FILTER, an arrangement for the separation of impurities from liquids, by passage through porous material. The filtering process is common in nature. The clearness of spring water is due to it; for such water generally comes from a considerable depth in the ground (as appears, e.g., from its equable temperature throughout the year); and having traversed a variety of porous strata, it has under-gone a straining action, producing the beautiful trans-parency we observe. This does not, of course, represent absolute purity, for the liquid retains in solution various substances acquired by contact with the strata through which it has percolated. The operation of filtration is extensively practised in purification of water, on a large scale for supply of towns and cities (now an important branch of civil engineering), and on a small scale for domestic purposes, and for the use of ships on a long voyage, &c. It is also a valued method in chemistry and the arts.
The mechanical action of straining, by which all particles larger than the interstices of the porous material are arrested, is one important function of filters, and it used to be commonly represented as their only function. They may act, however, in other ways to purify water. Not to speak of the further mechanical actions of subsidence on upper surfaces of particles of the filtering medium, and lateral attraction and adhesion of suspended matter (which doubtless occur in some measure), it is an important fact, now well ascertained, that a filter may separate from water a portion of the matter held in solution. On the other hand, it is often precisely such matter, when of organic origin and putrescent (and minute invisible disease-germs may here be included with the matter in solution), that it is specially desirable to remove from drinking water, as being prolific of mischief when taken into the system. In numerous instances an outbreak of virulent disease, such as typhoid fever, has been clearly traced to water so con-taminated. The danger is the greater that such water may be bright and sparkling, and peculiarly palatable.
That even sand has the power of removing dissolved matter from water was shown by Mr Witt's valuable experiments at Chelsea, described in 1856. In one of these, e.g., water containing T42 grains of chloride of sodium per gallon (70,000 grains) was deprived of 22 per cent, of that substance by filtration through a depth of 1 foot 9 inches of sand. The sand had no appreciable action on dissolved organic matter, as charcoal had, but the quantity of such matter originally present was small. It is probable that all finely porous material has such action, more or less. The efficiency of charcoal in this respect, and especially fresh animal charcoal, has been well demon-strated and utilized. The mode of the disappearance of dissolved organic impurities has been a disputed point. Som^ say they are retained and accumulated in the pores of the charcoal; but the experimental evidence seems to leave little doubt that they are mainly oxidized, so as hardly to impair the activity of the filter. In fact, the value of a filtering material will be found to depend chiefly on the power it has of bringing oxygen, stored in its fine pores or otherwise provided, into chemical union with the dissolved organic matter and destroying it. At the same time it is obvious that, chiefly by the mechanical action of straining, organic matter may accumulate in a filtering medium, and in course of time, through decomposition, render the water which passes through more impure and less wholesome than in its unfiltered state.
We may remark here that river water and shallow well water, while extensively used for water supply, are in general the most largely polluted. Rivers commonly receive large volumes of sewage, impure surface water from cultivated land, and other contamination. The water of shallow wells, especially in large towns, or when near churchyards, stables, cesspools, &c, is often contaminated with organic matter of the worst kind, in large quantity. While rain water, collected from the roofs of houses in butts, no doubt contains organic matter, this hardly bears comparison in amount to the organic impurity in rivers (thus Dr Hassall found it a hundredfold less). Deep spring water, again, is freed from much of its organic impurity through natural filtration. The advantages of a good lake-supply for large towns has been amply demon-strated in these days, notably in the case of Glasgow, which draws its water from Loch Katrine.
Though the arrangements for water supply of most of our large towns include filtering processes by means of which, as a rule, excellent drinking water is abundantly provided, so that in the opinion of some chemists a domestic filter may be superfluous, while it is sometimes a source of harm (owing to lack of proper attention), it is generally thought a wise additional safeguard to employ one of those instruments, in view, more especially, of some of the possible consequences of epidemics and floods, and the necessity of house-storage of the water received. In country places, and in various other circumstances, their use is often quite imperative if the laws of health are to be respected.
The Japanese and Egyptians seem to have used water filters at an early period. These consisted of sandstone or unglazed earthenware, and were of bowl or egg shape, with small projections at the top resting on a wooden frame.
The water poured into this vessel filtered through to a vessel below. About the middle of last century, slabs of a lias found in Picardy were used as a filtering agent, being fitted as a false bottom in water cisterns; the water was drawn off through a tap from the space below. A porous filtering stone of Teneriffe was at one time imported largely into England. The " alcarrhazas" are filter vessels of porous biscuit stoneware made in Spain.
