Nov., 1921
T H E JOURNAL OF INDUSTRIAL A N D ENGINEERING CHEMISTRY
mesh screens, before referred to as being placed between the layers, have heavy bronze centers that fit into these depressions and thus, if one layer is placed above the other, it makes a compact, elastic filtering column. Each layer is a filter in itself. The product does not pass through first one layer and then the other. On account of these compression centers, i t is impossible for the cloudy liquid to break through, even though the pressure may reach 100lbs. This type of filter will block with sediment and it is impossible for any unfiltered liquid to pass through into the outlet.
SIZESAND TYPESOF FILTERS The filters are made in various sizes and types to suit the requirements of the user. It is very esqential that the user secure from some source information as to the type and size of filter he should use. Often we find that the manufacturer hesitates to accept the advice or the help of the filter manufacturer, with the result that he is never adequately equipped. Certain chemicals must not come in contact with bronze. Alcoholic preparations must not be handled on iron filters. Products containing a great deal of acid should be handled through filters that have been heavily silver plated. If' the purchaser will €rankly put his problem up to the manufacturer of filters, he will receive unbiased and valuable information. Pulp filters to-day are a necessity for many manufacturers who desire to clear liquid products. The use of these filters is not always understood and perhaps a few illustrations may he of interest. At this time there is a great demand for filters for the clarifying of fruit juices. Manufacturers frequently desire to filter this juice immediately as it is pressed from the fruit, This, however, cannot be done. The fresh fruit contains a certain amount of pulp and glutinous matter that will block the filter almost immediately. The manufacturer is usually impatient and feels that some Gtering device must overcome the trouble. Frequently a great deal of money is spent in trying to secure such devices when they cannot be supplied. The trouble can be overcome by one of three methods: I-By pasteurizing or cooking the juice immediately, filtering, and then effecting a second pasteurization when the product in in the bottle. This method is not desirable, however, with many product.;, for the reason that the heat causes them to lose their fine flavor and aroma. 2-A very advantageous method is t o chill the product, bring it down to a temperature of about 34' t o 3 6 O F ,then filter, bottle, and pasteurize. 3-By running the product into tanks, and allowing it to stand from 4 to 12 hre., depending on the room temperature. While the product is thus a t rest, a considerable amount of precipitation occurs. Draw off and filter without disturbing the sediment, and immediately pasteurize in bottles.
PRESSURE TO BE EMPLOYED Another feature of filtration is that with certain products a high pressure can be used, while with other products a very low pressure of only a few pounds should be utilized. Olive oil is ti very delicate product and must be handled carefully with but little pressure. Recently there was brought to the attention of the writer an interesting experiment in the filtration of glue. A chemist of experience, a leader in his particular field of work, had devoted some time to the problem of properly clarifying a certain type of glue. During his research work he had endeavored to secure results by filtration, but had been unsuccessful. He then visited one of the large filter manufacturers and conducted his experiments in their laboratories. He found that a t just a certain temperature and pressure, excellent results could be secured. A few years ago it was found possible to clarify gelatin by means of a pulp filter. Necessarily the filters must be of ,large area and able to withstand the high pressure. I believe it safe to state that all the gelatin produced in this country to-day is filtered through pulp filters.
997
Let me leave this one suggestion: I do not wish to appear as holding a brief for any filter manufacturer, but I do commend that the prospective user of a filter present his problems to the filter manufacturer and secure his cooperation.
Leaf and Rotary Suction Filters By G. D. Dickey INDUSTRIAL FILTRATION CORPORATION, 115 BROADWAY, KEW Y O R K , N. Y.
