current

research

Water and Waste Water Filtration: Concepts and Applications

Kuan-Mu Yao,l Mohammad T. Habibian, and Charles R. O’Melia2

Dept. of Environmental Sciences and Engineering, University of North Carolina, Chapel Hill, N. C . 27514

particles t o the immediate vicinity of the solid-liquid interface

H A conceptual model for water and waste water filtration presented by the filter (Le., t o a grain of the media or t o
processes is presented and compared with the results of labora- another particle previously retained in the bed); and second,
tory experiments. Efficient filtration involves both particle the attachment of particles t o this surface (O’Melia, 1965;
destabilization and particle transport. Destabilization in Ives and Gregory, 1967).
filtration is similar t o destabilization in coagulation; effective Viewed in this perspective, filtration and coagulation pro-
coagulants are observed t o be effective “filter aids.” Particle cesses are quite similar. I n both processes, the particles t o be
transport in filtration is analogous t o transport in flocculation removed must be made “sticky” or, more formally, destabil-
processes. A particle size with a minimum contact opportunity ized. The considerable research on colloid destabilization
exists ; smaller particles are transported by diffusion while mechanisms during the last few decades can be used t o under-
larger particles are transported by interception and settling. stand these chemical aspects of filtration. In both processes,
Applications of these concepts t o water and waste water suspended particles must be transported so that contacts
filtration are presented. may be achieved. I n coagulation, the transport models of
Smoluckowski (1917) are used. These models predict that in
water the transport of particles larger than about 1p is accom-
plished by velocity gradients or Auid motion; for smaller

T
particles, Brownian diffusion is effective. In water filtration,
his paper is written with three objectives. First, mecha- transport models are being derived which are based on models
nisms for coagulation and filtration processes are developed by investigators in air filtration (e.g., Friedlander,
presented as analogous; similarities between these processes 1958). One such model is presented in this paper; others have
are particularly helpful in understanding filtration processes been presented by Spielman and Goren (1970) and Cookson
and in assessing their capabilities. Second, a theoretical or (1 970). These models predict that suspended particles larger
conceptual model of water and waste water filtration processes than about 1 p are transported t o the filter media by settling
is set forth; the results of experiments designed t o test this and interception; smaller particles are again effectively trans-
model are presented and discussed. Third, conclusions are ported by Brownian diffusion.
reached concerning the capabilities of filters for removing Transport Model. Let us begin by considering a single
pollutants in water and waste water treatment; suggestions spherical particle of the filter media. Assume that it is
are made for the design and operation of filters which accom- unaffected by its neighbors and is fixed in space in the flowing
plish this removal. suspension (Figure 1). This single particle of filter media
Coagulation and Filtration Processes. The overall rate of is a collector, emphasizing that the ultimate purpose of
aggregation in a coagulation process is frequently evaluated by transporting suspended particles from the bulk flow to the
determining the rate a t which collisions occur between par- external surfaces of media grains i n packed beds is the
ticles by fluid motion (orthokinetic flocculation) and by collection of these particles, thereby accomplishing their
Brownian diffusion (perikinetic flocculation), multiplied by a removal from the water. The main flow direction is that of the
“collision efficiency factor” which reflects the ability of gravitational force. A suspended particle following a stream-
chemical coagulants to destabilize colloidal particles and line of the flow may come in contact with the collector by
thereby permits attachment when contacts occur (e.g., Swift virtue of its own size (case A in Figure 1); this transport
and Friedlander, 1964; Birkner and Morgan, 1968; Hahn process is interception. If the density of the suspended particle
and Stumm, 1968). A similar approach is used herein to is greater than that of water, the particle will follow a different
describe filtration processes. The removal of suspended par- trajectory due t o the influence of the gravitational force field
ticles within a filter is considered t o involve at least two sepa- (case B). The path of the particle is influenced by the combined
rate and distinct steps: First. the transport of suspended effects of the buoyant weight of the particle and the fluid drag
on the particle. This transport process is sedinientcrfion.
Present address: Camp, Dresser. and McKee, Consulting Engi- Finally, a particle in suspension is subject t o random bombard-
neers. 1 Center Plaza. Boston. Mass. 02108. ment by molecules of the suspending medium, resulting in the
To whom correspondence should be addressed. well-known Brownian movement of the particle. The term
Volume 5, Number 11, November 1971 1105
 pattern is undisturbed by the presence of the grain, respec-
tively; d is grain diameter. Performance of a packed bed is
related to the efficiency of a single spherical collector:

