August, 1926
INDUSTRIAL AND ENGINEERING CHEMISTRY
The volume per cent of oxygen remaining in the effluent gas is given as a measure of the catalytic efficiency. Although the highest accuracy was not sought in these experiments, the results of blank and reproducibility tests indicate that the maximum error is equivalent to less than 0.01 per cent oxygen in the exit gas results. The results show a decided advantage for reduced oxide catalysts over massive copper. Copper from the fused oxide was the most active, although both of the reduced catalysts are capable of effecting practically complete removal of 2 or 4 per cent of oxygen a t 60,000 space velocity, which corresponds to a flow of 1000 liters per minute per liter of catalyst volume. It will be noted that the oxygen leakage a t equal rates of flow is less for gas containing 4 per cent oxygen than it is for 2 per cent of oxygen. This is readily explainable when it is
867
remembered that the temperatures given in the table are those of the vapor baths used and that the actual temperature of the catalyst was higher than this in proportion to the exothermicity of the reaction, which in this case, of course, increased with the oxygen content. I n conjunction with these experiments, a few tests were made employing 80-mesh platinum gauze as a catalyst. Extensive tests were not made, since a preliminary study indicated that it was much less promising than the reduced copper in granular form. It may be of interest, however, to state that a single layer of this gauze heated electrically to 750’ C. was able to decrease the oxygen concentration from 2 or 4 per cent, to about 0.01 per cent, with a gas flow of 1 liter per minute per square centimeter of gauze area.
The Suitability of Chloroform for Alkaloidal Assay’ By H. R. Watkins and S . Palkin BUREAUOF CHEMISTRY, WASHINGTON, D. C.
The usual purity tests t o determine the suitability of chloroform for alkaloidal assay are inadequate. Automatic continuous extraction afords easy and certain meam o f determining such suitability. HE presence and character of contaminating impurities in chloroform have been exhaustively studied by Baskerville and Hamor2 in their general investigations of anesthetics. In the last article of the series these authors present a scheme for testing chloroform, which forms the basis of present Pharmacopeial and analytical reagent tests now generally employed in the examination of this product. In this paper it is proposed to show that the ordinary tests employed for examining chloroform are not adequate to ascertain its suitability for alkaloidal assay. This report deals with the subject in a more or less empiric way, and no attempt is made to determine the origin or nature of contaminants. Chloroform is doubtless the most widely used of all solvents in pharmaceutical analysis. While it is well known that a high-grade chloroform is necessary in such analysis, the extent and manner of purification necessary to render it suitable for alkaloidal assay are hardly appreciated. The error that might be introduced by the use of faulty chloroform where the alkaloid extraction is carried out in the cold by the hand-shaken separatory funnel is variable, depending on a number of factors not easily controlled. Automatic, continuous e x t r a c t i ~ n however, ,~ where contact of solvent (hot) with the alkaline aqueous layer and with the alkaloid can be conveniently prolonged and controlled, affords a satisfactory means of ascertaining the effect of the solvent on the alkaloid being extracted. The character of the errors introduced indicates that in the main partial neutralization, and to some extent destruction also, has occurred. I n effect, however, a low titration result is obtained.
T
Solvents Used
Chloroforms 1, 2, 3, 4 , and 9 made by different manufacturers, all labeled “U.s. P.” and found upon examination to meet the Pharmacopeial requirements. Reagent Chloroforms R-5, R-6,R-7, R-8, obtained from different manufacturers of chemical reagents and labeled as reagents, sold as of analytical reagent grade. 1 2
*
Received May 10, 1926. THIS JOURNAL, 4, 212, 278, 362, 422, 499, 671 (1912). Palkin, Murray, rrnd Watkins, I b i d . , 16, 612 (1925).
Chlorofomns T-3and T-4 are chloroforms 3 and 4, respectively, which have been given a purification treatment involving refluxing for several hours with brucine and then removing from the alkaloid by distillation. Experimental Results
VARIATION IN CHLoROPonM-The fact that some u. 8. P. and reagent chloroforms may be unsuitable for quantitative alkaloidal extractions was forced to the writers’ attention by the low and variable results obtained in the course of an investigation, when a change was made in the brand of chloroform being used (Table I). Table I-Nux
Vomica Preparations ALKALOIDS PER 10 cc. FLUIDEXTRACT Chloroform 1 Chloroform 2 No. Method of assav Gram Gram 0.2380 46671 u. s. P. 0,2293 0,2284, 0.2193 7.7. S. P. modifieda 0.23io Automatic extractor 0.2417,0.2417 U. S. P.modifiedD 50603 0.1984, 0.2029, 0 . 2 0 1 8 0 . 2 1 5 5 , 0 . 2 1 6 4 , u. s. P. 0.2133 Automatic extractor 0.2155 Automatic extractor 0.0895, 0.0976, 0.0946 0.1088 Sol. Xb 0.1463 Sol. xx Automatic extractor 0.1274 a J. Am. Pharm. Assoc., 13 691 (1924). b Sol. X gave with benzene a; solvent-0.1086 gram alkaloid per 10 cc. Buid extract.