One of the earliest filters in England was that patented by Mrs Johanna Hempel in 1790. It was a supported basin made of tobacco pipe (or similar) clay and coarse sea, river, drift, or pit sand, and hardened in the fur-nace. In the following year (1791) the ascending prin-ciple was first applied by Mr James Peacock. Water from an upper reservoir was admitted through a pipe to the bottom of a box containing strata of sand, gravel, and a mixture of charcoal and powdered calcareous stone. Passing up through this, it was drawn off by a pipe at the top. The filter was occasionally cleaned with an exhaust-ing and condensing pump, which sucked up water rapidly through the filtering material and then sent it back with force, washing out the dirt.
The construction of filters is a matter on which inventive-ness has been largely exercised. All sorts of porous substances have been called into requisition, as may be seen by a glance at the patent records. Thus, to mention some of these, we have various kinds of stone, sand, gravel, powdered glass, clay, porous sulphur, preparations of iron, charcoal (vegetable and animal), cloth, felt, horsehair, skins, paper, silicated carbon, sponge, wood, wool, cane, capillary threads, and so on. Vegetable charcoal, we may note, was first employed in 1802, animal charcoal in 1818, and solid carbon blocks in 1834.
It is impossible here, and it would be tedious, to give anything like a detailed account of the changes that have been rung on the filtering principle for domestic filters. In the simplest and most familiar forms, of course, the water passes down by its own gravitation through the filtering medium to a reservoir below. The force of downward pressure has sometimes been augmented by a head of water, sometimes by a force pump, and sometimes by means of air condensed above the water to be passed through the filter. Or the air has been extracted from the vessel con-taining the filtered liquid, thus adding force to the atmo-spheric pressure above. The ascending principle has appeared in various applications; and in some cases the water has been caused first to descend and then to ascend through filtering material (a vertical partition, e.g., being fixed in a vessel so as to reach nearly to the bottom, and layers of sand and vitrified limestone being placed on either side of the lower part; the water is poured in on one side of the partition and drawn from the other side near the top). Lateral filtration has also been tried. Filters have been arranged to act in the cistern, or in connexion with the service pipe between the cistern and the water tap, or independently, and, in the latter case, either having the unfiltered water poured into them, or being placed in a vessel of it, and giving filtered water through a tube. Sometimes a series of separate connected vessels have been employed; and for very dirty water it is often advantageous to have one system of filtration for the coarser, and another for the finer impurities. Once more, filters have been ren-dered self-supplying by means of a ball-cock. These are some of the general forms which the filter has taken.
The application of pressure to filters cannot as a rule be pursued very far, for it tends to derange the apparatus and render the filtered liquid muddy. Enlargement of surface is a better means of obtaining rapid filtration. Upward filtration, while it offers some advantages over downward, has not hitherto come very much into use. It is open to
objection in that the water sent upwards has a tendency to force a passage through certain channels, without being uniformly disseminated in the material; and the deposit of any filth is excluded from view, and mostly also from smell, instead of being exposed and giving us warning.
In passing now to examine more closely some of the ap-proved forms of domestic filter at present in use, it should be borne in mind that while any of these filters will doubtless purify water both mechanically and chemically, more or less, it is only on condition of their being properly attended to, and the filtering material renovated at intervals depending on its nature and the nature and amount of impurity in the water. The term " self-cleansing," applied to some filters, may have a (limited) true sense, but if understood to imply that a filter, let alone, will go on ad infinitum giving pure water, it is quite inapplicable; solid impurities must accumulate and call for removal. The statement, occasionally made, that a filter is " warranted to remove all impurities" from water is absurd and hardly deserves notice. Absolutely pure water is a thing almost unknown ; careful distillation alone will give an approxima-tion to it. Again, the claim that a filter will remove all lime from water is often false ; filtration is capable of re-moving only a small quantity of lime. It must be allowed that sundry points in the process of filtration still remain in some obscurity, and it is matter of regret that the action of some common filtering agents has not been so fully cleared up by scientific experiment as others. Still enough has been ascertained probably to guide to the construction of a filter on rational principles.
In a large proportion of filters, as already indicated, some form or other of carbon is the chief filtering agent. The well-known filters of Lipscombe are cylindrical-shaped covered vessels of glazed earthenware, in which the filtering medium, a mixture of vegetable and animal charcoal, in granular form, is enclosed between two slabs cemented in the case. The upper (glazed earthenware) slab has a central aperture with raised border, and a small perforated basin immediately below it; into this is inserted a sponge to arrest the grosser impurities, which is taken out and cleaned at short intervals. The filtered water passes through the lower (and porous) slab to the reservoir below, which communicates above with the outer air by a narrow tube passing up within to the top of the apparatus, and delivers its water through a tap.