--- STREAMLINE I
(3)

where /'is the bed porosity, L is bed depth, and a is a collision

COLLECTOR efficiency factor which reflects the chemistry of the system.
Following practice in coagulation (e.g., Swift and Friedlander,
1964), a is defined as a ratio-Le., the number of the contacts
which succeed in producing adhesion divided by the number
of collisions which occur between suspended particles and
the filter media. Ideally, a is equal t o 1 in a completely de-
A INTERCEPTION
stabilized system. Equation 3 is similar in form t o the first-
B SEDIMENTATION order equation used by Iwasaki (1937), Ives (1960), and others
/ /
C DIFFUSION to describe the effects of filter depth on filter performance.
Integration of Equation 3 yields the following:

C 3
In - = - - (1 -f)ap (4)
CO 2

difusion is used to describe mass transport by this process where C, and C are the influent and effluent concentrations
(case C). for a packed bed.
The general equation describing the temporal and spatial An impression of the magnitude of p in real systems can
variation of particle concentration in such a system may be be obtained by using a numerical example. Consider a con-
written as follows : ventional rapid sand filter with a bed depth of 24 in., a bed
porosity of 40%, and containing media with a size of 0.6
mm. Assume that the suspended particles t o be removed are
completely destabilized ( a = 1). If such a filter removes 90%
of the partides applied t o it (CjC, = 0.11, 7 is 2.5 X
where C is the local concentration of suspended particles, c (Equation 4). One is then led to ask what parameters affect 7 .
is the local velocity of the water, t is the time, Dbm is the Subsequently it will be shown that p depends not only on such
diffusion coefficient of the suspended particles, p and pi> are parameters as the filtration velocity, media size, and water
the densities of the water and the suspended particles, re- temperature, but also in a significant manner on the size and
spectively, p is the water viscosity, m and d],are the mass and density of the particles to be filtered.
diameter of the suspended particles, respectively, g is the I n this paper the results of a theoretical model for the de-
gravitational acceleration, and z is the coordinate in the direc- termination of p are presented. The results of laboratory
tion of the gravitational force. Equation 1 is derived from a experiments in which the filtration performance of packed
mass balance of C about a n elemental volume of suspension. beds is characterized by influent and effluent concentrations
The first term on the left-hand side of the equation ( a C j dr) (C, and C) are included. Comparisons of theoretical pre-
represents the temporal variation of C at any point with dictions of p with experimental results are made using Equa-
coordinates x , y , and z ; the second term ([Link]) describes the ion 4.
effects of advection on the concentration at that point. On NUMERICAL DETERMINATION OF p . The numerical solution
the right-hand side of Equation 1, the first term (Db,,,V2C) to the problem of predicting 7 and filter efficiency involves
describes the effects of diffusion, and the second term char- four steps (Yao, 1968; Yao and O'Melia, 1968): (1) deter-
acterizes the effects of gravitational settling on the system. mining the distribution of particles in the region close to the
The influence of interception is included in the boundary surface of a single collector; (2) calculating the rate at which
conditions used in integrating the equation. The form of particles strike the collector surface; (3) computing the single-
Equation 1 has been widely used by engineers t o describe the collector efficiency; and (4) calculating the overall removal
fate of pollutants in the atmosphere and in streams and estu- efficiency of a given packed-bed filter.
aries, and has been applied t o air and water filtration processes. I n step 1, the diffusion equation (Equation 1) is integrated
Equation 1 cannot be solved analytically; numerical pro- numerically t o yield the distribution of particles in the region
cedures and (or) simplifying assumptions may be used. of interest. Several assumptions are made. First, a steady state
THESINGLE-COLLECTOR EFFICIENCY. The contact efficiency is assumed; Le., a C / d t = 0. Second, Stokes equations for
of a single media particle or collector (7) is a ratio-i.e., the fluid velocities in laminar flow around a sphere are used
the rate a t which particles strike the collector divided by the in the advective term ( c . VC). Subsequent experimental re-
rate at which particles flow toward the collector, as follows: sults and the work of other investigators (e.g., Spielman and
Goren, 1970) suggest this assumption is not justified for an
accurate description of packed-bed systems. Third, Einstein's
equation is used to estimate the diffusion coefficient of the
suspended particles:

