986
THE JOURNAL OF INDUSTRIAL A N D ENGINEERING CHEMISTRY
drying of the cloth concentrates such caustic as is present. Wool is the poorest material to use on caustics of any strength. This should be remembered where a cake filtered from an acid liquor, in which wool is a very good medium, is washed with a caustic. Strictly speaking, the world’s best, acid filter medium is silica, or other inert compounds as carborundum, alundum, etc. Filtros, a porous €used silica, is typical of this class of material. Mechanically these media do not lend themselves as well to the types of industrial filters most widely used. They are sometimes faulty on account of their lack of uniform porosity and the possibility of solids penetrating the surface, never to be removed. JlETALLrC
CLOTHS
One of the factors of safety provided in this weave is that any imperfection due to faulty workmanship can be reduced by rolling the cloth between heavy cylinders, thus closing up the imperfection. Much criticism has been leveled a t this rolling, on account of the injury to the wires. If the metal is soft>enough the rolling has only a small, if any, deleterious effect. If heavier wires of improperly annealed material are used it is quite evident, that rolling is a poor expedient. The improvements in twill weave instead of square weave, strength proportioned to the warp members, monel metal for iron, etc., are later-day improvements making the cloth a better medium. The wire cloth company who turned out the first commercial cloth later perfected a weave from the old Dutch cloth of commerce which has proved to offer some striking advantages. Much heavier wire can be used and the smooth finish of its surface as well as the evenness of its weave are some of its commending features. In some liquors the metal is slowly attacked so that; its life is definite. For such work there can be no discussion as to the kind of metallic cloth to use. The wire of maximum crosssection is desired and the cloth using it should consequently be selected.
CLEANINGOF FILTER CLOTHS The filter medium often becomes fouled as a result of incrustation, either from handling supersaturated solutions or from precipitation caused by lowering the pressure of the liquor. I n order that the porosity shall be maintained sufficiently to obtain production, the cloth must be cleaned with an agent that will dissolve the incrustation. This is particularly true of metallic filter cloth used on caustic liquor containing calcium compounds as precipitates. ’ The use of an acid, such as hydrochloric, to remove these incrustations is, of course, fraught with danger to the cloth. In most instances the incrustation is unnecessary. Calcium carbonate will often deposit because the liquor contains bicarbonate. If the temperature had been raised and held a t the boiling point the bicarbonate would have broken down to normal carbonate. This is very evident in beet sugar manufacture. There is a safety provision for even these liquors as they are generally handled. In closed outlet filters all that is required is that there shall be a back pressure on the medium above the actual point of precipitation. I n practice this back pressure can take the form of a pipe delivering the filtrate to an outlet some feet above the filter. I n a magnesia plant the scale that formed on the vertical pipe required a change of a section of the pipe each week. This, however, was a great improvement on having the filter cloth plugged up with this deposit. Of course, this remedy is not applicable to suction filters. I n this case pretreatment is the only outlet, and where this is not feasible a different type of machine is probably the solution for successful handling of this material.
Vol. 13, No. 11
Fundamental Laws of Filtration with Suggestions Regarding Research W o r k By D. R. Sperry 119 McKFE ST., BATAVIA, ILL.
[The material covered by the first part of this paper will be found in Chemical and Metallurgical Engineering, 15 (1910), 198.1
The fundamental law of filtration developed in this article is:
The discharge formula under constant rate of flow conditions may be derived from (1)as follows: If 9 = constant = M dT
P
+RAI K 2K = W or if -
R%
p =-2QM f R m M a n d Q =
E-
W 2M 2 where Q = flow of liquid; P = pressure; T = time; K = rate of deposition; R = resistance; yo = per cent of solids; Rm = resistance of filter base. 2K Using the terms W for - and N for KRm - the equation for
R% R %’ discharge under constant pressure conditions, developed in the original article, may be expressed as Q
=.\I WPT +NZ-N.
