Sidebilder
PDF
ePub

EXPERIMENTS ON RATE OF FLOC FORMATION.

We may safely assume from analogy that in any homogeneous mixture of alum solution and alkaline water, the reversible reactions proceed almost instantaneously to a state of equilibrium which would remain unaltered were there no separation of solid material. On the other hand, it is a matter of experience that the passage of a precipitable material from invisible to visible aggregations often requires very considerable time. Undoubtedly one determining factor is the following: It has repeatedly been shown for liquid drops forming in a saturated atmosphere, for crystals and so-called amorphous precipitates separating from solution, that the small aggregates are less stable than the large, their material "distilling" over to, or at least having a preferential tendency to deposit upon, the larger aggregates. In the case of raindrops it can be shown by very simple thermodynamic reasoning that the vapor pressure in equilibrium with a very small drop is greater than the vapor pressure in equilibrium with a larger drop. Therefore, a higher degree of saturation must be required to form the first minute drops than the subsequent larger ones. The same reasoning can be applied to the formation of bubbles of gas from a solution of a gas in a liquid and to the formation of a solid aggregate from a solution of the solid in a liquid. In each case appreciable time is required for migration and orientation of the material entering into the aggregate.

Now, then, if all other conditions could be held constant while the degree of supersaturation of the separable material is varied, the rate of aggregation would be proportional to the concentration in the liquid phase of the separable material. In the case at hand we could vary the degree of supersaturation of the water with Al(OH), (assuming this to be the precipitate) either by increasing the total alum added at a given final pH value or with constant aluminium by varying pH. Under the last condition the maximum should occur at the isoelectric point. Undoubtedly the experimental attainment of the condition that "all other conditions shall remain constant" is difficult, especially since we have very little detailed information upon the mechanism of a solid separating from solution. Nevertheless, with due reservation, we should expect minimum time of flocculation at the isoelectric point. As will be shown presently, we obtain experimentally a minimum time in a zone of pH within which we have found a provisional value for the isoelectric point.

It would be difficult, however, to formulate a priori the precise mathematical relations, because we know little quantitatively about the forces operating in the formation of precipitates. We prefer, therefore, to describe our experimental results and to leave them

formulated by means of empirical charts and equations which are sufficient to bring out the important features.

In the study of flocculation in waters it has sometimes been the custom to ascribe certain numbers to flocs of different character. This seems to us an irrational procedure, because, aside from the inherent difficulty in obtaining any mathematical relation between the qualities to which the numbers are assigned, the observer has to carry from experiment to experiment a very precise picture of several flocs to which he has ascribed numbers. If, however, the attention is fixed upon one state, as, for instance, the very first detectable formation of floc, the remainder of the measurement may be left to the impersonal accuracy of a stop watch. It is this method that we have employed, leaving all questions of the quality of floc to incidental notes. As we shall see, this was a fortunate choice of method. The experimental method followed will be briefly described. A definite volume of a solution of known composition was treated with varying amounts of aluminium sulphate in dilute solution. After mixing as rapidly as possible, the liquid was poured into a 100 c. c. cylinder. To increase the visibility of the floc, the cylinder was slightly agitated whenever observations were made. The initial turbidity, the time required for the first appearance of a "floc," the character of the "floc," and its rate of settling and its abundance were noted. Caliometric pH measurements were made with the mixture. The temperature at which the experiments were conducted did not differ appreciably from 22° C.

Following this procedure, the entire range of pH values within which a "floc" can be expected was investigated. At first the solutions used were the ordinary buffer solutions of the usual strength. In order to approach more nearly to the buffer strength of natural waters, experiments were also made with buffer solutions at various degrees of dilution. Experiments were also made using solutions of sodium hydroxide, sodium carbonate, and calcium hydroxide, although we have left for future investigation detailed studies of the effects of specific salts.

It was found to hold very strictly that the "floc" which appeared first in any given series was invariably the best as far as flocculent appearance was concerned. It was also the "floc" which appeared first which possessed the qualities of rapid settling and abundance in the highest degree.

If we select those cases in which the total salt content-in this case buffer salts-was constant and in which the concentration of aluminium varied, we find, on plotting the final pH values against the flocculation time, that we have a family of curves of the general type shown in Figure 2.

In Figure 2 the concentration of "alum" was 400, 300, 200, and 100 p. p. m. for curves A, B, C, and D, respectively. The buffer

strength (total salt) was the same in all cases. The optimum pH value for the production of a "floc" in a minimum time appears to increase slightly as the concentration of alum decreases. The following values are taken from Figure 2:

100

Using 400 p. p. m., the optimum pH value is about 4.95.
Using 300 p. p. m., the optimum pH value is about 5.10.
Using 200 p. p. m., the optimum pH value is about 5.25.
Using 100 p. p. m., the optimum pH value is about 5.40.