Here L',and C, are the water velocity and suspended particle

concentration upstream from the collector where the flow
1106 Environmental Science & Technology
where k is Boltzmann's constant, and T i s the absolute tem- clusions: There exists a size of the suspended particles
perature. Fourth, interception is included in the boundary for which the removal efficiency is a minimum. For the as-
conditions, which assume that C = C, at an infinite distance sumed conditions typical of conventional practice in water
from the collector, and that C = 0 at a distance equal to filtration, this critical suspended particle size is about 1 p.
(d + d,)/2 from the center of the collector. Finally, the For suspended particles larger than 1 p , removal efficiency
model applies most directly to clean filters, where deposition increases rapidly with particle size. Removal is accomplished
within the pores has not significantly altered the flow pattern by sedimentation and (or) interception. For suspended
or media characteristics. particles smaller than 1 p , removal efficiency increases with
In step 2, two additional numerical operations are used to decreasing particle size. Removal is accomplished by diffusion.
compute the total rate at which suspended particles strike (It is useful to note here that many suspended particles of
the collector. First, the particle fluxes at various points on interest in water and waste water treatment are about 1 p
the collector surface are calculated from the concentration in size or smaller. Included here are viruses, many bacteria,
gradients at the collector surface. These concentration a large portion of the clays, and a significant fraction of the
gradients are determined from the concentration distribution organic colloids in both raw and biologically treated waste
computed in step 1. Second, these particle fluxes are integrated water.)
numerically over the whole collector surface, yielding the ANALYTICAL DETERMINATION OF 7 . Equation 1 can be solved
rate at which particles strike the collector. analytically to determine the single-collector efficiency if only
In step 3, the single collector efficiency is calculated directly one transport mechanism is operative. Assumptions concern-
from the results of step 2 using Equation 3. Finally, in step 4, ing flow velocity (Stokes), steady state, and, where appro-
the efficiency of a packed-bed filter is calculated directly priate, the diffusion coefficient and the boundary conditions
from 7 using Equation 4. Here the sticking factor is generally are made which are identical to those used in the numerical
assumed to be 1. procedure. The case of diffusion alone has been developed by
The results of numerical calculations of 7 and filter efficiency Levich (1962) ; methodology for considering sedimentation
as a function of the size of the suspended particles are pre- and interception alone is presented elsewhere (Yao, 1968).
sented in Figure 2. These results lead to the following con- The following results are obtained:

5 (-J
3 d,
VI = (7)

- NUMERICAL SOLUTION

Here T D , v I , and 7~ represent theoretical values for the single-

collector efficiency when the sole transport mechanisms are
diffusion, interception, or sedimentation, respectively, and
Pe is the Peclet number.
Equations 6-8 are presented in Figure 3, where the appro-
priate single-collector efficiency is plotted as a function of the
size of the suspended particles. Included as points in Figure 3
d :0.5 mm are the results of the numerical analysis presented earlier
(Figure 2A). It is apparent that for the conditions used in
T = 25OC these calculations, the single-collector efficiency calculated
10-5

10-2 10-1 I 10

100

80

60
w )"'A
40

2o t a = 1.0

10-I I 10 I02

SIZE O F THE SUSPENDED PARTICLES

(microns 1 10.2 10-1 I 10 102
S I Z E OF THE SUSPENDED P A R i I C L E S ( m i c r o n s )
Figure 2. Theoretical model for filtration efficiency with single-
collector and removal efficiencies as functions of the size of the sus- Figure 3. Comparison of numerical and analytical solutions of Equa-
pended particles tion l