SUGGESTIONS AS TO RESEARCH It should be borne in mind that these expressions are for rigid solids. There should be an investigation as to the effect of pressure upon nonrigid solids. It will of course change both K and R, and to a certain extent R,. A table should be made showing the value of K for different substances, and a careful record made of size and shape of the suspended particles, etc. A table should be made showing the value of R for different substances with full data regarding particles, etc. An investigation should be made of filter cloths. The value of R, should be found for various weaves and weights of cloth. Measurements should be made of the strength of various cloths, their coefficient of shrinkage, their ability to stand acids and alkalies, their ability to wear under ordinary treatment, leakage a t beginning of runs, etc. The effect of viscosity upon the fundamental laws should be studied, along with the effect of change in temperature. The effect of sedimentation should be studied in so far as i t influences filtration. The effect of various filter-base supports should be studied. The effect of consolidation of adjacent cake formations should be studied, and the effect upon the time-discharge curves. The effect caused by the addition of different alloys to various mixtures should be investigated. A study should be made of filter paper and a table made similar to that made for cloths. Filter bases not made of cloth or paper should be studied. The effect of feeding filters with different kinds and sizes of pumps should be investigated.
The Feeding of Filters By J. F. Springer 618 WEST 136TH ST., N F W YORH,N. Y.
Feeding may be defined as consisting of the transmission under pressure of the unfiltered liquor from a point where it is received from storage to the inlet aperture of the filter.
Nov., 1921
.
T H E JOURSAL OF INDUSTRIAL A X D ENGIKEERIKG CHEMLSTRY
The necessary equipment will accordingly consist: (1) of the transmission pipe line; (2) the pump or other forcing device which directly or indirectly brings about the movement through the transmission line and effects the pressure of the unfiltered liquor against the filtering medium; and (3) the power apparatus through which or by which the drive of the forcing device is effected. TREATMENT OF THE RAWLIQUOR CHILLING-Filtration has for its object the physical separation of solids from the liquid proper. It is highly desirable, however, other things being equal, that the raw liquor be kept in the form of a simple solution while passing through complicated parts of the feeding equipment. Certain classes of filters are competent to deal with certain solutions. Such filters effect precipitation in the immediate vicinity of the filtering media, so that the raw liquor, which enters the filter as a solution, consists a t the moment of filtration of a liquid in which solids are suspended. This transformation may be effected, for example, in the standard filter press by providing for the circulation of a chilling brine back of and between the working surfaces of the plates. The incoming solution is chilled as it approaches thefilteringmedium, and soluble matter contained in it undergoes precipitation. ADDITION OF PREcIPITANT-It seems to the writer that another method might be used in the press for securing similar results. A great many substances are precipitated from solutions by the addition of certain precipitating substances, which may be solutions, simple liquids, or solids. It might be possible to introduce such a substance into the filter press a t points between leaves or elsewhere. The conversion of the solution into a liquid containing suspended solids would occur shortly before arrival at the filtering medium. The importance of maintaining the condition of solution up to actual entrance into the filter centers largely on the relief this gives to pump and transinission line. One of the greatest problems encountered in feeding is the maintenance of valves and other complicated parts free from deposits of solid matter. It is not absolutely necessary that the transformation from a solution to a liquid containing suspended solids be effected inside the filter. In many cases the transformation point may probably just as well be located a t any convenient point on the transmission line between the filter and the nearest valve or other part subject to clogging. This may call for some changes in the character, dimensions, and placing of that part of the transmission pipe between the transformation point and the filter. REMOVAL OF PRECIPITATZ-TO carry this point furthe?, there are probably a few, possibly many, raw liquors, in which certain unwanted solids are already in suspension, which may be reduced to simple solutions for passing through that part of the transmission line where clogging is possible. This