T H E J O U R N A L OF I N D U S T R I A L A N D ENGINEERING C H E M I S T R Y ation with cold alcohol t o exhaust t h e drug, a n d hence t h e hot maceration is advisable. @Six different samples of Podophyllum were assayed by this method, a n d from t h e same samples fluidextracts were made according t o t h e directions in t h e United States Pharmacopoeia, and were also assayed. T h e results are given below: Grams resin in 100 grams drug 1 .......................... 5.03 5.24 2 . . ........................ 3 .......................... 4.65 4 .......................... 4.92 5 .......................... 4.73 6.82 6 ..........................
Grams resin in 100 cc. fluidextract 4.96 5.16 4.66 4.89 4.83 6.71
These results show t h a t t h e method gives very concordant results on t h e assay of t h e drug a n d fluidex-
I
Vol. 6, No. 8
t r a c t a n d much higher results t h a n t h e precipitation method. If t h e method gives equally good results in t h e hands of other workers, t h e n i t would be advisable t h a t fluidextract of Podophyllum, U. S. P., be assayed a n d standardized. As good lots of Podophyllum drug contain j per cent of resin, a standard of j per cent resin for t h e fluidextract is suggested. T h e method can also be applied t o t h e assay of solid a n d powdered extracts of Podophyllum by dissolving weighed quantities of t h e extracts in sufficient alcohol t o render t h e solutions of a b o u t t h e same strength a s fluidextract. ANALYTICAL DEPARTMENT PARKS,DAVIS& Co., DETROIT
LABORATORY AND PLANT STUDIES ON FILTRATION’ B y J. W. BAIN AND A. E. WIGLE
I n connection with factory operation quite recently, one of t h e authors h a d t o form a n estimate, in advance, of t h e amount of moisture which would be retained by a finely divided solid on a vacuum filter. A search among t h e usual sources of information yielded no serviceable d a t a . When t h e filters were in actual operation, their performance in this respect was very much better t h a n h a d been anticipated, a n d had this fact been known in advance some economy in construction might have been effected. With a view t o gaining information on this point, t h e authors investigated t h e literature a t their disposal, a n d with t h e exception of t h e interesting a n d valuable paper b y Hatschek,2 t h e y were unable t o find a n y useful d a t a . When t h e experimental work had progressed t o a certain extent, an accident drew our attention t o t h e exhaustive monograph of King a n d Slichter, “ Principles a n d Con’ditions of t h e Movements of Ground Waters,”3 from which we have drawn freely in this discussion. I n t h e problem which is here under investigation, the solid is assumed t o be bathed by a liquid in which it is insoluble, such as, for instance, t h e mother liquor of a crystalline magma. It is proposed, therefore, t o investigate t h e amount of liquid retained by a mass of finely divided solid when filtration is carried out under atmospheric or other pressure a n d also in t h e centrifuge. T h e experimental work was considerably simplified by t h e condition laid down above, which permitted t h e use of a solid insoluble in water. A q u a n t i t y of pure well-rounded lake sand was carefully sieved, a n d t h e grains which were retained on t h e 4 0 mesh screen b u t which passed t h e 30 mesh, are referred t o throughout as 40 mesh sand. T h e screens used were not of very good quality in t h e regularity of t h e mesh opening, as will be seen from t h e d a t a given later, b u t this point is of no particular significance in this investigation. 1 Presented at the 6th Semi-annual Meeting of the American Institute of Chemical Engineers, Troy, New York. June 17-20, 1914 2 J . SOC.Chem I n d . . 1908, p . 538. 8 Nineteenth A n n Report, U.S. Geol. Survey.