Charcoal in the form of solid finely porous blocks, which can be conveniently brushed and cleaned externally, is now often moulded for filters. The convenient decanter filter, in which the water passes through the block to a central tube, forms an elegant addition to the sideboard. The annexed figure (1) represents one of these as made by Atkins, who also furnishes earthenware filter vessels having a division across the inside wherein a carbon block is fitted water-tight, which can readily be taken out and cleaned and replaced, or a fresh one substituted. Sometimes the block is fitted in a movable pan. Again, in the filter shown in fig. 2 a double filtration is effected, the water passing first through loose charcoal B, then through a charcoal block C, supported as shown. The block in this case is said to last longer without cleaning. The movable and perforated earthenware plate A, which is placed above the charcoal (see fig. 3), allows of easy renewal of the latter. The charcoal used in these filters is chiefly of vegetable origin. They are found to remove more or less of organic and inorganic matter dissolved in water.
It may be useful here to call attention to some of the conclusions arrived at regarding charcoal in the sixth report of the Royal Commission on Rivers' Pollution a short time ago. Fresh animal charcoal has been proved to act powerfully in the removal of organic impurity (considerably more so than vegetable charcoal), as well as of mineral matter. But, ' according to Dr Frankland, its reduction of the hardness ceases in about a fortnight, the removal of organic matter continuing even after six months, though to a much less extent, especially if the filter be much used. It was found necessary to renew the charcoal every six months when used for filtering the comparatively pure water of the London New River
Company, whilst the water of the Thames required renewal of the charcoal every three months. If this be not done, myriads of minute worms, we are told, are developed in the material passing out with the filtered water. Other statements, of scientific weight, regarding animal charcoal are more favourable to it, and seem to show that under certain conditions, perhaps imperfectly understood, it may give IQ* gte at A better results. °'
In Major Crease's system, which is adopted in the British army and navy, loose animal charcoal is com-pressed between plates by means of a screw, the amount of compression being determined by the degree of impurity in the water to be filtered.
The silicated carbon of Mr Dahlke's filters is obtained by mixing animal or vegetable charcoal with the residue of distillation of Boghead coal. By adding a little clay to the latter product, and saturating the whole with oily matter, it can be moulded, after which it is burnt. In one form of the filter, two carbon blocks are sealed into the interior of an earthenware vessel, granulated car-bon being placed between them ; in another, a double action is obtained by placing a carbon block over the entrance to a second carbon medium. These filters have been highly com-mended for their chemical properties, attributed to magnetic oxide of iron present in the medium. We give (fig. 4) a representation of the sili-cated carbon siphon filter with its case. Water may be sucked through FIG. 4.Silicated Carbon it from a stream directly into the Siphon Filter, mouth, or passed siphon-wise from one vessel into another. These and similar pocket filters of Atkins's were supplied to the Ashantee expedition.
A very powerful filtering medium was discovered and introduced many years ago by Mr Thomas Spencer. It is called magnetic carbide, and consists of protoxide of iron in chemical combination with carbon. It is obtained by roasting hematite iron-ore with granulated charcoal for twelve to sixteen hours at a dull red heat. Mr Spencer considers the purifying property of the oxide to be due to its power of attracting oxygen to its surface, without the latter being acted upon, the oxygen attracted being then changed into ozone, by which the organic matter of the water is consumed. The magnetic carbide is used in granular form. This filter gained prize medals at the last London and Paris ex-hibitions, and its effi-ciency was demon-strated by the Lancet Sanitary Commissioners' report on filters in 1867.
The only other system we shall here notice is that in which spongy iron is used. This substance is me-tallic iron which has been reduced from an oxide without fusion. It is in a spongy or porous state of extremely fine division. Its remarkable purify-ing action on water was discovered by Professor Gustav Bischof of the Andersonian University, Glasgow; and ex-periments made with his filters by the Eoyal Commissioners already referred to showed that their power both of re-moving organic matter and reducing the hard-ness of water even in-creased during upwards
Fla 5-Bisohof's Spongy Iron Filter.
of eight months' constant use. The general form of the filter is represented in fig. 5. An inner vessel containing the spongy iron is supported in a case, which, below, contains some prepared sand, a regulator A, and a receptacle C, for filtered water (with tap, not shown). The unfiltered water B is in this form supplied from a bottle which is inverted into the upper part of the inner vessel (a method familiar to chemists). After passing through the spongy iron, the water ascends through an overflow pipe in the direction of the arrows ; the object of this is to keep the spongy iron, when once wet, constantly under water, as otherwise it is too rapidly oxidized. The object of the prepared sand (which is generally in three layers, viz., pyrolusite at the top, then sand, then gravel) is to separate traces of iron retained in solution. The regulator A consists of a tin tube, cemented in the position shown; it is open at the inner end, which is below the perforated bottom supporting the sand, and closed by a screw cap at its outer end. It has also a small lateral perforation, through which alone the filtered water passes into the reservoir. Should the perforations get choked, the screw cap is removed, and a brush inserted ; on starting at first, too, the cap is un-screwed, that the materials may be well washed out with-out soiling the lower reservoir. With a ball-cock and con-stant supply of water, the inner vessel is dispensed with.