may often be effected in one of two ways or in a combination of the two: The suspended solids may consist entirely or wholly of material that may be made to settle, it may be possible to dissolve the solids, or both results may be brought about. When such measures are possible the passage through the pump and valve section of the transmission line is wonderfully facilitated. Actual practice today would, upon adequate search, probably show many instances where the foregoing program is partly carried out, more especially the removal from the raw liquor of sedimentary matter which can be settled quickly. Whether the second method of treatment is used, the writer is unable to state. The means of facilitating deposition of sediment will usually involve the provision of convenient settling basins. ,4nother method makes use of the centrifugal idea. The liquor containing sediment is made to swing round sharp
987
curves or is passed in and out of a machine which causes it to revolve. If the particles in suspension are really of' greater specific gravity than the liquid and if the retardation effected by the liquor is not too great, the particles will seek points away from the center of rotation. The same general method has been used in actual practice in freeing air of dust. KOXCORROSIVE CoNRTRUCTION Some of the raw liquors that must be transmitted in presentday practice will not, however, be made harmless by the measures just described. Some of them are highly corrosive. The obvious answer to this difficulty is to construct all parts in pump or transmission line coming into contact with the corrosive liquor from Eoncorrosive materia!, or at least line them therewith. Lead is a favorite metal for linings and for pipe lines. Tt interacts very slowly with most substances, as, for example, ordinary sulfuric acid. However, thwe are limitations to its use. For instance, rain water reacts with lead to form the soluble hydroxide, Ph(0H)z. Xitric acid attacks lead. Concentrated sulfuric acid (sp. gr. higher than 1 700) attacks lead even in the cold. With hydrochloric acid, lead reacts slowly. It is obviously ill-adapted to the construction of moving mechanical parts, because of the lack of the necessary hardness and tensile strength. Wood resists acids splendidly, but is also unsuited for many moving parts. Recently, hard rubber has come into consideration. It is highly resistant to many corrosive activities, and possesses considerable degrees of hardness, aiid tensile and compressive strengths. It is also resistant to alkalies. Cast iron enjoys a good reputation as a material resistant to acids and alkalies. It successfully resists the boiling acid solution of magnesium bisulfite employed in the manufacture of wood pulp. It has been used successfully as the material for the shell of a monte-jus through which is passed a mixture of 40 per cent nitric acid and fuming sulfuric acid. A montejus made of wrought iron is mid to have been completely destroyed in 24 hrs. Upon exaniiiiation of the interior of the cast-iron monte-jus after what appears to have been a fair period, no trace of corrosion is said to have been found. A special iron-like material which is well adapted for use in filters appears to be a kind of silicon-iron. Manganese, phosphorus, and sulfur are also present, though probably only in minute quantities. The material seems to be essentially a gray cast iron with a high siljco, content. This material has been subjected to prolonged cold and heat tests by the Rureau of Standards. The cold tests were carried out a t a temperature of 15" to 20" C . for about one-third of a year, and the depth of corrosion for the whole year calculated. A few of the results are fairly representative. Sulfuric acid of 95 per cent strength had only the most minute effect, calculated as a corroded depth of 0.0000206 in. in 1 yr. The diluted acid showed greater action, but even the 10 per cent solution effected a depth of corrosion less than 0.0001 in. Nitric acid was still less effective. A 25 per cent solution of hydrochloric acid effected corrosion a t the rate of 0.00805 in. per yr. Acetic, phosphoric, oxalic, pyrogallic, oleic, and picric acids were also tested. Other corrosive liquids, such as aluminium potassium sulfate, ammonium chloride, ferric chloride. and copper sulfate, had only insignificant effects, the maximium being less than 0 0001 in. per yr. A 30-day test a t 82" to 88" C. naturally produced greater results. With sulfuric acid, the 10 per cent solution effected corrosion a t the rate of 0.00156 in. per yr. and the more concentrated solutions a t diminished rates. Nitric acid in the higher concentrations effected corrosion at rates m7ell under 0 0005 in. per yr. The 10 per cent solution had more effect. The 25 per cent hydrochloric acid ate into the metal at the rate of 0 366 in. per gr.
