2 THE ANALYST. ON FILTRATION. BY OTTO HEHNER AND HENRY D. RICHMOND. IZead at Meeting, November, 1887. CONSIDERING that a large percentage of the analyst’s working time is spent upon various filtering operations, it is not a little remarkable how ill the conditions are understood which influence the rate at which a fluid passes through a filter. The text-books are generally silent upon the subject, or, when they do give instructions to the student, mostly direct that the filter-paper should be fitted into the funnel as closely as possible. (See, for instance, Fresenius’s (‘ Qual. and Quant. Analysis.”) I n many laboratories where students are instructed, this advice is followed. In others the paper is folded so that the angle formed by the edges of the filter should meet in an angle somewhat larger than 90”.In others, again, various forms of pleating the filters are in use. If we except the Bunsen Waterpump arrangement-which no doubt allows of the most rapid working, but is not in general favour owing to the transferring of filtrates from one vessel into another which it entails, and the not infrequent breaking of filters in the middle of an analysis which occurs,-for quantitative purposes, smooth filters closely fitted, or loosely laid into the funnel, are employed ; for qualitative work, various forms of pleated filters, It has long ago been observed by Fleitmann (“Zeitsch.f.Ana1. Chem.,” xiv.,p. 77) that in many cases a double filter filters more rapidly than a single one, because it presents a thicker passage for the filtrate to escape, With a similar intention, Ebermayr ( ‘ I Chem.Centralbl.,” 3, f. x., 176) recommends to lay under the filter a small piece of muslin into the funnel. Hempel (‘‘ Zeitsch. f, Anal, Chem.,” xiv., p. 308) etches, with hydrofluoric acid, a few channels into the funnels used in connection with a Bunsen pump, whilst De Mollins (“Zeitsch, Anal. Chem.,” xix., p. 334) uses a perforated cone placed into a funnel. A similar arrangement has recently been patented in England by Nickels. When a filter-paper of sufficient coherency is placed tightly into a funnel in the manner directed by Fresenius, and the rate of filtration ascertained, and the paper, when empty, is then cautiously lifted so that a glass rod may be placed between it and the funnel, the velocity of filtration is in,all cases much increased.It is evident that by placing the paper close to the walls of the funnel the fluid which is capable of passing through theTI€E ANALYST. 3 pores of the paper cannot find an exit with sufficient rapidity to allow of the total filter- ing capacity of the paper being brought into play. The smoother the funnel and the more nearly its angle approaches GOO, the more complete the fit and the greater the blockage. The old-fashioned fluted funnels were no doubt constructed to counteract this blockage, but in the case of the thin filter-paper now generally in use, this object is not attained, the fluting being too shallow to prevent the paper, expanded as it is by the moistening, from clinging closely to the funnel and barring the flow of ths filtrate.We therefore got some funnels made which have a number, generally four, high ridges running straight on the inner surface from the edge of the funnel for some dis- tance into ths shaft. It is impossible for the filter-paper, however thin, to adapt itself to the sides and to obstruct the flow of the filtrate, whilst a t the same time sufficient support is given to the paper t o enable it to bear the heaviest precipitates without risk of breakage. The following table gives the time, in seconds, required for filtering 250 C.C. of cold water. I n all cases the filters mere kept full of water during the duration of the experiment. For the purpose of strict comparison, we attempted to use one and the same piece of filter-paper for the different modes of filtration, but we found that even pure water gradually blocks up or contracts the pores of the paper, each succeeding litrs requiring a somewhat longer time than the preceding one, as will be seen from the following set of figures, which represent such successive times :- 217, 345, 267, 282, 293, 295, 296 seconds.For cadi experiment, therefore, ,z new piece of paper was taken, all from the same bundle, and four separate sets of trittls were made to equalise as much as possible differences in thickness and porosity. Column 1 shows the kind of filter-paper used; 2, the rate of filtration in the case of the old smooth funnel with narrow shaft, the paper laid closeIy into the funnel, and the shaft kept full of water; 3, filtration fhrough smooth funnel, with shaft cut short, the paper being folded 5 3 as not to fit well; 4, new funnel with four high ridges, paper dropped in anyhow; 5, old fluted funnel, no par- ticular care taken t o fit the paper ; 6, pleated paper for qualitative analysis; 7, paper pleated in the following manner, and shown in the figure :-Fold the paper across the centre as usual, open it, and fold it again at right angles; press the parts be tween the diameters thus indicated inward to the middle; the paper when looked a t from the top now forms a four-pointed star.Press it flat; the outer edges now meet at an angle of 90" Double back each pair of edges, so that they meet in the middle, n:ld open the filter as shown in the figure. All papers were of the same size.4 TI-IE ANALYST.1 2 3 4 5 6 Smooth ; Folded larger New Old Fluted Pleated "per' Shaft full. than 90°. Funnel. Funnel. Paper. 682 Best Swedish,j 200 Thin . .. . I 480 I- 218 Average 395 ( 328 -- 306 English, Thin< 156 c 181 -- Average 285 Schleicherand ( 563 S~hG11(590),( 450 washedwith I 545 H.FI. . . .. 750 Average 577 -- (1070 Schleicherand I 700 SchGll(59 7), -( very Thick c 6o -.- Average 886 Schleicherand I 412 SchB11(597),-( 388 Thick .. . . 430 (- 54s -- Average 445 592 5 48 138 85 341 137 135 79 287 160 840 160 180 950 357 415 368 1125 280 547 517 182 252 252 30 1 -- -- -- -- -- 95 95 97 74 90 51 60 49 65 56 160 145 192 162 165 262 350 332 322 317 188 212 168 160 182 - - -- -- -- 854 1290 600 115 715 152 90 217 152 153 338 543 582 508 492 295 288 420 488 3 73 132 455 188 507 321 -- -- -I -- -- 56 70 61 84 68 52 45 80 48 61 98 100 90 188 119 192 210 180 202 198 85 88 135 80 97 - - I- - 7 Richmond's Pleat.70 73 69 60 68 75 83 91 55 76 120 200 155 150 156 155 172 175 160 150 92 100 145 92 107 - - -- -- -- It is seen that, with all classes of filter-paper examined, the worst plan which can be adopted is to fit a filter closely into the funnel. The paper does not get a chance to work, and the loss of time is very great, the greater the thinner the paper. When the filter is folded so that it lies loosely in the funnel, excellent results are frequently ob- tained, but these are somewhat erratic, especially with thin papers, because the paper is apt to slip closely into the funnel when fully weighted with water. In all cases the velocity in Class 3 was tbe greater, the greater the angle a t which the fold was made.It disposes of any chances due t o imperfect folding or fitting of the paper ; it allows more freely than any of the other methods the passage of the fluid into the stom, and this far more cer- tainly than the old fluted funnel, which, indeed, with thin paper gives very bad results. Paper folded in pleats in the usual manner filters, especially in the case of very thick paper, faster than smooth filters are capable of doing, but the long, uneven The new ridged funnel gave the best and most concordant results.THE ANALYST. 5 edge renders washing so diflticult that they cannot well be used for exact quantitative work. The folding used in the tests stated in Column 7 is intended t o obviate this diffi- culty as far as possible, and to provide a pleated filter with a smooth edge. It will be seen that its filtration velocity equals that of ordinary pleated paper. The funnels with deep ridges, described in this paper, have been made by, and may be obtained from Messrs. Townson and Mercer, 87, Bishopsgate Street, at a price differing but very little from that of ordinary smooth funnels.