Vacuum filtration was first successfully employed in the latter half of the 19th century in Belgium. An ordinary rotating drum with a perforated surface was covered with a filtering medium and hung in a container. In the interior of the drum, vacuum was created, and it was found that the liquid was readily drawn through the filtering medium, while the solids formed on the surface of the drum an even layer which could be scraped off on the opposite side, leaving the filtering medium relatively free. This trial was made on bicarbonate of soda, which is, of course, a very free-filtering material, and therefore the presence of a thin layer constantly remaining on the filter drum did not retard the rate of flow. The old drum type of apparatiis is still used on this class of work, and for many years no attempt was made to apply it to other substances. Suction filtration, like all other methods of separating solids from liquids, is dependent upon pressure differences. Offhand, therefore, it would seem as if creating a vacuum should, in effect, be the same as the application of direct pressure. Thus the pract(ica1limitation of 30 in. of vacuum would mean that the maximum results obtained by vacuum filtration would be equal to those of a press exerting 15 lbs. pressure. A careful study of the many underlying factors which enter into filtration shows that this does not hold true, principally since the formation of the cake itself plays a large part in the results obtained. The success of vacuum filtration is due to a large extent to the fact that the cake built up is of uniform density and thickness throughout, thereby causing equal resistance to the flow, both of the mother liquor and of the wash water. This uniform cake formation is caused by the fact that the vacuum is operative behind the filtering medium, is equal a t all points, and cannot be increased with increasing resistance. Thus, packing of the first layer of the cake against the filter cloth with increasing thickness of formation is avoided, and as the atmospheric pressure is relatively low, a porous cake is formed. Microscopic examination of a cross-section of filter cake formed by vacuum shows almost perfect uniformity throughout. There are, of course, many kinds of material which, because of their finely divided state or colloidal nature, will not form any appreciable layer under vacuum. This discussion therefore applies only to those materials which are of a crystalline base, or slurries to which foreign crystalline material has been added as a filtering aid, and which are amendable to vacuum filtration. DETERMINING FACTORS IN SELECTING VACUUM FILTRATION The ability to form a cake from 0.25 to 0.5 in. or more in thickness, even though many hours are required, is a determining factor as to whether or not vacuum filtration can be employed, and the time required for this cake formation determines what type of apparatus may be used. Two extreme examples are: that of starch, where a cake of 2-in. thickness can be built up in approximately half a minute, and that of some ore slimes, where 8 to 10 hrs. are required for building a 0.5-in. cake. There are two other determining factors in cake formation. The first is the percentage of solids to liquids present, and
998
THE JOURNAL 011"INDUSTRIAL A N D ENGINEERING CHEMISTRY
the second is the temperature a t which filtration may be carried on. Of course, the greater the proportion of solid., to liquids the more quickly is the cake formed. It would seem that the old law which states that "when particles approach each other by a distance less than three times their diameter they have a tendency to clump" holds good in most cases. This is especially noticeable in ferric hydrate, for instance, where vacuum filtration is practically impossible in suspensions as dilute as 1 to 16, but when a concentration of 1 to 1 is obtained, the material filters so freely that the rotary filter can be employed. Temperature is also a very important item in filtration. General13 speaking, materials filter much more freely when hot than when cold, owing to the lowered viscosity of the mother liquor, and its consequent freer passage through the filtering medium. Certain oils, for example, cannot be handled a t all when cold, but filter quite freely a t a temperature of 90". Despite the fact that, theoretically, vacuum cannot be maintained on boiling solutions, there is always enough condensation in the pipe lines to enable 28 in. to be maintained on slurries which are filtered a t the boiling point. An example of the increase in the filtering rate is in the handling of calcium carbonate, the filtering rate, and subsequent cake formation of which is increased 150 per cent by raising the slurry from room temperature to that of boiling. There is another aid to cake formation which enables many slurries to be handled which otherwise could not be, and that is the addition of some foreign material, such as Filter-Cel, in the handling of fruitjuices. Ordinarily, if Filter-Cel in the proportion of 1 to 2 per cent of the weight of the solids is added, a marked increase in the cake deposit, as well as in the rate of flow, is noticed. This is due to the separating of the particles, so that a more porous cake is formed, which builds more quickly and offers less resistance to the flow of the filtration. The addition of from 1 to 1 to 1 to 5 parts of Filter-Cel to solids enables practically m y slurry to be handled. When such a large amount is added, the cost is often prohibitive, and therefore a balance between the economy of increased and porous cake formation, and the cost of the foreign material added, must be worked out. I n many cases, the foreign material can be subsequently recovered and re-used. The choice of filtering medium should always be carefully considered. However, if the general policy of getting a medium sufficiently coarse so that the solids form their own bed and filter through themselves is followed, good results can be depended upon. Often, the filtrate will be cloudy a t first and it will be necessary to refilter, but this can be done readily unless too large an amount of solids has run through. WASHIXGIX THE SUCTION FILTER An equi-resistant, porous