988
T H E JOURNAL OF INDUSTRIAL A N D ENGINEERING CHEMISTRY
PRESSURE REQUIREMENTS All filtration is by pressure, whether the pressure a t the filter medium is effected directly by the creation of a positive pressure on the raw liquor on the precipitate side, or whether it is secured indirectly through the removal of the natural air pressure on the filtrate side. I n the latter case, 14.7 Ibs. per sq. in. would appear to be the maximum attainable under ideal conditions. In elevated regions even this theoretical maximum would be reduced. However, the vacuum method is doubtless vigorous enough for multitudes of cases. Direct pressure may be effected through the force of gravity, by the drive of pumps, and by the tension of compressed air. Still another method, which is said to have been introduced into small filters and in laboratories, depends upon the creation of a centrifugal pressure on the raw liquor, causing filtrate and precipitate both to press against the filtering medium. As the precipitate cannot get through and the Sdtrate can, the ordinary filtering separation occurs. It is precisely the same phenoinenon as occurs under more usual circumstances. GRAVITY FILTRATION-T~~S requires pumps or an eqliivalent to put the raw liquor a t the proper elevation. Except for the considerable losses through friction, etc., the energy expended in elevating the raw liquor is regxerted in effecting the filtration. Theoretically, in accordance with theory based on a feu! assumptions as to conditions, each foot of elevation will bring about an increment of pressure per sq. in. amounting to the product of the specific gravity of the liquor by the weight of 12 cu. in. of water. This principle may be stated as: P = 0.434 X s X E = 0.434 3 s where P = pressure ia lbs. per sq. in., E the elevation in ft., and s the specific gravity of the raw liquor. This formula disregards friction, viscosity, leakage, etc. It may serve, however, as the basis for a broader formula containing a factor to be determined by experiment and then to be used tentatively as a single independent variable is varied. Used as it stands for highly concentrated sulfuric acid, for example, it means that a 50-ft. head would produce a pressure of 40 lbs. per sq. in. That is, if we assume the highly concentrated solution of 168" Twaddell, the formula gives: P = 0.434 X 50 X 1.840 = 39.9 lbs. per sq. in. While this formula does not take into account all the conditions that obtain, it is nevertheless useful a t the present moment in showing us that considerable elevations will be required, even with heavy liquors, if one is seeking high pressures. The effect of the factors not included in the derivation of the formula would generally tend to the increase of E required to produce a given pressure. To get the maximum pressure required in filtration operations, usually around 150 lbs., involving such a liquor as sulfuric acid, the tank would have to be set a t a height sufficient to enable the level of the contained liquor to be maintained a t about 1S7.5 ft. above the filter. It should be remembered in this connection, that high elevations, while they do incur the construction and maintenance of towers, do not reduce the power requirements, since all pressure that gravity exerts in the filtration operation must be expended in putting the raw liquor a t the necessary elevation. However,. gravity feeding provides for a steadiness of flow and a steadiness of pressure that are valuable. By a sufficient increase in the plan section of the tank, the pressure fluctuation may be reduced almost to zero. For example, if we assume a tank to be 6 ft. in diameter and to contain as heavy a liquor as concentrated sulfuric acid, a variation between delivery of liquor to the tank and withdrawal for feed would have to amount to 2 . 3 6 cu. ft. to bring about a 1-in. change in level. This 1-in. change in level, moreover, would cause merely a pressure change of a trifle over 1 OZ. (1 -06 oz.) per sq. in. It will be seen, then, that with generous
Vol. 13, No. 11
horizontal sections, a wonderful steadiness of pressure may be produced. Of course, there will be no pulsation from the pumps or other means of delivery to the tank. Naturally, the pipe line from tank to press must be sufficiently generous in size and so moderate in the number and sharpness of its bends as to insure a good flow of liquor. Furthennose, an advantage to be credited to gravity feeding is that, for moderate pressures involving moderate elevations, it will often be possible to locate the filter in the basement and the gravity tank in the attic or on the roof, thus obviating the construction of a special tower. Pums-The types of pumps usually employed in connection with feeding are: I-Steam pumps 2-Power-driven reciprocating pumps 3-Power-driven rotary pumps (including centrifugal)
