chemical composition of a substance is an experimental procedure and has substantial specific features in addition to those found in many measurement experiments, because the substance (physical system) undergoes various physicochemical transformations in chemical methods, in which it completely ceases to exist. Nearly always, such a measurement experiment breaks up into two essentially different stages. In the first stage, one frees an extracted chemical component from its physicochemical links with the system (matrix) and identified it, which may involve several steps. In the second stage, one measures the amount (mass or volume) of the isolated component, where the physical quantity to be measured interacts with the means of measurement. This scheme is characteristic of most standardized measurement methods (see for example the series of standards on methods of analyzing elements in nickel and copper-nickel alloys: GOST (AllUnion State Standard) 6689.1-80 to GOST 6689.23-80).
273
The procedure fer isolating a given component is a complicated and elaborate microtechnology, which is based on advances in analytical chemistry, but which is not always adequately standardized. Some of the intermediate steps in separating components do not have metrological criteria, so it is not possible to perform metrological checks operation by operation or to draw up an error equation. Therefore, the error in the result is often more dependent on the care with which the technology is carried out and the skill of the experimenter than on the error in the measurements at the final stage in the measurement experiment. This interweaving of analytical and measurement operations greatly complicates the theoretical and practical evaluation of errors in the final results, and the given metrological characteristics of the MM are often insufficient to provide final evaluations. The most reliable way of checking such measurements is check analyses of SS, these being metrologically certified analogs or close equivalents of the relevant physical system [Ii]. When these are used in a check experiment, one repeats all the stages in determining the composition, including ones not related to the direct measurements, in order to provide high reliability in the check procedure. As SS are metrologically certified analogs for test objects, not means of transmitting units, we have an explanation for the large number of SS in use (over 3000 types). In fact, one does not need such a large number of SS types to transmit physical quantities. For example, the 136 existing test schemes for all areas of measurement define a total of about 450 forms of standard measure, which exhaust the requirement for transmitting units for basic physical quantities. Certifying SS during production involves the same procedure as in technical measurements. However, they are carried out by particularly well-qualified staff in certain laboratories that have accumulated considerable experience in such measurements. The results obtained in the various laboratories are passed for statistical processing, which enables one to consider the SS certification results quite reliable. The results can be considered as very reliable information on the accuracy attainable in practical measurement of the corresponding characteristics. The range of confirmed SS types gives this information on the accuracy for tens of thousands of characteristics. In certifying SS, one uses the most thoroughly tested chemical methods, which in turn should undergo metrological certification. The latter requirement is laid down in the documentation governing the sequence of certifying SS [12]. For example, to certify SS for the main range of materials in ferrous metallurgy, these are 15 analysis methods [13] used to determine the mass proportions of elements such as carbon, silicon, manganese, chromium, and phosphorus (a total of 30 components). Each component is analyzed by three to seven different methods. Table i gives the standard deviations in estimating the mass fractions of certain major components in certifying SS for carbon and alloy steels. According to [13], this level of accuracy is equal to that attained in the Federal Republic of Germany, Britain, and the USA. An additional confirmation of the certified characteristics is provided by international comparisons conducted in recent years for numerous types of SS, which have always confirmed the reliability of the error estimates. When SS are certified, one may use a subordination system for them. In particular, studies are being conducted on the scope for setting up a restricted range of SS for certain pure substances certified by absolute methods with the highest possible accuracy, which would be used in certifying a wider range of subordinate SS. The choice of substances for the SS is based on the scope for the strict realization of stoichiometric relationships between substances differing in chemical nature. Practical methods of using SS are mentioned in most national and sectional standards for measurement methods, and also in techniques for carrying out measurements and in instructions for using SS, which must be supplied with each SS. The use of SS enables one to check the errors in technical measurements and has an effect on detailed experiments designed to put errors of measurement in accordance with the set standards.
274
SS are used in periodic checks on measurements carried out by chemical methods, in constructing particular calibration characteristics for measurement devices (spectrometers, spectrographs, and so on) not having fixed scales, and for direct use in measurements by substitution or comparison methods. We omit here a further function of SS, namely checking means of measurement, since that application is primarily designed to provide unity in means of measurement. We have considered methods of using SS for checking measurements in general form above, such checks being often carried out very frequently: in laboratories in ferrous metallurgy, for example, not less often than once per shift. In the same area, wide use is made of a set of control cards, on which one records the spreads in the results from monthly measurements on SS characteristics over a long period, which gives a reliable indication of the overall measurement level in a particular laboratory ([14], pp. 91-111). One also constructs or checks calibration characteristics in analytical equipment periodically during the measurements. The frequency and scale of the check calibrations are dependent on the features of the particular method. For this purpose, one uses calibration SS or sets of them, which cover a certain concentration range. The choice and use of check standards is considered in detail in [14], pp. 83-86. These methods of checking accuracy in technical measurements on compositions and of ensuring unification are widely employed in the analytical laboratories of the Ministry of Ferrous Metallurgy, the Ministry of Geology, and the Ministry of Nonferrous Metallurgy. These ministries have set up extensive sets of SS, which in the Ministry of Ferrous Metallurgy amount to 1170 types, and in the Ministry of Nonferrous Metallurgy, to 675. The Ministry of Ferrous Metallurgy is also conducting an operation to generate systems of sectional and inhouse SS, and over 800 types of SS have already been officially confirmed. This is an extensive and powerful metrological basis for providing unified composition measurements, and hundreds of sectional laboratories and organizations have collaborated in setting it up. Their activities on a unified methodological basis represent a very valuable contribution to unifying measurements. The final and most effective phase in efforts to unify measurements is provided by evaluating the state of technical measurements by checking the accuracy of the results directly in the laboratory. One can use SS as certified models or analogs for the test systems, which is the best method of metrological checking in composition determination. Effective and constantly operative checking enables one to transfer the center of attention in measures to support unified measurement from formal factors to experimental checks on measurement procedures (the formal factors are the observation of certain measuring-instrument checking periods and the use of confirmed measurement methods, etc), which enables one to ensure that the measurement results meet the requirements of the technical documentation. These checks enable one to make corrections during intervals between tests on means of measurement and to modify the measurement methods and other components of the metrological support system. LITERATURE CITED i. 2. 3. 4. 5.
