CHAPTER III

GENERAL OPERATIONS

33. The general operations with which the analyst has to deal are in their usual order of occurrence: sampling, weighing, measuring volumes, solution, fusion, evaporation, precipitation, digestion, filtration and washing, ignition and drying. These operations will be discussed in detail in the following paragraphs except that the discussion of the topics of weighing and measuring volumes, on account of the extensiveness of detail, will be expanded into separate chapters.

It is to be remarked at this juncture that the key-note to the accomplishment of results in quantitative analysis is not haste in carrying through these operations individually but rather the careful planning of work so that two or more operations are being conducted or are taking place simultaneously; thus while one precipitate is being washed another is being ignited, or while a precipitate is cooling in the desiccator, a weighing is being made. It will soon be found that there is really no such thing as "speed" in quantitative analysis; it is careful planning and thoughtful coördination that bring results. The great thing is not to be required to repeat a determination, and in this respect there are certain points the observance of which will do a lot toward cutting down the chances of a repetition. The desk should at all times be kept scrupulously clean, and it is a good plan to begin the day's work by wiping off the desk with a wet sponge or rag and to close the day's work the same way. All apparatus not in use should be cleared off the desk, cleaned at once, and put away, while samples, precipitates, filtrates and all solutions should be marked systematically by means of a marking crayon or small labels so that no confusion may arise as to the identity of things. In brief, systematic orderliness and cleanliness are the two watchwords of accomplishment.

34. Sampling. It is of the utmost importance that the sample submitted to analysis should have a composition which is sensibly the same as the average composition of the original mass from which the sample was derived; otherwise the results of analysis will not give the true percentages of the constituents in the original mass. The object of sampling is to obtain a sample any portion of which shall have a composition truly representative of the average composition of the original lot, and it cannot be too strongly stressed that an analyst should always assure himself of the history of the sample submitted to him and make certain that the precision of the sampling is sufficient for the particular purposes of the analysis. If he is not certain of this point he should direct the taking of a new sample or else attend to it himself.

If materials were strictly homogeneous in their composition, i.e., if for every specimen all the small unit masses into which we might imagine each specimen to be divided showed precisely the same composition, the problem of sampling would be easy because then it would make no difference which particular portion we took as its composition would be the same as that of the lot. It rarely happens in practice that materials are homogeneous 1 so that a portion withdrawn from one part of a lot will not have the same composition as that withdrawn from another part, while neither portion will be representative of the lot. However, by drawing a goodly number of portions in a systematic manner from different parts of the lot and combining them, it is possible to obtain a sample whose average composition will approach very closely to that of the lot. The number of portions to be drawn and their relative size depend upon two things, the nonhomogeneity of the lot and the precision with which the sample is to represent the lot; the more the non-homogeneity of the lot and the more the precision required the greater the number of portions that must be taken and the greater their relative size.

1 Thus to cite a few illustrations: minerals, metals, alloys and other products which have solidified from the molten state almost invariably show a segregation of their constituents; chemical salts are seldom free from the effects of their deliquescent or efflorescent properties; liquids will often show a variability in their composition due to the loss of some volatile constituent; mixtures of gases, particularly flue gases, will often show great non-uniformity owing to lack of sufficient time for the several gases to interdiffuse among themselves thoroughly.

In the absence of specific data as to the non-homogeneity of a lot but assuming a fair degree of homogeneity, and assuming further that the composition of sample shall represent the average composition of the lot to a precision of one part per 1000, the analyst may take the following directions as a general guide. The number of portions withdrawn from the lot shall not be less than twenty; they shall be approximately of the same size and shall represent regions of uniform spatial distribution in the lot; their aggregate mass with respect to the original lot should never be less than 1/100, and if it is convenient to obtain a larger ratio so much the better although it is seldom that a ratio greater than 1/20 is ever employed.