The r a t e of flow of a given liquid under a constant head through a filter-mass of a finely divided solid will obviously be dependent upon t h e amount of space which is not occupied by t h e grains, i. e., what is commonly called t h e “pore space.” On first consideration, i t would appear t h a t t h e pore space would vary a good deal according t o t h e size of t h e grains composing t h e mass, a n d t h e results of computation a n d experiment are a n astonishing contradiction t o this idea. The pore space is almost independent of t h e size of t h e grains, a n d t h e arrangement of t h e l a t t e r is of chief importance. By considering a number of small spheres of uniform diameter packed as closely as possible in a given space, it is possible t o arrive a t a mathematical formula from which t h e pore space may readily be calculated. Slichterl has shown t h a t if t h e spheres are so arranged t h a t their centers lie a t t h e corners of a cube, t h e pore space will be 47.64 per cent; while if t h e centers of t h e spheres lie a t t h e corners of a rhombohedron which permits t h e closest possible packing, t h e pore space is 2 5 . 9 5 per cent. Between these limits we may expect t o find t h e porosities of all ordinary materials. With actual materials, in t h e case where t h e grains are of approximately equal size,. t h e pore space a n d also t h e diameter of t h e particles may be readily determined b y counting a number of t h e grains, determining their combined weight a n d the specific gravity of t h e material; t h e t o t a l volume may be ascertained b y adding t h e sand in small quantities t o a cylinder, tapping gently with a flat-faced pestle until no further decrease in volume takes place. T h e results of this procedure on our sands are presented in Table I. Mesh screen
No.,of grains
30 40
(400 ,,,
50
{%I
60
80
{:E {igg
TABLEI Total One grain Sp. Pore space Diam. wt. grm. grm. X 10-6 gravity per cent Mm.
::;:::
~
2:; ]
~
:
2.74
~3 5 . 4 ~0 . 4 2 0] 34.1
0.354
E;:; :;:I
2.68 2.73
36.4
0.318
o,0172 0 .0238 0 ., 0 210526 o
2.82 2.85
36.8 37.7
0.269 0.257
o,0253 0.0251
::) :;)
The comparatively slight variation in pore space 1
LOG.cit.. p. 309.
A -VD E ATGIAVE E RI N G C H E M I S T R Y is worthy of note; a n d i t m a y be added a t this point t h a t mixtures of small a n d large grains show a surprising similarity in their porosity t o t h a t of either t a k e n alone. For all practical purposes, t h e pore space of masses of crystals, such as are commonly produced b y rapid cooling, may be placed a t 3; per cent of t h e t o t a l volume occupied. PI L T R A T I 0 N
N D E R A T M 0S PH E R I C P R E S S U R E
673
of a mass of grains in a filter, becomes i m p o r t a n t when t h e question of washing away a n impure mother liquor has t o be considered: ,4 series of experiments was performed with t h e object of ascertaining t h e a m o u n t of water retained in t h e sands a t different levels while under vacuum. T o carry this o u t , a t u b e a b o u t 80 cm. long a n d provided with side t u b e s closed with corks a t I O cm.
This p a r t of t h e subject has been so carefully worked o u t b y King' t h a t i t suffices t o reproduce some of t h e results, slightly modified t o suit t h e present purpose. Cylinders 8 feet long, j inches i n diameter, were filled with special, sorted sands, wire gauze being used as a support a t t h e b o t t o m . Water was introduced from below, a n d when t h e t u b e s were full, percolation was allowed t o commence, a n d t h e water which drained a w a y was collected a n d weighed a t intervals. T A B L EI1 Effective size of grains M m . 0.4745 0.1848 0.1551 0 . I183 0.0826
Water retained--per cent of d r y sand Pore space Per cent ,38.86 40.06 40.76 40.57 39.77
I
After 1 hour 11.23 12.72 14.73 19.30 20.15
After 9 days 4.24 5.05 7.25 9.41 11.82
/o
2
3
4 5113 4 Percent cf mojsture
5
6
F I L T R A T I O N IVITH V A C U U M
Experiments were carried out b y t h e authors with t h e idea of approximating t o factory conditions. T h e s a n d was poured into a Buchner funnel provided with a piece of wire gauze, a n d gently t a m p e d down with a flat-faced pestle; t h e d e p t h of t h e layer was inches. T h e t ? p of t h e funnel was closed b y a glass plate ground t o fit a n d provided with a central aperture through which air could be admitted. T o avoid t h e error of surface evaporation during filtration, this air was drawn through a tower, down which water trickled slowly. T h e funnel was placed in a suction flask a n d a simple gauge enabled t h e vacuum t o be read. When t h e sand h a d been under v a c u u m f o r a given period, i t was thoroughly mixed a n d a sample removed; water was once more poured on a n d t h e vacuum was maintained for a longer period. T h e results are given in Table 111. Mesh screen 30 40 50 60 80
L O C . Lit.