Volume 5, Number 11, November 1971 1107

numerically can be approximated by the sum of the analytical The final suspensions had a p H of 8.3. Absorbance measure-
expressions. I n other words, ments were made at appropriate wavelengths with a Beckman
Model DB spectrophotometer to determine particle concen-
7 = 1)D -k ?I1 $- 770 (9) trations in samples from the filter influent and effluent.
The analytical expressions (Equations 6-8) combined with Glass beads supplied by the Minnesota Mining and Manu-
Equation 4 provide a convenient picture of the effects of con- facturing Co. were used as filter media. These beads were
ventional filtration variables on filter efficiency as predicted sieved for uniformity; the mean size of the sieved beads was
by the model. The right-hand side of Equation 4 is seen to 0.397 mm with a standard deviation of 0.0145 mm. Filter
vary with cQ t o c-l, p 0 t o p-l, d-l t o d-3, and d p - 2 / t3o d p 2 beds were generally 14 cm in depth, 2.6 cm in diameter, and
depending o n the transport mechanism which is operative. had a porosity of 36 %.
These results correspond t o the range of results observed by Both the suspended latex particles and the glass beads are
other investigators in laboratory experiments and in practice negatively charged in water. To provide for efficient attach-
(Ives and Sholji, 1965). The effects of filtration velocity, water ment (a approximating l), a destabilizing chemical must be
viscosity, media size, and the density of the suspended used. I n the experiments described in this paper, a cationic
particles o n the single-collector efficiency are presented graph- polymer, diallyldimethylammonium chloride (Cat-Floc) sup-
ically in Figure 4. F o r particles larger than 1 p , this model plied by the Calgon Chemical Corp. has been used. This
predicts that the density of the suspended particles exerts polymer is reported by the manufacturer t o have a molecular
significant effects on filtration due to settling (Figure 4D). weight in the order of 5 X loj; its positive charge is constant
Other conventional filtration parameters exert considerably below p H 11.
less effect on the process (Figures 4A-C). Two methods of applying this cationic polymer have been
used. I n one series of experiments the filter beds were pre-
coated with a concentrated polymer solution (10,000 mg/liter)
Experimental
prior t o use; no additional polymer was applied during the
Latex beads supplied by the Dow Chemical Co. have been filtration runs. The purpose of these studies was t o determine
used to prepare suspensions for filtering. Polystyrene latex experimentally the single-collector efficiency at the start of a
particles with 0.091-, 0.357-, and 1.099-p diameters and styrene filter run when the theoretical model could be expected t o
divinylbenzene copolymer latex particles with 7.6- and 25.7- characterize the system. In a second series of experiments the
p diameters were selected for use. Experiments were thus filter beds were precoated and, in addition, polymer was fed
conducted within and on both sizes of the critical size where it throughout the duration of the filter runs. The dosage of
was expected that filter efficiency would be poorest. These polymer was selected on the basis of jar tests similar to those
particles have a density of 1.05 gm/cc. used in coagulation processes. In these experiments the head
Suspensions for testing were made by diluting the stock loss developed during the filtration process was measured
supplied by the manufacturer t o a suitable latex concentration using piezometer tubes connected to the inlet and outlet of
(10 to 200 mg/liter) with 10-3M NaCl and 10-3M NaHCO,. each filter.

IO" 10-1

10-2 10-2

10-3 I0-3

10'4

Figure 4. Theoretical model 10-2 10-1 I 10 I02 10-2 10-1 I 10 I02

for the single-collector
efficiency: effects of filtra-
tion velocity (v,,), tempera-
ture (T), media size (d),
and the density of the sus- -
pended particles ( p p ) L

10-2
__ d = 0.5rnrn
d : [Link]
vg: P g p m / s q . f t .

10 - 3

10-2 10-1 I 10 IO2

SIZE OF THE SUSPENDED SIZE OF THE SUSPENDED

PARTICLES (microns) PARTICLES (microns)
1108 Environmental Science & Technology
Results the single-collector efficiency and the size of the suspended
Typical results of those experiments in which only a pre- particles is in reasonable agreement with the model. The
coating of polymer was used are presented in Figure 5. comparison suggests that the transport mechanisms used in
Removal efficiency is plotted as a function of filtration time. developing the model are in fact operative in filtration.
Data for both coated and uncoated media are presented for In other words, diffusion is operative in transporting small
comparison. In these experiments clean water is passed particles, while settling and interception are able to transport
through the filters for several minutes while the flow rate is particles larger than about 1 ,u in size.
established; the latex suspension is then introduced into the Experimental filter efficiencies are higher (better) than
apparatus at zero time (Figure 5). Clean water in the filter theoretical predictions. Possible reasons for these discrepan-
apparatus requires about 2 t o 3 min for displacement; after cies are discussed subsequently.
this time a n additional 1 or 2 min is required t o elute latex The results of experiments using several filtration rates are
suspension which has mixed with the original clear water. presented in Figure 7. The single-collector efficiency is plotted
After about 5 min, the effluent from the filter represents un- as a function of the filtration velocity. The latex particles used
diluted latex suspension. in these experiments had a diameter of 0.091 ,u so that trans-
Very little removal is accomplished by the uncoated filter ; port by diffusion was operative. The predictions of the Levich
after 5 min the effluent concentration equals or exceeds 95 z model (Equation 6) for diffusion alone are plotted in Figure 7
of the influent concentration. Using a polymer-coated filter, for comparison. According to this model, 7 varies with u - * / ~ ;
4 4 zof the latex particles is removed; this corresponds t o a the experimental data are in reasonable agreement with this
single-collector efficiency (Equation 4) of 1.6 X In this prediction. Again, experimental filters are observed to ex-
system, negatively charged latex particles are able t o adhere to ceed the performance predicted by the conceptual model out-
the positively charged filter media. lined previously in this paper.
Results of experiments using five different sizes of latex
particles are summarized in Figure 6. The removal efficiency
of the packed beds and the single-collector efficiency are
plotted as a function of the diameter of the suspended latex
particles which are filtered. The predictions of the theoretical
model are also presented in Figure 6 for comparison. This ~-2-i- EXPERIMENTAL DATA