Any of these may be direct-connected to the filter press or open filter or gravity tank. I n the case of the steam pump it is necessary t o consider whether heat is permissible in the particular case. One of the principal builders of filter presses favors the steam pump for all-round service. The speed is under a nice control, and the equipment is inexpensive to purchase and requires but little space for its installation. It is said to be "fool-proof," in that the ratio between pump and pistons may be so selected that the pump will stall a t a predetermined pressure. This is a point of especial value, where the steam pump is directly connected with the press and its pressure is directly transmitted to the plates. Where a power-driven pump is employed it will generally be possible to introduce on the transmission line between pump and filter a specinl safety valve with pressure indicator and weighted control lever. This valve may in a simple manner be set a t a predetermined pressure and thus provide relief inside the press, particularly a t the working surfaces of the plates. Ordinarily, the power-driven pump will be without speed control. This may be corrected, however, by using a variable-speed electric motor. These are more or less expensive, and it would seem that, so far as protection against the hazard of breaking plates in the press is concerned, the safety valve would cover requirements. It may be worthy of note that some concerns using pumps for feeding glycerol, grease, oils, and lard have apparently found a duplex pump suitable. Another concern is reported to be using the same design of pump for handling sulfate of manganese. This pump is fitted with ball valves. When glycerol, grease, or oils are in question, bronze and brass may be used in various parts liable to suffer otherwise. A power-driven rotary pump is often an inexpensive form of feeding unit, but it lacks ease of control. Sometimes, a valve is placed between pump and press, the object in view being the cracking of the valve when a predetermined pressure has accidently or otherwise been built up. Objection has been made to this form of control on the ground that the valve is liable to clogging from deposited solids. COMPRESSED ArR-The monte-jus is the equivalent of a pump in which operation is secured through the pressure developed against the liquor by compressed air. There are two general types, one being hand-controlled and accordingly requiring personal attention, the other being automatic. I n a representative monte-jus there is a case or tank having several openings above. For instance, immediately in line with the vertical axis may be a central opening through which compressed air is admitted. T o one side may be another opening, through which the liquor is admitted to the tank. On the opposite side may be the discharge opening, through which the compressed air a t the proper moment forces the liquor. This opening may have passed down through it a prolongation of the discharge pipe. We will now assume
Nov., 1921
THE JOURA’AL OF INDUSTRIAL AND ENGINEERING CHEMISTRY
989
FILTRATION OF RIGID PARTICLES Problems of the first class are the easiest to handle. It is necessary merely to choose a proper filtering medium, either cloth or screen. Usually the medium has a sufficiently fine mesh to retain all of the suspended solids from the beginning of the filtration period, but a t times it is advantageous to use a medium with larger openings. I n this case the suspended solids are depended on to build up, and the cloth or screen becomes, more properly speaking, the retaining medium. The filtration proper starts only when a properly arranged layer of larger sized particles has been retained! p?eventing the finer particles from passing through and giving true filtration. I n many problems in filtration the suspended solids have been produced by precipitation and crystallization. It is an advantage to secure the largest possible crystals, in such cases, so as to secure a cake having larger pores, thus offering less resistance to the flow of liquid. I n a case of this kind, provided the cake is not to be used, or if it can be subsequently separated from the filter aid, a precoating of the cloths with material of very fine porosity will give initial clarity. According to Hatschek,l if the mean distance from pore to pore of the filtering medium is larger than the diameter of the particles being filtered, the first layer will be drawn by the vortices into the orifices of the pores and the medium throttled. It can be seen that if an initial film of very porous material be deposited on the medium the orifices will not be throttled. The network of pores prevents the meshes from being closed up, and each succeeding layer of deposit mill form a network so that the whole thickness of the layer or film will be permeable. In addition it should always be kept in mind that it is advisable to work with reduced pressure a t the beginning and gradually to ’ increase the pressure as the thickness of the deposit increases. By this means the first layers of the deposit are spongy and porous, and offer little friction to the passage through of By C. P. Derleth liquid, and the high final pressure tends to dry the cake. CELITE PRODUCTS Co., MONADNOCK BLDG..CHICAGO, ILL. This precaution will also increase the length of life of the 9 filter aid is defined as a substance which may be admixed cloths by preventing the sharp corners of the crystals from ‘with liquids to be filtered