This showed the need for investigating the various brands of reagent chloroform on the market. Chloroform 1, which, as shown in Table I, gave erroneous results, was subjected to an even more rigid examination for compliancewith reagent requirements by the usual reagent tests and found to be superior, if anything, to chloroform 2. The alkaloid extractions were made, as indicated in the table, by several different methods, both in the hot continuous method and cold hand-shaken separatory funnel method. Considerable variations are obtained with the two different chloroforms. It is also apparent that even in cold extraction (U. S. P. method) the error is appreciable. Incidentally, it may be observed that the better chloroform (No. 2) on prolonged extraction with hot solvent, using automatic extractor, not only showed no loss of alkaloid, but the results so obtained were even higher than the highest by the separatory
INDUXTRIAL AND ENGINEERING CHEMISTRY
868
funnel method, thus indicating that a suitable grade of chloroform is not affected by the process of continuous extraction. Nux VOMICAPREPARATIONS (Table 11)-Data on the combined alkaloids of nux vomica, obtained by extraction with other brands of chloroform, including "treated" chloroforms and special chloroforms of "analytical reagent" grade show that not even "reagent" chloroform can be depended upon to be suitable, that both chloroforms 3 and 4 are improved by the purification treatment, and that extraction with benzene yields the highest results. Table 11-Fluid Method of assay
E x t r a c t Nux Vomica
Solvent ' Chloroform 3 Chloroform 4 Chloroform 2 Chloroform T - 3 Chloroform T-4 Chloroform R-5 Chloroform R-6 Chloroform R-7 Chloroform R-8 Benzene Chloroform 2 Chloroform T-4 Chloroform 4 Chloroform T-4 Chloroform 2 Chloroform 4
Automatic extractor
U. S. P. modified
u. s. P.
Alkaloid Gram/lO cc. 0.2410, 0.2402 0.2300, 0.2322 0.2432, 0.2424 0.2424, 0.2424 0.2424, 0.2432 0.2439, 0 . 2 4 3 9 0.2453, 0.2450 0.2453, 0,2453 0.2373, 0,.2366 0.2461 0.2432 0.2424 0.2432, 0.2424 0.2424 0.2432 0,2424
BELLADONNA ALKALOIDS (Table 111)-When employed for the extraction of belladonna alkaloids the chloroforms behave in the same way. It is of interest to note that the poorest chloroform, No. 4, sometimes, although not always, yields good results in cold extraction (U. S. P. method). Here, too, the treated chloroforms show an increased yield of alkaloid. The most interesting point noted in this series (see also Atropine, Table VI) is that good results are invariably shown by benzene extraction even when compared with the best chloroform. This was pointed out in a previous publication,*and is somewhat surprising in view of the reputed instability of belladonna alkaloids. The higher extraction temperature (approximately 18 degrees higher than that of chloroform) evidently causes no destruction of alkaloids under the conditions. One of the poorest chloroforms (R-8) is of the reagent grade. T a b l e 111-Fluid Method of assay
Automatic extractor
u. s. P. u. s. P.
U.S. P. modified
Extract Belladonna R o o t Alkaloid Solvent Gram/lO cc. Chloroform 3 0.0463, 0 . 0 4 6 7 0.0440, 0.0435 Chloroform 4 0.0500, 0.0500 Chloroform 2 Chloroform T-3 0.0491, 0.0495 0,0454, 0 . 0 4 4 9 Chloroform T-4 0.0518, 0.0523 Chloroform R-5 0.0618, 0.0523 Chloroform R-7 0.0382, 0.0384 Chloroform R-8 Benzene 0.0504 0 . 0 4 9 5 0 . 0 5 0 9 0.0497: 0.0486' Chloroform 2 0.0497, 0.0440, 0.0434 Chloroform 4 0.0495 Chloroform 2
I i
STRYCHNIXE AND BRUCINE (Tables IV and V)-Brucine seems to be more susceptible to error than strychnine. This was one of the reasons why brucine was employed in the "purification refluxing treatment" of chloroforms 3 and 4. Benzene is also a reliable solvent for extraction purposes. Table IV-Strychnine (Automatic extractor method, except where otherwise indicated. Approximately 0.2 gram of alkaloid used for each determination.) Alkaloid recovered Solvent Per cent 9 8 . 4 1 , 98.57, 9 6 . 8 3 Chloroform 4 100.00, 100.46, 9 9 . 2 9 Benzene of chloroform 4" 98.78, 100.00, 99.85 Chloroform T-4 99.64, 9 9 . 1 3 , 9 8 . 9 3 Chloroform 3 100.15, 100.00, 100.00 Benzene of chloroform 3" Chloroform T-3 100.00, 100.00 Chloroform 4 99.3, 98.2 Chloroform 2 100.4, 99.3 Benzene 100.00, 99.85, 100.00 a T h e alkaloid extracted with chloroform was titrated in t h e usual way and t h e titrated solution was made alkaline and extracted with benzene. 4
J . Am. Pharm. Assoc., 14, 1099 (1925).