comparison reveals:
A suspended-particle size with a minimum opportunity
for removal is observed to exist; this is in agreement with the
THEORETICAL MODEL
/
model. Furthermore, the magnitude of this critical particle
size (about 1 ,u) is in good agreement with the predictions of
the model.
The general trend in the observed relationship between

5w 100

3
-I
LL
LL 1 I I
w 80
z
a=:
60 IO 0 I I
zs
z: -5-2- EXPERIMENTAL DATA
52
E O Ll.
z Y 40 0
E:a
20 MUNCOATED MEDIA
z
W
0 MCOATED MEDIA

'
Z
0 0
0
0 5 IO

lf=0'36
\J
TlME(rninutes)

20
L: 5.5 in.
yo= 2 g p m / s q . f t .
I
vo= 2gpm/[Link]. T = 23OC
a = I (assumed )
d = 0.397 mm f = 0.36
dp = 1.10 microns L = 5.5 in.
pp = 1.05 grn/cm3 PolymerCooting =CAT-FLOC
1 I 1 I
(no polymer used after t = O 1 10-2 10-1 I 10 io2

Figure 5. Typical experimental results for coated and uncoated filter S I Z E OF SUSPENDED PARTICLES (microns)
media Figure 6. Comparison of theoretical model and experimental data

Volume 5, Number 11, November 1971 1109

 10-1 , I I The filtration capacity is exhausted after only a few minutes.
At the optimum polymer dose (filter no. 4), the effluent
concentration is significantly better than that observed in the
earlier experiments. When polymer is added continuously, a
removal efficiency of 93% of these 0.09-p latex particles is
*
0 observed after about 1 hr, compared with a bed efficiency of
Z
w_ 61% when only a precoat of polymer is used (Figure 6B).
10-2 This is probably not due to coagulation in the filter pores
2
LL
LL since particle growth of these small particles (0.09 p ) would
LI
a lead to less efficient filtration (6B).
0 During the first hour of filtration a n initial relatively
I-
o high concentration of particles appears in the effluent, after
W
J
J which the removal efficiency improves considerably. This
0
0 initial breakthrough would be less noticeable if a deeper and
I