without affecting the physical. or cutting the fibers of the filter cloths, as a result of the pump chemical characteristics of the filtrate and which facilitates pressure applied. Where corrosive liquid is being filtered the the removal of suspended matter in the filtering process. precoating of the cloths will give cleavage, allowing the cake Some of the materials which have been used for this purpose to drop off readily, and will prevent the finer crystals from are: Asbestos fiber, wood pulp, paper pulp, Filter-Cell diato- being forced into the pores of the retaining cloth or screen. maceous earth and infusorial earth (both known as kieselguhr), Where fine mesh screens or metallic cloth are used, a premagnesia, talc, lime, calcium sulfate, powdered silica, saw- coating with a filter aid is often desirable to give initial dust, and, partly for this purpose as well as for their decolor- clarity to the filtrate. The precoating may be done by chargizing properties, vegetable carbon, animal carbons, and fuller’s ing the filtering surface with a suspension of the filter aid in water or clarified liquid, returning the filtrate by by-passing earth. The material trade-named “Filter-Cel” is a cellular siliceous for a second filtration. It is easily possible with a good product which is especially prepared for filtration purposes, filter aid to retain finely divided crystals, such as barium and of exceptional purity. Kieselguhr, which includes both sulfate. FILTRATION OF NONRIGID PARTICLES diatomaceous earth and infusorial earth, is a broad term covering material from various deposits found throughout I n a few instances precipitates of the second class can be the world. The impurities occurring in these earths affect converted into a crystalloidal state either by proper regulathe filtering efficiency and contaminate the liquid with which tion of the precipitation or by subsequent treatment. This they are admixed. Strata of clay are found in most deposits should be done wherever possible, thus converting the proband, since they consist largely of alumina, have the effect lem into one of the first class. of giving gelatinous precipitates, especially in alkaline soluThose nonrigid suspended solids which do not lend themtions, which decrease the rate of filtration. Many of these selves to crystallization and which have many of the properclays are in such a colloidal state that the filtering efficiency ties and characteristics of colloids (although they may be of the material is completely counteracted. much larger in size) constitute the most difficult problems Filtration problems may be divided into three classes: in filtration. A problem is often rendered more complex by the viscosity 1-Those in which the suspended solids are rigid, and either of the liquid, as in the case of concentrated sirups or raw crystalline or amorphous. 2-Those in which the suspended solids are nonrigid bodies, linseed oil. I n the past, gravity filtration has been depended i. e., those approaching the colloidal state, although they may upon in coping with this type of problem. The downward be far greater in size than those properly classed as such. that no compressed air is entering the tank and accordingly that the air valve is closed. Upon opening the inlet valve controlling admission of liquor through the feed opening, liquor is admitted. If this valve be closed when a proper charge has entered the tank and the air valve opened, conditions will be right for a discharge of the liquid contents. The discharge valve is opened, and this passage of liquid takes place. I n the automatic monte-jus, the opening and closing of valves takes place without attention. Monte-jus may be made of special chemical stoneware or of metal. I n the latter case, they may be specially lined with lead, or the metal itself may be selected so as to provide suitable resistance to corrosion. A monte-jus is especially adapted for use with acids and liquors containing acids. It may be a small affair or it may be of considerable size. Monte-jus often fit into special conditions. I n one instance a filter press had to be fed with a thick granular and acid mixture. No pump could be found for the job. Gravity feed was unsuitable because of the probability that the connecting pipe line would suffer clogging. A monte-jus with a lead lining gave good service. Monte-jus are not cheap, either n t the beginning or during operation. The first cost is considerable, and compressed air must be applied to secure action, They are not largely used in the United States but have found considerable application abroad. One great objection to the apparatus is its intermittent action. Under the special conditions where a gravity feed to the filter is permissible, however, the monte-jus may be used to supply the elevated tank without anything but inconsiderable fluctuations in the pressure in the press, provided the plan section of the tank is given a generous area. A monte-jus that is to be used under considerable air pressure may often with advantage be imbedded in concrete.
Filter Aids
3-Suspensions containing mixtures of the above classes.
1
J . SOC.Chem. I n d . Z 7 (1908), 639.