[blocks in formation]

FIG. 2. The relation between time required for the first appearance of floc in solutions buffered at various pH values when the total salt concentration is constant and the alum concentration is varied A-400 p.p.m. of alum; B-300 p.p.m. of alum; C-200 p.p.in, of alum; D= 100 p.p.m. of alum.

When less than 100 p. p. m. are used, the optimum pH hovers about a value of 5.5, and it is this value which is of particular significance, since the amounts of "alum" ordinarily used are always considerably less than 100 p. p. m. (6 g. p. g.).

It is to be noted that as the concentration of "alum" decreases, the width of the curves also decreases. Using 400 p. p. m. of alum, the final pH value might be varied over a range of about 2.0 pII units, without any appreciable effect on the flocculation time. Using 100 p. p. m., however, the width of the optimum zone would be less than 1.0 pH unit. Using less than 50 p. p. m., it is necessary to adjust the solutions quite carefully in order to secure a "floc" within a reasonable time.

As stated above, these results were obtained with only slight agitation of the liquid made at the moment of observation and in 100-c. c. cylinders. Now, it is most interesting to note that the time of flocculation can be varied by forming the mix in vessels of different size. For instance, a large, shallow tank containing a large volume of buffer solution was allowed to deliver the buffer solution through a 1-inch pipe to an inclined trough. At the upper end of this trough there was delivered a small constant stream of alum solution. At the lower end a crosspiece formed a "hydraulic jump." The mixture, as it ran off the trough, was collected in vessels varying in capacity from 10 c. c. to 5,000 c. c., care being taken to avoid a graded order of collection. It was invariably found that the time for first appearance of floc was least in the large vessels. The same result was obtained by hand mixing and subsequent pouring of the mix into vessels of different size. The difference in time required for flocculation is remarkable, a precipitate forming within one minute in a large vessel and often requiring hours to become visible in small vessels. Though we recall no published data upon this phenomenon, we know that it is well recognized by plant operators, who have frequently observed that tests in laboratory vessels do not correspond to the flocculation in the basin. We have in one case plotted the time. of flocculation against volume of collecting vessel and found the curve to become flat near the 2,000-c. c. volume. While we have not pursued this problem in sufficient detail for practical application, we are confident that, if necessary, a relation could be found for calculating the basin time from a series of small-volume experiments. However, the "volume" effect was constant in our laboratory experiments, and this factor therefore does not alter our conclusions regarding relative pH effects.

We believe that this "volume effect" is little more than the effect of the volume-surface ratio upon circulation, for mechanical circulation will decrease the time required for floc formation. We conducted a series of experiments like those summarized in Figure 2 and, in addition, imparted to the solution a slow rotary motion. The total amount of floc produced seemed not to be affected, and the optimal pH remained the same, though, of course, the time values were altered from those shown in Figure 2.

A few experiments made with very small quantities of alum indicate that the curves relating pH and flocculation time would be of the type shown in Figure 2, but that the two branches would tend to be closer together than those shown. Since the branches of the curve tend in all cases to become parallel, it is evident that no floc would become apparent for a very long time in vessels of laboratory size when the pH lay beyond the asymptote to either branch. With very dilute alum solutions the region of pH within which floc might appear in laboratory vessels might well narrow to a zone less than 0.3 pH units wide. Experimentally it has been found that if a floc does not appear within a few hours with slight occasional agitation, it will not appear within a greatly reduced time when there is mechanical agitation. It would thus appear that under large-volume conditions a rigid control of the final pH is necessary for floc formation from extremely dilute alum solutions.

We stated in a former section that, without much confidence in its accuracy, we obtained a provisional value of 5.5 for the isoelectric point. With still less confidence in the sufficiency of the reasoning, we then suggested that a study of flocculation time would reveal the isoelectric point. We then arrived, by experimental methods, at the point pH 5.5 for the minimum coagulation time in dilute solution. An examination of the data of Buswell and Edwards (1922) on residual alum found under commercial conditions suggests that if there were included data for effluents of pH values lower than those observed, there would be obtained a curve relating pH and residual alum passing through a minimum residual alum point at about pH 5.5. Thus, whatever uncertainty there may be in each mode of approach, there is substantial agreement in the tentative results of each method.

Up to this point we have considered only those experiments in which the various experimental conditions were kept as constant as possible and, so far as was practicable, only the hydrogen ion concentration and the total aluminium were varied. Since experimenters in other fields have reported the effects on coagulation of the total salt, we have studied the effect of varying the concentration of the buffer solutions used.

The influence of the salt content is evident from an inspection of Figure 3. In the derivation of the experimental values, 100 p. p. m. of alum were used in all cases. The buffers for curves A, B, and C M M

are

M

20' 100 and 200' respectively. The width of the optimum zone

decreases as the buffer concentration, i. e., total salt, decreases. We would expect that a water of low alkalinity or low total salt content should require a much more delicate adjustment of the dose of alum than a water of high alkalinity or high salt content.

« ForrigeFortsett »