Vol. 18, No. 8
Table V-Brucine (Automatic extractor method, except where otherwise indicated. Approximately 0.2 gram of alkaloid used for each determination.) Alkaloid recovered Solvent Per cent Chloroform 4 95.47, 94.52, 9 5 . 9 2 Benzene of chloroform 4' 98.42, 98.42, 9 7 . 9 2 Chloroform 3 9 6 . 5 2 , 97.02, 9 5 . 9 7 Benzene of chloroform 3" 9 9 . 1 2 , 97.92, 9 8 . 6 2 Chloroform R-5 9 9 . 9 5 99 95 Chloroform T - 4 100.14: 98:OO. 97.34, 96.75 Chloroform T - 3 97.70, 9 7 . 7 0 Benzene of chloroform T-3" 97.70, 9 7 . 7 0 Benzene 100.14, 99.85 Chloroform 4 , hand-cold 98.81, 98.45 Chloroform 2 98.46 98 12 Chloroform R-6 98.95: 98:OO Chloroform R-7 9 8 . 0 0 98 00 Chloroform R-8 98.71: 98:71 a The alkaloid extracted with chloroform was titrated in the usual way and t h e titrated solution was made alkaline and extracted with benzene.
ATROPINE (Table VI)-Consistent variations are exhibited by atropine with the different chloroforms. Benzene proved to be a reliable solvent with belladonna. The unreliability of some reagent chloroforms is apparent. MORPHINE,CODEINE,QUININE (Table VI)-These are less susceptible to error with the various brands of chloroform than strychnine, brucine, and atropine. The danger of using poor chloroform is nevertheless obvious even here, particularly with quinine. T a b l e VI-Extraction
of Alkaloids-Automatic
Extractor Method
SAMPLES TAKEN C P Atropine 83.5 mg (equivalent to atropine sulfate H20, 100 mg.) C' P' Morphin; sulfate '(+5Hz0) 200 mg. C: P: Codeine sulfate (+5H20), io0 mg. C. P. Quinine (+3HtO), 100 mg. MORPHINE CODEINE ATROPINE SULRATE SULFATE QUININE Alkaloid Salt Salt Alkaloid, recovered !+5Hz0) (f3HnO) anhydrous Solvent Per cent Solvent Mg. Solvent Mg. Solvent Mg. Chl. 1 b 76.8 Chl. 3 192.6 Chl. 3 9 9 . 7 Chl. 3 86.88 84.1 191.9 100.0 86.21 Chl. 2 99.2" Chl 4 193.0 Chl.4 9 9 . 7 Chl. 4 90.58 191.9 100.0 81.47 Chl. 3 98.1" Chl. 2 195.8 Chl. 2 101.7 86.21 193.8 103.3 88.22 195.4 102.1 Chl. 4 87.2" 102.5 102.5 Chl. T-1 95.5" Benzene 103.7 Chl. 2 86.21 104.1 86.21 Chl. T-2 99.0" Benzene 9 0 . 6 8 91 5 9
Chi. T-3 Chl T-4
95.65 92.4 95.5 Benzene 99.7" Chl. R-5 96.7a Chl. R-7 95.0 Chl. R-8 9 8 . 1" Chi. R-9 97.8" a Average of several closely agreeing determinations. b Chl. = chloroform.