W
10-3 more conventional filter bed were used.
_I
(3 At the optimum polymer dose, filtration (transport
z DIFFUSION
and attachment) is so effective that the available head loss
i
3 T = 2 5 ' 2 2OC
dp: 0.091 micron is utilized in about 3 hr. It is significant t o note that this occurs
d :0.337,rnm even with particles with a size in the order of 0.1 p.
L = 5 . 5 in.
f :0.36 These results suggest that when conventional filters fail
a = Ilarrurnedl to produce efficient filtration, effective improvements can be
10.~ made by altering the chemistry of the system. I n other words,
10-1 I attention should be directed toward increasing a , rather than
FILTRATION R A T E ( g p rn / s q . ft.)
merely changing, such conventional filtration parameters as
d, t i , and L.
Figure 7. Comparison of theoretical models and experimental data Transport is so efficient in water and waste water filters
that when polymers are used t o improve a , the filtration pro-
cess will be so effective that short filter runs will result with
conventional beds due to rapid clogging of the filter pores.
It is probable that effective filtration without excessive head
I n these experiments where precoated filter media were
loss can be achieved with polymers and dual medium filters,
used without any additional use of destabilizing chemical,
upflow filters, biflow filters, moving bed filters, etc.
it was expected and observed that the ability of the filter
Two additional statements can be made on the basis of
beds t o remove particles would be exhausted in a short time.
other experiments. First, if polymers are used continuously
I n this system the filter media can retain only a monolayer of
but without a precoat of polymer applied to the filter media;
particles; after this time the latex particles in suspension
the time for filter ripening can be several hours. Latex par-
collide with negatively charged latex particles previously
ticles with polymer adsorbed at the optimum dose for co-
removed in the bed. I n any real filtration process, the particles
agulation may be only partially removed by negatively charged
to be removed must be able to adhere to each other on con-
media. In practice such precoating could easily be achieved by
tact; this can be achieved by continuous addition of polymer.
adding polymer to the backwash water (Harris, 1970). Second,
An important question then arises-how much polymer is
the optimum concentration of polymer required for filtration
needed? To answer this question, it is proposed that the chem-
depends on the concentration of colloids t o be filtered.
ical aspects of filtration are similar to the chemical aspects of
Here again, filtration is analogous to coagulation. Stoichi-
coagulation. If this is so, then jar tests used to determine chem-
ometry in coagulation has been reported by many investigators
ical dosages for coagulation could serve the same purpose for
(e.g., Black and Vilaret, 1969; Stumm and O'Melia, 1968).
filtration.
The results of experiments designed to test this hypothesis
are presented in Figure 8. The results of jar tests are depicted Discussion
in 8 A ; residual turbidity after settling is plotted as a function Let us consider here some plausible causes for the dis-
of the applied dosage of cationic polymer. An optimum poly- crepancies between model and observation (Figures 6 and 7).
mer dosage of 0.07 mg,'liter is observed for this suspension. First, the assumption that Stokes equation for the velocity
Vertical arrows correspond t o dosages selected for study in pattern about an isolated sphere can describe the velocity
filtration. The results of a filtration experiment are presented in distribution in a packed bed with a porosity of 36 % is prob-
8B and C; effluent concentration and head loss are plotted as a ably unrealistic. Pfeffer (Pfeffer, !964; Pfeffer and Happel,
function of filtration time. Clean water again requires a period 1964) has used the cell model developed by Happel (1958)
for displacement. Based o n these results the following state- to describe the velocity terms characterizing mass transfer by
ments can be made : diffusion in packed beds having porosities of 40 % and higher.
Effective filtration is achieved using the optimum poly- Cookson (1970) has applied this model t o the filtration of
mer dose observed in the coagulation (jar) tests. viruses. The results of the Pfeffer and Happel model are similar
Underdosing and overdosing with polymer are observed. t o the Levich equation (Equation 6), with the addition of a
Again, these phenomena are observed in jar tests. Overdosing porosity term:
in this case is probably due to sufficient adsorption of the cat-
ionic polymer to produce charge reversal of the latex particles. 7 = 4BPe-*I3 (10)
When no polymer is added to a precoated filter bed
1- y5 19
(filter no. l ) , the variation of effluent concentration with
time is similar to that observed in earlier tests (e.g., Figure 5 ) .
where B = 1.26 (-w-) ; w = 2 - 3y + 3y5 - 2 ~ ~ ;

1110 Environmental Science & Technology

 1

120 - -
n e
f
5 6
-
60 - -
40 - ~ 0 2 9 . 7m g / l -
-
A

-
I I
FILTER SYMBOL POLYMER
-

-
NO. DOSE(mg/l)
1 - 0

- -
2 0.0014
3 0.007
4 0.07 -
5 0.7
6 - 7 -
6
I I

100 I I I I I I
I
POLYMER :CATFLOC
80
d p : 0.1 m i c r o n s
C O S 9.7mg/1
60 d = 0 . 3 9 7 mm
L :5 . 5 in.
40 vo = 2 gpm / [Link].
I d f 0.36 I
20 T : 25OC
C
0 I
0 50 100 150 200 250 300
FILTRATION TIME (minutes)
Figure 8. Comparison of jar test results (A) with filter performance (Band C)