T A B L EIV-PERCENTAGEMOISTUREAT ENDOF 15 M I N U T E S Depth of sample from top in cm. Mean Pressure Mesh In. 7-per cent 70 moisture 20 30 40 50 60 screen mercury IO 3.97 3.13 3.78 3.77 4.28 3.56 4.13 4.90 30 4.5 4.60 1 . 8 2 4 . 1 0 4 . 3 3 4.08 4 . 0 8 5 . 0 8 6 . 4 5 40 4.0 5.30 3.90 4.40 5.00 4 . 8 0 5.30 5.60 7.60 50 6.5 5.30 4 . 1 4 4.95 5 . 3 2 5 . 2 0 5 . 2 0 5 . 6 3 6 . 6 0 60 7.0 1
30 40 50 60
-
~~
T A B L EV-PERCENTAGE MOISTUREAT 4.5 2.84 3.08 3.10 3.31 3.03 3.24 3.30 3.76 5.0 6.5 3.40 4.00 4.50 4.65 3 . 5 8 4.37 4 . 5 6 4 . 6 0 7.0
ENDOF 30 MINUTES 3.35 3.19 3.51 4.56 3.85 3 . 8 3 4.20 5.13 4.70 4.95 5.00 6.50 4.70 4.60 4.65 5.75
These results were plotted a n d curves were drawn as shown in t h e accompanying illustrations. The
T A B L EI11 Moisture a t the end of
_.--_
5 min 7.20 8.20 8.65 8.42 9.15
7
15 min. 5.69 6.84 2.50 i .38 7.52
30 min. 4.75 5.19 6.41 6.90 7.37
Vacuum
I n . mercury 1.5 1.75 0.75 2.0 2.25
I t is seen from these results t h a t t h e moisture content increases inversely as t h e diameter of t h e grains of sand. I n each experiment t h e water p u m p was worked at full capacity, a n d as might be predicted, t h e vacuum increases slightly as t h e size of t h e grains decreases. B y way of comparison, a single experiment with s a n d of mesh jo m a y be quoted. Water was poured on t h e layer a n d no v a c u u m was used; after I j minutes' standing, t h e moisture content was found t o be 27.4 per cent against t h e 7. j o per cent under vacuum. T h e a m o u n t of liquid retained b y different portions 1
intervals, was filled with each sand, a n d connected a s has been described in t h e case of t h e Buchner funnel. A powerful water p u m p was r u n t o full capacity a n d t h e pressure, as before, varied with t h e size of t h e grain. T h e results are given in Tables IV a n d V.
Percent
+ moisture
individual points were sometimes decidedly off t h e curves, b u t although t h e experiments were repeated n these cases, no better agreement could be obtained;
T H E J O U R N A L OF I N D U S T R I A L A N D ENGINEERING C H E M I S T R Y
674
t h e accurate determination of small amounts of moisture in these sands proved t o be difficult, probably owing t o sampling. T h e average per cent of moisture was determined by measuring t h e areas under t h e curve, a n d dividing this by t h e height which gave t h e width of t h e rectangle of equal area. FILTRATION
W I T H CENTRIFUGE
This well-known method of separating solids from liquids was next subjected t o test for t h e sake of comparison with t h e previous experiments. A small hand centrifugal, 41/4 inches inside diameter, was used; i t could be run a t jooo r. p. m. without inches a n y trouble. A cylinder of wire gauze, in diameter, was placed over t h e axis of t h e machine a n d t h e sand was poured i n t o t h e annular space t h u s formed; t h e layer had, therefore, a thickness of 1'/4 inches which was t h e same a s t h a t in t h e Buchner funnel. As a preliminary expefiment, t h e sand was thoroughly wetted, a n d t h e centrifugal run a t 2 0 0 0 r. p. m. for 2 minutes. T h e percentages of moisture are given in Table VI. TABLEVI--PERCENTAGES OF MOISTURE Centrifugal 2 min. Mesh screen Vacuum 15 min. 50 7.50 2.56 60 7.38 2.55
T h e marked efficiency of t h e centrifugal is noteworthy a n d t h e method of procedure was altered t o show this more forcibly. Sand was placed in t h e Buchner funnel, wetted a n d vacuum applied for j minutes. After sampling, t h e s a n d was placed while still moist, in t h e centrifugal which was t h e n run for z minutes a t 2 0 0 0 r. p. m. Table VI1 shows t h e percentages of moisture. TABLEVII-PdRCENTAGES Mesh screen Vacuum 5 min. 30
t E:
40 50
i9 . 3 5
60
18.40
{?:E
80
OF
MOISTURE Centrifugal 2 min. 2.26 2.20 1.93 2.30 2.56 2.80 2.36 2.65 2.49 2.46