The sample so collected will approximate very closely in its average composition to that of the lot, but, like the lot, it will be non-homogeneous. It, in turn, might serve as a source for a second smaller sample by the method outlined above, but in order to procure an ultimate sample that shall be uniform in every subportion, the preceding sample must be made as homogeneous as possible before a further sample is drawn from it. For solids this means successive subdivision and intermixing of the particles or pieces; for liquids and gases it will naturally mean just thorough intermixing as the subdivision already exists.

For the subdivision of friable substances various instruments are used: where the pieces range in size from 2 in. diameter down to 1/8 in. diameter, jaw crushers operated by power will be found very serviceable or if the quantity of material is small a chilled steel muller may be employed; for sizes under 10 mesh, say down to 200 mesh, small pebble mills operated by power are to be preferred; for small quantities of material, steel or agate mortars work very nicely. It is to be remembered that any operation of subdivision always introduces more or less impurity into the sample from attrition of the grinding instrument. The size to which the subdivision is carried depends upon the weight of the sample. Thus to quote the figures of Bailey, J. Ind. & Eng. Ch. 1, 161 (1909) for the sampling of coal to a precision of 1 part per 1000, the relationship given herewith is necessary:

For intermixing the particles or pieces of a sample the following devices, among others, are used. If the quantity of material is large a pile of it is built up as a cone about a rod as a center. The material is thrown by shovelfuls on top of the cone in such a manner that it slides down and distributes itself more or less evenly over the surface of the cone. After the cone is completed it is flattened out into a disc or flat circular plate by means of shovels working from the apex to the periphery round and round the cone. The disc is then separated into quarters, two opposite quarters rejected, and the remaining quarters piled up into another cone, which is flattened out and again quartered. This process is repeated until the sample is reduced to the proper size mass, when it is subjected to further grinding and intermixing.

If the quantity of material is about 5 lb. or less, intermixing is accomplished by the method known as "tabling." This consists in spreading the finely divided material in the center of a piece of oil-cloth or paper about 30 in. square and pulling each corner in succession over to its diagonal partner, thereby causing the mass of particles to roll over on themselves. The lower portions are constantly brought to the top of the mass and are intermixed thoroughly; at the same time an interchange of material between the four quarters of the mass is effected. Each corner is manipulated in turn until the operator has rotated the manipulation around the cloth a sufficient number of times, say at least ten.

After the sample has been made as homogeneous as possible it is ready to be used in turn to furnish a smaller sample if necessary. In this second sampling, or any subsequent sampling for that matter, the same circumspection must apply to obtain a repre

sentative sample and then to make it homogeneous as applied in the first sampling. The final sample for the laboratory should usually be not less than 25 grams and in many cases larger amounts even up to 1000 grams will be required, depending on the purposes of analysis. All samples should be put and kept in air-tight bottles as soon as obtained. Even after the proper preparation of a sample it should be remembered that the principles of sampling still apply, so that in weighing out the small milligram portions for analysis the analyst should not dip into the sample at random but should spread the sample out and select 15-20 equisized portions distributed equispatially throughout the mass.

35. Weighing. The operation of weighing is of fundamental importance in quantitative analysis as all determinations are either directly or indirectly based upon the measurement of the relative masses. For first-class quantitative work weighing must be made to a precision of 0.1 milligram. While the attainment of such refinement is in itself not a matter of great difficulty for the masses ordinarily encountered, there are certain details which must be carefully observed and certain errors which must be studiously avoided. As these cannot be adequately described in brief, their discussion, along with the general theory of weighing, has been set apart for Chapter V.

36. Measuring Volumes. The general problem of measuring volumes is one with which the analyst is concerned almost as frequently as he is with that of weighing, because of the fact that there are so many methods of analysis which employ the procedure known as titration. By this method of analysis the constituent sought is not precipitated and weighed but is determined by allowing a solution of known composition to react with it, and from the volume of solution used the amount of constituent is calculated by the laws of chemical equivalence. With respect to titration methods it may be mentioned that some of them possess extreme accuracy, while all of them possess the advantage that for routine work they are very rapid. In so far as the precision of such methods depends upon the measurement of volumes, it is to be noted that the precision falls off rapidly with the smallness of the volumes employed. In general we try to arrange matters so that we will employ volumes that can be meas