and f = 1 - [Link] a packed bed with a porosity of 36%, will deviate from a n undisturbed streamline (Le., a fluid
B = 3.81, and 7 = 15.24 Pe-*i3. streamline unaffected by the presence of the suspended par-
The results of the Pfeffer and Happel model are presented in ticle) due to the difficulty in draining fluid from between the
Figure 7, together with the Levich model and the results of collector and the particle as the distance between them nar-
experiments describing the effects of filtration on 7. As noted rows. London-van der Waals forces are invoked t o overcome
earlier, experimental points are computed from observed this otherwise slow drainage. These authors developed a model
removal efficiencies of packed beds using Equation 4 and which predicts that when interception alone is operative the
assuming Q = 1. The observed data are consistent with the filter coefficient can in fact vary with the first power of the
Pfeffer and Happel model if 01 is 0.5. I n other words, if half of suspended particle size, as well as inversely with the 2.5 power
the contacts between the negatively charged latex particles of the collector size and the 0.25 power of the filtration rate.
and the polymer-coated filter media are successful in achiev- Next let us consider that experiments using a continuous
ing attachment, these data are quantitatively consistent with dosage of polymer indicated that filtration efficiency after a
the Pfeffer and Happel model for mass transport by diffusion ripening period was considerably better than that observed
in packed beds. when a precoated filter media was used alone, and a n initial
All terms o n the right-hand side of Equation 4 are com- ripening period always occurred. Possible explanations for
monly combined in a filter coefficient, X (e.g., Ives, 1960); these phenomena include one or more of the following:
in the model described herein, X = 3(1 - f)a7,'2d. When Latex particles coated with the optimum dosage of
interception is the dominant transport mechanism this model polymer may find it more difficult t o adhere to polymer-
predicts that the filter coefficient is independent of the flow coated glass beads than to previously retained latex treated
rate and varies as the square of the suspended particle size with the optimum polymer dosage. I n other words, CY for a
and the inverse cube of the media size. Limited experimental latex-glass bead interaction could be lower than for a latex-
results (Figure 6A) suggest that 7 and consequently X vary latex interaction, even when all surfaces receive the appro-
with the first power of the suspended particle size when inter- priate polymer treatment.
ception and settling are dominant. Spielman and Goren For particles greater than 1 p, coagulation within the
(1970) note that a suspended particle approaching a collector pores could produce a n increase in the effective size of these
Volume 5, Number 11, November 1971 1111
particles and enhance their removal. This cannot explain the particles present in dilute but objectionable concentrations;
results in Figure 8 but could contribute to similar results prior coagulation is not required, but a destabilizing chemical
observed in other experiments with larger suspended particles. must be added t o the filter influent. Such processes could save
Coagulation within filter pores will depend upon the concen- capital costs and some chemical operating costs, but would
tration of particles in suspension; for dilute suspensions suffi- require close operating control.
cient contact opportunities would not be available. Unlike the quality of the filtrate, the head loss developed
I t is possible that particles which are removed may act as during filtration is very dependent upon media size, filtra-
collectors themselves. Recall that X varies with d-’ to d P 3 ; tion rate, and the concentration of particles t o be filtered.
if small particles which have been removed can act as col- The ability of polymers t o enhance attachment in filters re-
lectors, X and 7 can be large. Billings (1966) has photographed quires that conventional filters be redesigned t o provide larger
the accumulation of latex particles on glass rods in air filtra- pores. Multimedia beds, radial beds, moving beds, or other
tion and showed the development of fibrous strands of indi- modifications are required. For example, it seems plausible
vidual latex spheres which extended considerable distances that direct filtration of secondary effluents could provide
into the airstream. After a n initial period most removal was both long filter runs and efficient removal if upflow filters with
achieved by attachment t o other latex particles previously depths in the order of 4 ft and media having a size of about
retained, rather than t o the considerable amount of remaining 1.5 m m were used at filtration rates in the order of 5 gpm/ft2.
available surface o n the glass rods.
Literature Cited
Conclusions Billings, C. E., “Effects of Particle Accumulation in Aerosol
Based on the conceptual model and the experimental re- Filtration,” W. M. Keck Laboratory, California Institute of
Technology, Pasadena, Calif., 1966.
sults presented in this paper, we conclude that conventional Birkner, F. B., Morgan, J. J., J . Amer. Water Works Ass. 60,
sand filters provide ample contact opportunities for the re- 175-191 (1968).
moval of all particles applied to them. When such filters are Black, A. P., Birkner, F. B., Morgan, J. J., J. Amer. Water
not producing efficient removal, the chemistry of the system Works Ass. 57, 1547-1 560 (1965).
Black, A. P., Vilaret, M. R., J. Anier. Water Works Ass. 61,
( a ) should be changed. A great variety of destabilizing cheni- 209-214 (1969).
icals is currently available for this purpose. Examples in- Cookson, J. T., Jr., ENVIRON. [Link]. 4,128-134 (1970).
clude hydrolyzing metal salts [e.g., salts of Al(l1I) and Fe(ll1)I Friedlander, S. K., Znd. Eng. Chern. 50, 1161-1164 (1958).
and natural and synthetic polymers which may be organic Hahn, H. H., Stumni, W., J . Colloid Interjace Sci. 28, 134-144
or inorganic. These latter materials can be cationic, anionic, (1968).
Happel, J., [Link].E. J . 4, 197-201 (1058).
or nonionic and can contain one or more of several types of Harris, W. L., J . Amer. Water Works Ass. 62, 515-519 (1970).
functional groups. These chemicals probably enhance attach- Ives, K. J., Proc. Inst. Cicil Eng. 16, 189-193 (1960).
ment in filtration by adsorption to produce charge neutraliza- Ives, K. J., Gregory, J . , Proc. SOC. Water Treaf. Exam. 16,
tion and (or) bridging. These destabilization mechanisms have 147-169 (1967).
Ives, K. J., Sholji, I., Proc. Amer. [Link] Eng. in J. Sanit.
been shown to be effective in many coagulation processes Eng. Dir., 91 (No. SA 4), 1--18(1965).
(e&, Stumni and O’Melia, 1968; Black et al., 1965; Hahn and Iwasaki, T., J . Amer. Water Works Ass. 29, 1591-1597 (1937).
Stumm, 1968). As in full-scale coagulation processes, the Levich, V. G., “Physicochemical Hydrodynamics,” Prentice
type and optimum dosage of chemical for filtration can be Hall, Englewood Cliffs, N.J., 1962, pp 80-85.
determined using jar tests. The wide diversity in the chemicals O’Melia, C. R., discussion of “Theory of Water Filtration”
by T. R . Camp, J . Sanit. Eng. Dic., Proc. Amer. [Link]
available should provide adequate means for filtering such Eng. 91 (No. SA 2), 92-98 (1965).
substances as bacteria, viruses, the colloidal calcium phosphate Pfeffer, R., Ind. Eng. Chem. Fundum. 3, 380-383 (1964).
precipitates often produced in tertiary treatment for phosphate Pfeffer, R., Happel, J., [Link].E. J . 10, 605-611 (1964).
removal, and organic biocolloids present in secondary emu- Smoluchowski, M., [Link]. Chem. 92, 129-168 (1917).
Spielman, L. A., Goren, S. L., [Link]. TECHNOL. 4,
ents, in addition to the clays and metal hydroxide flocs which 135-140 (1970).
are filtered in conventional water treatment plants. Stumm. W.. O’Melia. C. R.. J. Amer. Water Works Ass. 60,
To a first approximation, the removal eficiency of a packed- 514-539 (1968). ’