I t is seen from t h e above results, t h a t t h e moisture content under vacuum varies inversely a s t h e diameter of t h e grains; t h e moisture content after centrifuging, however, is nearly t h e same for t h e finer as i t is for t h e coarser sands. T h e distribution of t h e water a t several points in the annulus of s a n d was also investigated a n d Table VI11 presents t h e results in percentage of moisture. TABLE VIII-PERCENTAGE Mesh screen 40 50
OF MOISTURE Distance from,,center of basket 1 11/?" 2.9 2.72 2.43 3.0 2.90 2.76 '/2"
T h e variation, while sufficient t o permit measurement, is small a n d might be neglected for practical purposes. T h e objection may be raised t h a t these results, obtained in t h e laboratory with a small centrifugal, are of little value for comparison with t h e larger machines used in t h e factory. While with t h e hand centrifugal, t h e diameter is small, t h e speed is high,
Vol. 6 , No. 8
a n d we have calculated t h a t a weight of I lb. revolving a t a z inch radius a t 2 0 0 0 r. p. m. is subjected t o practically t h e same centrifugal force as a weight of I lb. revolving a t a radius of 1 2 inches a t 600 r. p. m. The comparison is, therefore, justifiable a n d a good idea of t h e behavior of a moist mass when centrifuged in t h e factory, may be obtained beforehand in t h e laboratory. Using t h e formula given by Griscom,' we have calculated t h e pressure as t h e periphery of t h e 4l/4 inch centrifugal running a t 2000 r. p. m. a n d find it t o be 7.66 lbs. per sq. in. THEORETICAL CONSIDERATION
Hatschek? has discussed t h e behavior of very finely divided substances on t h e filter, a n d has pointed out t h e value of a microscopic examination in this connection. The probable arrangement of t h e particles, with respect t o the pores of t h e septum, are pointed o u t , a n d t h e influence of t h e flexibility of t h e l a t t e r is taken i n t o codsideration. T h e retention of small quantities of liquid in t h e mass of fine grains is due, undoubtedly, t o capillarity. T h e extraordinary difficulty in removing t h e last few per cent is well known a n d is again set forth above. I n considering t h e reasons for this, it seemed t o be worth while t o calculate what would be t h e thickness of t h e film, if all t h e residual water were assumed t o be distributed uniformly over t h e superficies of t h e grains. For this purpose, sand of 30 mesh with 6 per cent moisture was selected; t h e thickness of t h e film of water on each grain was found t o be 0.011 6 mm. It would be interesting t o calculate what stress must be applied t o a grain t h u s coated, t o overcome t h e surface tension of t h e liquid in so far as t o allow t h e removal of a t least p a r t of t h e water; such a computation, if i t could be effected, might furnish a scientific basis for t h e prediction of t h e behavior of finely divided solids on centrifuging. T h e authors have been unable t o find time t o carry this o u t , b u t hope t o do so in t h e future. . T h e above discussion assumes t h a t all t h e water is present on t h e superficies of t h e grains, b u t t h e capillary action of t h e small spaces between t h e grains is undoubtedly of great importance. I n t h e case of t h e sand just quoted, which has a pore space of 35.4 per cent, t h e moisture present would fill 30 per cent of this; t h a t is, 7 0 per cent of t h e pore space is filled only with air. This gives some idea of t h e comparatively poor performance of t h e ordinary filter a n d of t h e vacuum filter; in each case, air channels form a n d t h e downward pressure on t h e water-filled pores is t h u s relieved. I n t h e case of t h e centrifugal, each particle of water experiences practically t h e same stress, a n d only t h e capillarity of t h e finest pores a n d t h e surface tension of t h e films on t h e grains are sufficient t o resist its action. SUMMARY
I-The pore space in a mass of fine grains averages a b o u t 37 per cent of t h e total volume. 1 2
Metal. and Chem Eng.. April, 1913. Loc cit.