Errors Due to Use of Impure Chloroform
The nature of the errors introduced by faulty chloroform is apparent, when the alkaloid extracted with chloroform and titrated is reextracted with benzene (Tables IV and V). The high benzene recovery values would indicate that the errors were due largely to partial neutralization of the extracted alkaloid during the chloroform extraction. In the case of strychnine that would appear to be the only effect. With other alkaloids, however, the benzene extraction (that is of previously chloroform-extracted residues) did not give a complete return of the original quantity of alkaloid, indicating that some destruction had also occurred. Benzene as a n Extracting Solvent
The excellent results obtained with benzene where continuous extract,ion is employed point to the desirability of using this solvent by continuous extraction in place of chloroform for all the alkaloids of this series except morphine. Further advantages in the use of benzene as the solvent, not
IiVD USTRIAL A.VD ESGINEERI,VG CHEMISTRY
August, 1926
obvious from the results alone, become apparent in the practical carrying out of the assay, as the alkaloidal residues obtained from fluid extracts, tinctures, etc., are much cleaner and more sharply titratable than those obtained with chloroform. \ Conclusions
Purity tests commonly used to determine the suitability of chloroform for alkaloidal assay are inadequate. Some chloroforms which comply with the usual tests (IT.S. P. and reagent) as to suitability for this work cause appreciable errors. Automatic, continuous extraction affords an easy, ready, and certain means of determining the suitability of chloro-
869
form for alkaloidal assays. This method of testing suggests a possible manner of treatment to render unsuitable chloroform fit for alkaloid assay-namely, prolonged refluxing with alkaloid (preferably brucine). Although this automatic continuous extraction method effects an improvement, it has not proved fully satisfactory. In view of the variable quality of chloroform with respect to this particular use, the question of a substitute solvent deserves serious consideration, especially for auxiliary and check methods. Experiments with brucine, strychnine, atropine, codeine, quinine, and galenical preparations of nux vomica and belladonna show that benzene when used in continuous extraction apparatus gives excellent results and is recommended as a substitute for chloroform.
Some Factors Influencing Sedimentation’ By Clark S. Robinson MASSICHUSETTS INSTITUTE OF
TECHNOLOGY, C A M B R I D G E , MASS.
HE settling of fine susalso be proportional, not to The p o b l e m of settling fine sludges and precipitates pensions of solids in the viscosity of the fluid but is found in almost all chemical processes. A n analysis to the viscosity of the suswater or other fluid of the rate of settling of such suspensions should has been divided by Walker, pension of the particles in the therefore be of general interest. Lewis, and McAdams2 into fluid surrounding the particle. It is shown that a modification of the Stokes law of three stages-i. e., the first Finally, the settling coeffisettling is applicable to fine sludges, permitting an part, where the rate of Jettling cient should be other than analysis of settling rates and the prediction of the is constant and the settling 2/9, depending on the shape time required to settle suspensions. c u r v e of h e i g h t of sludge of the particles. against time is a straight line; A modification of the Stokes thelast part, where the sludge is approaching its ultimate level law was therefore written: and the rate is getting slower and slower, along a logarithmic dH ka2(D-d) curve; and a n intermediate part, which is a transition zone d s = Z between the first and the last. It was thought that it should where H = height of sludge be possible to obtain a single relation which would describe e = time of settling the whole settling process. k = a proportionality constant
T
= average dimension of particles D = specific gravity of particles d = specific gravity of suspension z = relative viscosity of suspension Q
Stokes’ Law
Stokes’ law3 states that the rate of settling of a, very small spherical particle in a viscous fluid may be expressed by the following equation: V = where V a D d g p
2 a*(D--d)g 9P
= rate of settling = radius of spherical particle = density of particle = density of fluid = acceleration due to gravity = absolute viscosity of fluid
Modification of Stokes’ Law
The difference between the rate of settling of an actual sludge and that of a lone spherical particle may be stated to be (1) the sludge particles are not spherical, and (2) a large number of particles are present in the fhid. Under these conditions, therefore, it seems obvious that the potential or driving force which causes the settling is proportional, not to the differencebetween the specific gravity of the solid and of the fluid, but to the differencebetween the specific gravity of the solid and of the suspension surrounding the particle. The resistance to flow of the particle should
* Received April 24, 2
8
1926. “Principles of Chemical Engineering.” Camb. Trans. t. IX, p 58 (1851).
In this equation, k should be the same for any given kind of solid independent of the size of the particle, provided the shapes of the large and small particles are similar. Experimental Verification
Adams and Glasson4 studied the rate of settling of fine silica particles in water. The following examples were taken a t random from their work. Silica particles, carefully sized by hydraulic classification and measured for average dimension microscopically, were suspended in water by agitation, and then allowed to settle in a graduated cylinder, the height of the top of the sludge being measured as a function of the time. The true and the bulk densities of the silica particles were determined by water displacement methods in the usual manner. The relative viscosity of the suspensions was determined by comparing the time required to discharge the sludge from a pipet through a capillary tube with the time required to discharge an equal volume of water at standard temperature, 20” c. Sample settling curves are shown in Figure 1, curve A being for particles whose average dimension was 0.00174 4
M I. T.Thesis, 1925.