bed filter is independent of the applied concentration of par- Swift, D. L., Friedlander, S. K., J . Colloid Sci. 19, 621-647
ticles (Equation 4). Here a significant difference arises in (1964).
Yao, K’. M., “Influence of Suspended Particle Size on the
comparing coagulation and filtration processes. The detention Transport Aspect of Water Filtration,” unpublished Ph.D.
time required t o achieve a given degree of aggregation in dissertation, University of North Carolina, Chapel Hill,
coagulation depends upon the concentration of particles t o be N.C., 1968.
aggregated. Consider, for example, a water containing 10,000 Yao, K. M., O’Melia, C. R., “Particle Transport in Aqueous
Flow Through Porous Media,” presented at 16th Annual
viruses per milliliter and n o other colloidal particles. Using Conference of the Hydraulic Division of ASCE,MIT, Cam-
Smoluchowski’s analysis for perikinetic flocculation, 200 bridge, Mass., Available as ESE publication no. 210, Dept.
days would be required t o halve the initial particle concen- of Environmental Sciences and Engineering, University of
tration even if all the virus particles were completely de- North Carolina, Chapel Hill, N.C., 1968.
stabilized (cy = l). Furthermore, the resulting aggregates
would still be too small t o be removed by gravitational settling. Receired for review October 26, 1970. Accepted February 11,
The removal of viruses within a reasonable detention time in 1971. This research was supported in part by the Bureau o j
coagulation processes requires the presence of a large number Water Hygiene, Department of Health, Education and Welfare,
of other colloidal particles or enmeshment in a voluminous under grant no. EC-00296, and by the Federal Water Quality
Adniinistratim, Department of Interior, under grunt no. 17030
precipitate of metal hydroxide. This suggests that filters can FQU. One of us (M.T.H.) receiced fellowship support from the
be effectively used in water treatment t o remove colloidal World HealtA Organization.

1112 Environmental Science & Technology