AW.9 1914
T H E J O I:R N A L O F I N D 17s T R I A L A ATD E N G I N E E RI N G C H E M I S T R Y
675
1
a-The a m o u n t of water retained when a n ordinary filter is used varies from 11 per cent, with 2 0 mesh material, t o 20 per cent with I O O mesh material, one hour being allowed for drainage. 3-The a m o u n t of water retained on a filter with z in. vacuum averages 7 per cent after I j minutes for material varying from 30 t o 80 mesh. 4-111 a layer of material 7 0 cm. deep on a filter, with j in. v a c u u m , t h e t o p layer will average, after I j minutes, 4 per cent moisture, a n d t h e b o t t o m 6.5 per c e n t ; t h e size of t h e grains is n o t of importance within t h e limits discussed. ' If t h e vacuum be maintained for 15 minutes longer, t h e above figures will be reduced b y another half per cent. j-By t h e use of a centrifugal, t h e percentage of moisture, i n all t h e materials employed, m a y be reduced t o a n average of 2 . 5 per cent. 6-In t h e case of a s a n d of 30 mesh with 6 per cent moisture, if all t h e water be distributed over t h e surface of t h e particles, each grain would h a v e a film 0.0116 mm. t h i c k ; or t h e water would fill 30 per cent of t h e pore space. FACULTY OF
factory. Incidentally, i t m a y be s t a t e d t h a t all eft'qrts t o get t h e hard rubber people t o provide a material which would s t a n d u p against cold concentrated sulfuric acid have been futile. Hence, i t m a y be well t o warn others as t o their claims in this respect. A pear-shaped screen of copper gauze was placed in t h e opening of Tower B leading t o Tower A t o prevent clogging, in t h e event a rushing action of t h e p u m p sucked pieces of t h e pumice. We draw off a n d replenish t h e acid once or twice a year. Tower A is half filled with angular pieces of commercial caustic soda, in size from a hazel n u t t o a n egg. T h e mass rests upon a copper wire gauze screen
APPLIED SCIENCE
UNIVERSITY OF TORONTO TORONTO, CAXAIIA
SCRUBBER FOR CHEMICAL LABORATORY VACUUM SYSTEM' BY CHARLESBASKERVILLE
I n order t o protect t h e vacuum p u m p of our laborat o r y from t h e corrosive action of t h e gases drawn therein, I devised t h e installation described herewith. T h e pump-an improved Packard Vacuum P u m p , 2 cylinder, I a in. diam., motor-belt driven-has been in more or less continuous service for seven years without a n y expenditure thereon for repairs, as a result of this protection. It seemed safe, therefore, t o present a n account of it. T h e installation is a n application of t h e simple principles usually applied o n a small scale with glass a p p a r a t u s in t h e laboratory. T h e towers are made of cast iron, porcelain-lined, a n d set into t h e system with a by-pass, which we have used only during t h e short time necessary for recharging. T h e towers are connected b y hard rubber pipes ( a in. internal diameter). A t t h e b o t t o m of each tower is a h a r d rubber drain cock, bolted t o a flange. At t h e t o p of B a n d C are plates bolted t o flanges, which m a y easily be removed. T h e opening is of sufficient size t o a d m i t dropping a s t r u n g incandescent bulb for inspection. Tower B is three-quarters filled with pumice stone in egg-size pieces. T h e pumice is thoroughly saturated with concentrated sulfuric acid. I believe lead pipe would be better in this cylinder as t h e h a r d rubber softened on contact with t h e acid. So far, however, t h e weight of t h e pumice a n d acid has not been sufficient t o cause t h e h a r d rubber pipe t o collapse. T h e decomposition of t h e rubber compound became so pronounced with t h e drain cock in a short while t h a t i t was replaced b y a lead plate, which has proven satisPresented at the 6 t h Semi-annual Meeting of the American Institute of Chemical Engineers. T r o y , Xew York. June 17-20, 1914.
SCRUBBER FOR CHEMICAL
A-Caustic
LABORATORY VACUUM
Soda, B-Sulfuric
SYSTEM-SCALE,
1
IN. =2
FT.
Acid and Pumice; C-Trap
supported on a n d b y t h e tapering bottom of t h e tower. T h e drain cock admits of drawing off a n y liquefied caustic which m a y accumulate. A metal pipe leads from t h e t o p t o t h e pump. Tower C is a safety reservoir t o catch t h e fluid f r o m B i n t h e event of a leak beyond or other cause for increase in pressure on t h e p u m p side in t h e line. So far, no indication of its real need has been apparent, as t h e maximum a n d minimum contacts of t h e automatic regulator of t h e motor have never failed. A gauge in t h e system beyond t h e scrubber serves, by comparison with t h e gauge on t h e p u m p , t o show leaks i n t h e scrubber. Xone was observed until t h e h a r d rubber drain cock on Tower B failed, a n d none