cake is very easily washed and dried. The washing is practically by displacement, the mother liquor held in the cake being pushed ahead of the wash water as a wall. This has been proved in actual practice by plotting a curve of the Baume value of the filtrate. The filtrate maintains its high degree for a certain length of time after washing is commenced and then drops abruptly to about 2 per cent of its former strength. Sampling of the cake a t this point shows a washing of about 98 per cent, and the remaining percentage of mother liquor is apparently held as a film around the particles as they lie one upon the other in normal piling. The last portion of mother liquor is removed more or less by diffusion, but can nevertheless be accomplished with a minimum of wash water, because there is an equal passage of wash water through the cake a t all points. I n filters of the open tank type, where any quantity of wash water desired can be drawn through the filter cake,
Vol. 13, No. 11
an exact amount can be determined upon, if i t is desired to save the washings. On the other hand, when the object is to get a thoroughly cleansed filter cake, washing can be carried on for a long period of time. With a machine of the rotary filter type, the time of washing is more or less fixed by the revolution of the filter drum, and very exact requirements in washing can be accomplished only by the discharge of the filter cake, pulping, and refiltering. The ordinary limit on the rotary filter, without recourse to pulping, is around 1 per cent. This can be definitely figured out by determining the percentage of mother liquor in the original filtrate, and the percentage of moisture in the a t e r cake. Mathematically, with the increase in the proportion of solids to liquids due to fiitering, there is a corresponding decrease in the percentage of mother liquor present. If to this is added a 50 per cent reduction for washing, and the cake is now re-pulped to a 1 to 1slurry, there will be a certain proportion of mother liquor in this second slurry. Following the same procedure for the second filtration, the amount of the mother liquor left in the second discharged cake can be closely figured out without making any filtration tests. Very often, the efficient washing of the filter cake i s entirely destroyed by the cracking of the cake before washing commences. There are several ways of obviating this difficulty. In the open tank filter, the basket may be raised and the cake dried until cracking commences, and resubmerged one or more times before the cake building is completed. This method takes u p the shrinkage of the filter cake to a large extent, and permits transfer to the wash-water tank before cracking can occur. The cutting down of the vacuum, of course, is a great aid to reducing cracking, but often this cannot be done because the cake would be so wet as to fall off during transfer. In rotary filters, cutting down of the vacuum may be employed more successfully to eliminate cracking, as there is not so much opportunity for the cake to slough off. Rolling or smoothing down of the filter cake by external attachments is often used to iron out the cracks. Finally, if an unlimited amount of wash water can be used. the cake can be enveloped in the wash water from the point of emergence to the point of discharge, so that there is no chance for cracking to occur. DRYINGAND DISCHBRGING THE FILTER CAKE The drying of the filter cake is dependent to a large extent upon the formation of the cake itself. A cake which is uniform in porosity and thickness can be more or less readily dried by drawing air through it, and, because of its porous nature, is easily calcined or otherwise reduced to powder form after discharge. Ordinarily, the thinner the filter cake the more easily it is dried, although there is not a great difference between cakes of 0 . 2 5 to 0 . 5 in. In discharging a cake from the open tank filter, the use of steam, water, or compressed air frees the filtering medium cleanly, and no scraping is required. In the rotary filter, if a thin layer left on the filter drum materially cuts down the rate of flow, air pressure may be applied just prior to the scraping of the cake or just afterwards, as the filter drum enters into the unfiltered slurry. The object of the air pressure, of course, is to clean the filtering medium by reversing the action of cake building. TYPESOF SUCTION FILTERS The three main types of suction filtering apparatus are: open tank filters, continuous rotary filters, and continuous rotary hopper dewaterers. OPEN TANK FILTER-This consists of a unit or basket composed of a number of plates or leaves, fastened together and connected to a common header. Each of these leaves is
Nov., 1921
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 CHEMISTRY
constructed of a pipe frame, usually in the form of a rectangle, over which is drawn the filtering medium, sewn in the shape of a bag. The open or loose ends of the bag are held tightly together a t the top by means of wooden clamps. The two sides of the filter cloth are held apart by parallel wooden strips set upright, and the filtering medium is stitched in such a manner as to hold the strips in position. Along the bottom of the pipe frames are perforations, causing the suction to pull uniformly a t all points when applied behind the filtering medium. . In operation, the basket is submerged in a tank containing the slurry to be filtered, and suction is applied behind the filtering medium. The clear filtrate is drawn through the filtering medium, pipes, and header to the desired point, while the solids are deposited on the surface of the filtering medium. After a cake approximately 0 . 