The nature of the action of the metal on organic matter is rather obscure. Mr Bischof considers there are both reducing and oxidizing agencies constantly at work, and that the oxides of iron, being present in their nascent state, must be very energetic in their action. Probably ferric hydrate, the last product of oxidation, takes an active part in separation of the organic matter, transferring oxygen to it. Again, spongy iron is known to be very energetic in precipitating any lead or copper. Its reduc-tion of the hardness of water presents some difficulty.
This filter, we may add, recently gained the prize medal for general excellence given by the Sanitary Institute of Great Britain, and has been otherwise commended.
The subject of cistern filters (of which there are many varieties) need not detain us long here. The arrangement adopted by Lipscombe, shown in fig. 6, may be taken as an example. The impure water chamber of the filter, thence upwards through a porous stone, then through powdered charcoal into the pure water reservoir, whence it may be drawn off cold by the pure water tap, or hot and pure from the boiler.
When the filter is in action, the grosser impurities are stopped by the porous stone, while the finer pass into the charcoal and are chemically acted on. Each time the unfiltered water tap is opened to obtain water, unfiltered water enters the inlet and scours out the impurities in the chamber. To clean the filter, once every six months or other period, a plug is drawn up by means of a galvan-ized chain (attached as shown) from the top of the filter and dropped into the inlet. The unfiltered water tap is then turned on several minutes, so that the cistern water rushes down through the filter and out of the tap, carry-ing out impurities from the filter.
In the filtration of water for supply of towns, galleries of masonry are often constructed in the sand and gravel forming the river bed or banks; the water percolates through and enters these tunnels at the bottom and by side channels; it is thence pumped to the town. Genoa, Toulouse, Lyons, and Perth present examples of this system, which is apt to prove rather costly. The system of artificial filter-beds of sand resting on gravel, &c, is now more generally adopted. It was first introduced into London by Mr Simpson, in 1839, after a study of various works in the north, and especially of the costly experience of Glasgow. Eor the Chelsea Water Company he had a series of tunnels built of brick without mortar; these were covered with a layer of fine gravel 2 feet thick, then a 2-feet stratum of fine gravel and coarse sand, and lastly 2 feet of fine sand, or 6 feet in all. The sand was periodically removed to the depth of about half an inch. To the filter bed, covering 1 acre, were attached two reservoirs, slightly higher; into these river water was pumped, which, after time for subsidence, was admitted to the filter beds through small pipes.
The eight Water Companies of London all use similar beds, increasing in coarseness downwards, of various depth and proportion of materials. A sharp siliceous sand is preferred for the upper bed (the true filtering agent), and the stratum is seldom made less than 2 or more than 3 feet thick. Sometimes this bed is laid immediately on a bed of small shells.
It is considered that filtration through sand, to be effective, should not proceed more rapidly than 6 inches of descent per hour (in the London beds the rate varies from 2-5 to 10-7 inches), and that there should be about 1J square yards of filtering area for each 1000 gallons per day. The depth of water maintained on filter-beds varies between 1 foot and 7 or 8 feet. The inlet arrangements should be such as to produce little disturbance of the sand in charging; thus the water may be admitted into a long trough from which it gently overflows, or through an inlet pipe carried to the centre of the bed and turned upwards (at Chelsea the water enters through a wall of gravel between two horizontal concentric arches of brickwork with vertical joints). The beds are drained variously, e.g., by means of perforated stoneware pipes, or pipes with open joints, sometimes leading into a brick culvert which traverses the bed. A good method of draining is that of Mr Muir, adopted by the New Eiver Company. It con-sists of two courses of bricks laid flat and dry; in the lower, the bricks are placed end to end in series alternating with half-brick spaces which serve as drains leading to a central culvert; in the upper course, the bricks are laid close together, forming a floor, on which a thin layer of fine gravel supports a bed of sand.
