INTERMITTENT SAND FILTRATION OF SEWAGE
Part I-Current Operational Practices
and Preliminary Laboratory Investigations
D. L. EMERSON, JR.
Associate Research Engineer
Bulletin .of the
COLLEGE OF ENGINEERING
UNIVERSITY OF FLORIDA
FLORIDA ENGINEERING AND INDUSTRIAL EXPERIMENT STATION
Bulletin No. 9, November, 1945
The Florida Engineering and Industrial
The Engineering Experiment Station was first approved by
the Board of Control at its meeting on May 13, 1929. Funds
for the Florida Engineering and Industrial Experiment Station
were appropriated by the Legislature of the State of Florida in
1941 The Station is a Division of the College of Engineering
of the University of Florida under the supervision of the State
Board of Control of Florida. The functions of the Engineering
and Industrial Experiment Station are:
a) To develop the industries of Florida by organizing and
promoting research in those fields of engineering, and the re-
lated sciences, bearing on the industrial welfare of the State.
b) To survey and evaluate the natural resources of the
State that may be susceptible to sound development.
c) To contract with governmental bodies, technical societies,
associations, or industrial organizations in aiding them to solve
their technical problems. Provision is made for these organ-
izations to avail themselves of the facilities of the Engineering
and Industrial Experiment Station on a co-operative financial
basis. It is the basic philosophy of the Station that the indus-
trial progress of Florida can best be furthered by carrying on
research in those fields in which Florida, by virtue of its loca-
tion, climate, and raw materials, has natural advantages.
d) To publish and disseminate information on the results
of experimental and research projects. Two series of pamphlets
are issued: Bulletins covering the results of research and in-
vestigations by staff members; and Technical Papers, reprinting
papers or reports by staff members which have been published
For copies of Bulletins, Technical Papers or information on
how the Station can be of service, address:
The Florida Engineering and Industrial Experiment Station
College of Engineering
University of Florida
JOSEPH WEIL, Director
PUBLICATIONS OF THE FLORIDA
ENGINEERING AND INDUSTRIAL EXPERIMENT STATION
As long as the supply is adequate, copies of available publi-
cations are free for general distribution. Address all requests
to: The Director, Florida Engineering arid Industrial Experi-
ment Station, University of Florida, Gainesville, Florida.
No. 1 "The Mapping Situation in Florida", by William L. Sawyer.
No. 2 "The Electrical Industry in Florida", by John W. Wilson.
No. 3 "The Locating of Tropical Storms by Means of Associated
Static", by Joseph Weil and Wayne Mason.
No. 4 "Study of Beach Conditions at Daytona Beach, Florida,
and Vicinity", by W. W. Fineren.
No. 5 "Climatic Data for the Design and Operation of Air Con-
ditioning Systems in Florida", by N. C. Ebaugh and
S. P. Goethe.
No. 6 "On Static Emanating from Six Tropical Storms and Its
Use in Locating the Position of the Disturbance", by
S. P. Sashoff and Joseph Weil.
No. 7 "Lime Rock Concrete Part 1", by Harry H. Houston
and Ralph A. Morgen.
No. 8 "An Industrial Survey of Hides and Skins in Florida", by
William D. May.
No. 9 "Studies on Intermittent Sand Filtration of Sewage -
Part I", by D. L. Emerson, Jr.
TECHNICAL PAPER SERIES
No. Heats of Solution of the System Sulfur Trioxide and
Water, by Ralph A. Morgen.
No.2 The Useful Life of Pyro-Meta and Tetraphosphate, by
Ralph A. Morgen and Robert L. Swoope.
No. 3 Florida Lime Rock as an Admixture in Mortar and Con-
crete, by Harry H. Houston and Ralph A. Morgen.
No. 4 Country Hides and Skins, by William D. May.
No. 5 An Empirical Correction for Compressibility Factor and
Activity Coefficient Curves, by Ralph A. Morgen and
J. Howard Childs.
No. 6 Crate Closing Device, by William T. Tiffin.
Florida Engineering and Industrial Experiment Station
University of Florida
Bulletin No. 9
INTERMITTENT SAND FILTRATION OF SEWAGE
Part I-Current Operational Practices
and Preliminary Laboratory Investigations
D. L. EMERSON, JR.
Assuriate Research Engineer
COLLEGE OF ENGINEERING
UNIVERSITY OF FLORIDA
This is the first of a series of Bulletins from the Sanitary
Engineering Division of this Station. One of the projects of
this Division is a study of sewage treatment under Florida
conditions and the development of complete systems of sewage
disposal adapted to the needs of the smaller communities of
the state. One such system, the sand filter, appears to have
merit, and a program has been set up to study sand filtration
with respect to the availability of suitable sands and the results
to be obtained from their use.
The initial step in such a program involves a study of the
previous work in other areas and the possible processes to use.