5 in. or more in thickness has been built up, the basket is transferred, by means of a hoist and trolley, to a tank of clear water. Suction is kept on during this operation. During mashing the effluent can be switched to another tank, thus making possible a separation of filtrate from wash water. After washing, the basket is shifted to the discharge platform, suction still being kept on, and the filter cake is dried by drawing air through it. Discharge is accomplished by shutting off the suction and applying air, water, or steam pressure through the same line. This discharge not only frees the solids from the filtering medium, but leaves it clean for the next cycle. One of the disadvantages of the open tank type of filter is that in discharging the filter cake, the compressed air or steam causes a portion of the entrained liquor to be blown back into the cake, and a higher moisture content is obtained than would be the case if scraping could be successfully employed. The second disadvantage is that considerable head room is required for the hoist and trolley which must be used for shifting the filter basket. Finally, a cake of a t least 0 . 5 in. should be built up upon the surface of the filtering medium in order to insure a clean discharge. The advantages are that any period of time of cake building, washing, and air drying that is desired can be decided upon by the operator. The formation of the cake can be followed very closely by raising the filter basket from time to time. Increased filtering area can be obtained by the simple addition of filter leaves, so that a very large square area can be obtained in a comparatively small space. The fact that the filter can be cleaned automatically and the entire operation carried on without any set limits gives this filter a great advantage over any fixed type. At the same time, it is semi-automatic, requiring but a small amount of labor. The uniform porous cake formation gives maximum working with minimum wash water, and the discharge of a cake of this nature means economy in final drying. ROTARY FILTER-The rotary filter of the multiple compartment design consists of a drum whose periphery is divided into a number of uniform compartments, each being connected by a separate pipe line to the central valve hub, which is cored out to receive these lines. The compartments are covered with a wire screen suitably supported, and over this, encircling the entire drum, is stretched the filtering medium, each compartment being kept separate and distinct, and giving a smooth peripheral surface to the drum. The valve hub rotates against a stationary valve cap, which is so arranged that each compartment can be subjected to suction as well as pressure during any portion of the cycle. Provision is made in the valve cap for applying suction and for drawing off the filtrate and wash water separately if desired. The filter drum is hung in a hopper bottom container, which is provided with mechanical agitation. The
999
scraper constantly removes the filter cake as the drum revolves. Washing is accomplished by means of sprays on the surface of the drum just after it emerges from the container with its accumulated solidq. In operation, the container is kept filled with the slurry to be filtered, and suction applied to the compartments through the valve. The solids are thus picked up, washed, dried, and carried away during every revolution of the drum, while the filtering medium is automatically cleaned by air pressure. either just before or just after the solids are scraped off. The limitations of rotary filtration, of course, are that the solids must be relatively free-filtering and must be in sufficient proportion to enable a cake of 0.375 in. or more to be built up during a 4-min. period, because it is usually not economical to revolve the drum slower than one revolution in 10 min. A balance must also be worked out between the thickness of the cake formed and the washing and drying of this cake, in order to determine how fast the drum should turn to get the best results. The great advantage of the rotary filter is that its operation is continuous and automatic, thereby practically eliminating labor and saving an immense amount of time. Large capacities can be obtained upon a relatively small filtering area because of the continual ailtomatic cleaning, and the cost of operation is remarkably low. The advantages of a uniform porous cake for washing and drying, outlined under the open tank filter, are equally true in this type of filter. ROTARY HOPPER DEW.4TERER-There are other types Of filters similar to the rotary filter; for instance, the rotary hopper dewaterer which is used for solids which are too heavy to be made to adhere to the surface of the filter drum by suction. This machine consists of a series of hoppers provided with filter bottoms arranged radially about a central shaft. Each hopper, just below the filter bed, is connected to a valve hub by means of pipe lines which permit suction to be applied to the hopper during any desired arc in the revolution. I n operation as the hoppers rotate they are charged from an overhead chute about 30" before they reach the zenith. At this point suction is applied and continued through an arc of approximately 90". Here suction is automatically cut off and the hoppers discharged either by gravity or by the aid of air pressure. The solids may be washed before discharge by means of a sprinkler equipment. The points of advantage and disadvantage are practically the same as those of the rotary filter, except that the depth of the hoppers can be varied according to the material to be filtered, thereby increasing or decreasing the capacity of a given machine. CAPACITIES OF DIFFERENT TYPEs-In conclusion, there may be cited a few examples of what capacities may be anticipated with the three types of suction apparatus. Of course, as previously stated, the capacity of any apparatus varies with the material to be handled: and even with materials which are of the same nature, there is sufficient variation in plant manufacture to cause a coneiderablc degree of variation iu filtration results. In handling sulfur black in the open tank type of filter, a capacity of 200 lbs. per sq. ft. per day of 24 hrs. is usually averaged; or in clarifying lime water 1500 gal. per sq. f t . per day of 24 hrs. In the case of the rotary filter a capacity of 2000 lbs. of calcium carbonate in caustic soda manufacture or 1000 lbs. of cement from cement slurries can be averaged per day of 24 hrs. With the dewaterer, a capacity of 2 tons per sq. ft. per day of 24 hrs. upon the dewatering of slate ground by the wet process or 1 ton of crystals in copperas manufacture is the usual tonnage obtained.