Filter beds require to be cleaned, at intervals varying from one week to six or eight, by removal of about half an inch of sand. The clean sand remaining is loosened with a rake and exposed to the air some time, then smoothed over. The filter bed may be made with several compart-ments, some of which may remain in action while others are being cleansed. There are various contrivances for washing sand previous to its replacement. Filter beds are some-times arranged to be cleansed by a reverse current sent upwards with force,an operation which may be aided by stirring the surface sand after the water has come above it. Such a system is practiced, e.g., in the Greenock, Paisley, and Dunkirk water works. At the last-named place, washed coke is among the filtering agents used. (For an account of the works at Dunkirk see Engineering, vol. xiv. p. 206.) In the system of M. Maurraz the water is filtered both per ascensum, et descensum, the two portions of water, which flow in opposite directions, uniting at the middle. The sand is retained in closed and perforated boxes. This filter also permits of cleansing by reversal of the current. Mr Spencer's carbide system has been applied successfully on a large scale at Wakefield and other places, the carbide layer being combined with others of sand and gravel. With carbide, the total thickness of bed may be considerably reduced.
Filtering is frequently practised by the chemist. And whereas in the ordinary filtration of water above described the purified liquid is the object of the process, while the matter retained is merely to be got rid of and destroyed, the reverse may be the case in the laboratory, the retained matter being sought, while the " filtrate," as it is called, is disregarded. For most laboratory preparations the material used is unsized paper. Swedish filtering paper (which is prepared with very pure spring water) possesses the advan-tage of filtering very rapidly, and of being singularly free from inorganic matter. Cloth is employed in the case of viscid liquids such as syrup or white of egg ; while corro-sive liquids may be filtered through pounded glass. Asbestos is a valuable filtering material, since, by making it red hot, all organic matter may be destroyed, and acid and alkalies have scarcely any action in it. Being nearly indestructible, it can be repeatedly used. Glass-wool has also of late years been recommended.
Paper filters, to be placed in a funnel, are sold ready cut of circular shape. The paper is folded twice to the form of a quadrant, and this, when half opened, forms a cone, whose edges meet at an angle of 60°. To facilitate passage of the filtered liquid, small folds are sometimes made in the filter all round. In a filter devised by Bunsen, the neck of the filter is inserted in the caoutchouc stopper of a lower vessel, and through this stopper also passes a tube con-nected with an exhausting apparatus. The production of a partial vacuum below accelerates the filtering process. Sometimes substances have to be filtered under the influence of heat, as they solidify at ordinary temperature. In such cases the funnel may be surrounded by a sleeve containing water, which is heated with a lamp.
In Robinson's oil filter, oil is forced up from a cask into and through the filtering apparatus (containing charcoal or other medium) by water entering below from an upper reservoir. Sundry modes of filtration practised in the arts (sugar-refining, &c.) will be referred to elsewhere. In some of them centrifugal force is employed.
Circumstances are not uncommon in which it is very desirable to remove impurities from air by a process of filtration. Cutlers and other grinders use respirators to arrest the small particles which would otherwise find their way into the lungs. For steel particles magnetic gauze is an efficient protective. Professor Tyndall's respirator for firemen consists of an iron cylinder, attached to a mask, and containing charcoal, and three layers of cotton wool, one moistened with glycerine ; the ends of the case are of wire gauze. With this respirator it is possible to enter an atmosphere of dense smoke and remain in it over a quarter of an hour. The disinfecting properties of charcoal have been turned to good account by Dr Stenhouse for purifying
air, the substance being used in the construction of respira-tors, or at the outlet of sewers, &c. An interesting applica-tion of the principle was made in the justice room of the Mansion House, London, in 1854, where offensive smells originating just under one of the windows were effectually removed. There are various devices in existence for purify-ing the air admitted to railway carriages and other inclosed spaces. Thus the air may be passed through wire screens, or through a spray of water, &c. In connexion with biolo-gical research and the germ theory of disease, the removal, by filtration, of minute foreign particles from air is a mat-ter of great moment.
For further information see, among other works, Humber's Water Supply of Cities and Towns, 1876; Sixth Report ofthe Royal Commis-sionon Rivers Pollution (published in 1875); Lancet Sanitary Commissioners' Report on Filters, 1867 ; Wanklyn and Chapman's Water Analysis; Philosophical Magazine, 4th series, vol. xii. p. 30 (Witt); Proceedings of the Institution of Civil Engineers, May 1867 (Byrne); Paper by Mr Pearse, on Water Purification, Sanitary and Industrial, to Society of Engineers, March 4, 1878 ; Paper by M. Bischof, on The Purification of Water, to Society of Arts, April 25, 1878 ; Registrar-General's Returns for 1876, &c.; Chemical News, vols, xxxiii. and xxxiv. (Wanklyn, Hildebrand); Dictionaries of Ure, Knight, Tomlinson, Watts, &c. (A. B. M.)