10 mi. below the slit and adjusting the pulse height selector on the scintillation counter, the necessity for making the correction can be eliminated. Sample 1, Table I, n a s taken from the material as it came from the ball mill. The distribution of particle size is nearly linear when plotted on lognormal paper, with the n eight mean size a t 1.3 microns. The distribution as obtained by radioactivation is illustrated graphically as the upper histogram in Figure 2. The radioactivation analysis in Table I is the average of five runs on t w o different I ortions of sample 1 and the fractional extraction analysis is the average of three different portions of sample 1. The agreement between thc two niethods is considered satisfactory. Samples 2 and 4, Table I, and sample 3, Tahle 11, were prepared from a large batch of saniple 1 by fractional extraction. I n these samples the distributions of particle size were distorted in order to give an additional test of the radioactivation method. The distortion introduced into the distributions is illusstrated graphically in Figure 2 where the histograms for saiiiples 2 and 4 are plotted to compare with the distribution of sample 1 , The results for sample 2, 3, and 4 alqo skion excellent agreement between the two methods. The sources of error peculiar to the radioactivation niethod are counting statistics, inconiplete dispersion. and scattered radiation. 4 s mentioned above, scattered radiation can be largely eliminated by suitable geometrical arrangement and electronic discrim-
Table 11.
Particle Size Determination of Sample 3
Particle Keight Fraction Range, RadioFractional Microns activation Extraction 0
0 to 2 2 to 3
0.372 0 073
0.379
b
0.395 0 059
0.053 3 to5 0.180 0.221 0.213 5 to 10 0 08T 0 079 0.100 0.268 0.232 > 10 0.287 a Prior t o radioactivation analysis. Xfter radioactivation analysis.
in :ic.curnc>- \\-it11 a i i i r ~conventional ~ scdinientation technique. The method is rapid, capable of being adapted to automatic operation, and to taking close size intervals. I t is particularly suited t'o reactive systems that must be handled in vacuo or in an inert atmosphere. An additional application suggests itself-namely, the deterniination of the particle distribution of the components in a niisture by elect'ronic discrimination of the respective radiations. thereby avoiding the necessity of separating tlie pon-der into its components. LITERATURE CITED
ination. If nonuniforni dispersion isuspected, this can easily be checked by scanning the length of the tube with tlie counter while stirring. The counting statistics actually limit the accuracy of the method; this means that the largest weight fractions are known with the greatest precision. Here the activation technique has a distinct advantage over fractional extraction or pipet methods, for if the bulk of tlic material in a particular size interval is found near the point where a cut is to be made. there can he c:trry-over to the adjacent size brackets. Furthermore. close size intervals may be obtained n-ith the activation method without the tinieconsuming tediousness required by the other methods. SUMMARY
The radioactivation method for particle size deteriiiination is coniparalile
(11 Gronv@ld, F., J . Inorg. ie. Sziclear C h e m , 1,357-70 (1955). ( 2 ) Herdan, G., "Small Particle Statistics,' ' Eleevier, S e w York, 1953. ( X , Inst. Chem. Engrs., (London) and Soc. Chem. Ind., Road andBuilding Materials Group, Symposium on Particle Size A\nalysis, Trans. Inst. C'hem. Engrs. (London) Suppl., pp. 128-45 (Feb. 4, 1947). (4) Inst. of Physics, Brit. J . .Ippl. Phys. Suppl. s o . 3 (1954). ( 5 ) "International Critical Tables,'' Vol. 3; p. 25, 3IcGraw-Hill, Set? l o r k . 1928.
Co.. Sew'Tork. 1954.
-
RECEITEDfor review October 20, 1956. Accepted 1Iay 1, 1957. Presented at the Second Annual Meeting of the American Nuclear Society, Chicago, Ill., June 1956. V o r k performed under the auspices of the IT,S.Atomic Energy Commission.
High-Frequency Determination of Ferromagnetic Metals FRANCIS J. SCHMIDT1 Nickel P rocessin~Corp., Nicaro, Orienfe, Cuba
The ferromagnetic metal content of reduced oxide i s determined b y its magnetic permeability, which i s measured b y its influence on the induction of a coil forming p a r t of a tuned oscillator circuit. The method i s calibrated in an empirical manner b y use cf chemically analyzed samples. Large, coarsely crushed samples reduce sample preparation time, and permit completion of a determination in several minutes.
I
s THE ferrous metals industry the reduction of oxides to metals is a common process. Many process variables must be closely controlled to ob1060
e
ANALYTICAL CHEMISTRY
tain products within specified limits of reduced metal content. iinalytical determinations lose much of their effectiveness for process control TI hen they require considerable tinip. If sintering occurs, a product does not lend itself to rapid analytical niethodi, such as spectrography, because of sampling difficulties. Individual particles range from coarse pieces to finc dust :tiid the chemical compositions of the fractions vary. Fine dust is incvitahly lost during size reduction :ind splitting. Brittle particles are miwd together with soft, malleable metallics; the latter have a tendency to flatten out during crushing. Frequently many hours are required for the preparation
of 1-gram samples, and the sampling error is often several times greater than the analytical error. This pa1:cr describes a method capable of using larger arid coarser laboratory samples. I t is based on permeability inpasuremerits at high frequencies and i4 suitable fur the determination of iron, nickel. or cobalt, when only one of these nietals is pre*ent as a major constituent. The composition and temper of the eventual alloys formed can be regarded as constant. Smong the magnetic properties, permeability measurements offered the adl Present addresp, Rey1101ds 1Ietals Co., Sheffield, .\la.
vantages of easy determination and sufficient precision. Only in the determination of ferrosoferric oxide were nieasurements based on remanence and on coercive force, after exposure to fields of several hundred oersted