6. 7.
All-Union State Standard (GOST) 16263-70: State System of Measurements: Metrology: Terms and Definitions [in Russian]. G. A. Burdun and V. N. Markov, Principles of Metrology [in Russian], Standartov, Moscow (1972), p. i0. N. P. Mif and M. A. Zemel'man, Planning Technical Measurements and Evaluating their Errors [in Russian], Standartov, Moscow (1979), pp. 4, 5. All-Union State Standard (GOST) 1.25-76: State System of Measurements: Metrological Support: Basic Concepts [in Russian]. All-Union State Standard (GOST) 8.010-72: State System of Measurements: Basic Specifications for Standardizing and Certifying Methods of Carrying Out Measurements [in Russian]. All-Union State Standard (GOST) 8.011-72: State System of Measurements Accuracy Parameters for Measurements and Representation Forms for Measurement Results [in Russian]. All-Union State Standard (GOST) 8.190-76: State System of Measurements: The State Special Standard and All-Union Test Scheme for Means of Measuring the Volume Water Contents of Petroleum and Oil Products [in Russian].
275
8. 9. i0. ii. 12.
All-Union State Standard (GOST) 8.480-82: State System of Measurements: The State Test Scheme for Means of Measuring Water in Grain and Grain Products [in Russian]. Normative Technical Documentation in Metrology: An Index [in Russian], Standartov, Moscow (1982). A. V. Pokrovskii et al., Izmer. Tekh., No. 6, ii (1980). N. G. Semenko, Izmer. Tekh., No. i, 22 (1983). RD 50-270-81: Statements of Method: Sequence for Carrying Out Interlaboratory Certification Analyses and Establishing Basic Metrological Characteristics of State Standard Specimens for Compositions of Substances and Materials [in Russian], Standartov, Moscow
(1982). 13. 14.
Yu. L. Pliner (ed.), Metrological Support to Monitoring the Compositions of Materials: Handbook [in Russian], Metallurgiya, Moscow (1981). Yu. L. Pliner et al., Quality Control in Chemical Analysis in Metallurgy [in Russian], Metallurgiya, Moscow (1979).
A MICROPROCESSOR CONTROL DEVICE FOR A NEUTRON WATER CONTENT METER FOR POWDERS UDC 543.812.08:539.125.5
A. K. Stroikovskii, V. A. Pronyakin, A. A. Pershin, and V. D. Savelov
The neutron method is one of the most effective for determining water contents of powder materials under industrial conditions. Neutron gauges based on this method provide rapid measurements over wide ranges. Also, they provide an integral evaluation for a comparatively large volume of material. The working principle is that neutrons are moderated by the hydrogen in the water molecules when the material is exposed to fast neutrons. The water content W is determined by measuring the amount of water 9W in the material and the poured mass (density) of the dry material Pc using the formula
( 1~ --b0)a
Pw
(fw--a~)'q
~c
~jD--Oo)tq - (fw--a0)bl '
(1)
where JW and Jp are the intensities of the neutron fluxes, which are proportional to the change in bulk water content and the poured density [i, 2], while a~,b~,a~b~ are characteristics of the sensitivity of the slow-neutron counters to changes in those two parameters, and 0 and b 0 are constant coefficients. To keep the technical and metrological characteristics of the system within set limits, and also to calibrate the instruments during use without resort to laborious operations, current gauges employ special calibration devices, which simulate water content and density at several points in the measurement range. During the calibration one determinesa0, a~, a~ bo.b~.b2; although the calibrators simplify the operation considerably, periodic calibration during use is fairly laborious. Ionizingradiation protection is important during the calibration. This makes it necessary to automate the process. Previously, these neutron instruments were based on series K 155 integrated circuits, and the instruments were specialized, i.e., a particular processing device had to be developed for each material (coke, iron-ore concentrate, etc.) and for each set of measurement conditions, in which the processing operations were implemented by hardware. The hardware required to implement the computations even when using the basic algorithm based on (i) was up to 75% of the complete set of integrated circuits and logic modules. The values of ~,aba~, bo, bl, b2 were determined by the operator and supplied manually, which complicated the calibration. Translated from Izmeritel'naya Tekhnika, No. 3, pp. 50-51, March, 1985.
276
0543-1972/85/2803-0276509.50
9 1985 Plenum Publishing Corporation