Since the information is so widely scattered in the literature,
it was felt advisable to summarize it in concise form for the
benefit of the sanitary engineers and interested communities of
the state. This compilation should provide a handy reference
to those who plan sewage disposal plants at this time.
It is well known that Florida possesses many sand deposits.
However, when sand is used for sanitary "filter" purposes it
must have certain properties. Data are presented on a variety
of deposits located in many sections of the state. Those com-
munities which do not have adequate sand beds "in situ" may
be enabled to locate a desirable sand within reasonable hauling
As more information is obtained the results will be published
in future bulletins.
JOSEPH WEIL, Director
TABLE OF CONTENTS
Foreword .... ........... .......................-.... ...................... .....-- ..........-- 2
Introduction ................. ........... ...................... .......................................
H historical ........ ........ ................ ....................-...................- ......... .... ........ 5
General Considerations ..................................................... ..................... 6
D definition ... .................................................... ........ -.......... -. .............. ... 6
Natural and Artificial Filters .......................................... ............... 7
Application ....... ........................................................... ...... 7
Efficiency .................. .... .......................................................... ...................... 8
Operation and Maintenance ....................................... .............. 9
Dosing Filters .............. ......... ..... .......... ..................... ........... 9
Cleaning Filters .... .. ...... ........................ ......................................... 10
Resurfacing ............................................................. 10
Overloading .................. ... ....... .. ............................................. ..... .. 10
Pending on Filter Surface ....................................... ................ ...... 11
Temperature Effect on Operation ....................................... ........... 11
Controlling Factors .............................................................. ............. 11
Design and Operational Experiences ........................................... ....... 12
Sand Characteristics ........... .............. .. ....................... ....... ..... 12
Homogeniety ............................. .. .........................- ..-.......-... 12
Uniformity and Size ...... .................................. ..... ........ 12
Uniformity Coefficient ................................................ .......... 18
Effective Size ........................ ...................................... ........ 18
C leanness .................................................................................. ............. 14
Shape of Sand Grain and Porosity of Filter .................................. 14
Depth of Sand Beds and Underdrainage ................................... .......... 1
Rate of Distribution on Filter ............................................................... 17
Sewage Loading ...................................................................................... 18
Preliminary Laboratory Investigations ............ ....... ............................. 19
Properties of Natural Florida Sands ..................................................... 19
Laboratory Intermittent Sand Filters ............................................ ........... 22
Reduction in 5-Day B.O.D. in Laboratory Filters .................................. 24
Preliminary Observations and Conclusions .............................................. 24
Acknowledgments ................... ................................................................ 27
Bibliography .... .. .... ... .................................................. .. 7
The Florida Engineering and Industrial Experiment Station
has begun a program of research directed toward the develop-
ment of economic sewage and industrial waste disposal processes
particularly suited to the requirements of this state. Most of
the research in sanitary engineering processes heretofore has
been confined to northern temperate climates (England, Ger-
many and northern United States) and the resulting develop-
ments have naturally conformed to the demands of those climates
and with the materials available in those locations. The sani-
tary engineers in Florida have long recognized the need for
modifying those conventional processes to produce greater
economy and operating efficiency under Florida's particular set
of conditions. It is also desirable to investigate the possibilities
of utilizing Florida materials for economy in the initial cost of
construction. Large quantities of material, i.e., fine and course
filter aggregate, etc., have in the past been imported from other
states at high cost of transportation due to the lack of this
latter information. This has made the cost of small installations
prohibitive and has retarded the development of sanitary pro-
grams throughout the state.
Since many of Florida's smaller communities have inade-
quate sewage disposal facilities due to the high cost of installa-
tion and operation, the Engineering and Industrial Experiment
Station has incorporated in its research program a project to
investigate intermittent sand filtration of sewage as applied
to conditions met in the smaller communities of this state.
Where sufficient quantities of sand of the proper physical char-
acteristics are available, it is expected that this method will,
by comparison with other methods, result in the lowest cost of
installation, operation and maintenance, and produce the highest
degree of treatment. Florida's mild climate greatly reduces
the cost of maintenance of this type of treatment over that
required in colder climates.
Sewage filtration was first employed as the entire treatment
method in the irrigation of land at Bunzlau, Germany, in 1559.
The General Board of Health of London in 1854, recommended
that sewage be spread on the land at safe distances from towns.
An Act of 1854 gave to the towns facilities for transporting
the sewage to outlying districts and the development of land
irrigation or filtration methods followed. The cultivation of
the land, however, was soon made secondary to the securing
of adequate treatment.
One of the first scientific approaches to sewage filtration was
made by Sir Edward Frankland, a member of the Royal Com-
mission on Sewage Disposal, in 1868. This was not the first
time that science had been applied to sewage treatment, but
was the first record of sewage disposal practice based on scien-
tific work. Sir Edward (5)* showed in the report of the Rivers
Pollution Commission of Great Britain for 1870 that the process
was not entirely mechanical, but that oxygen or aeration was
a necessary factor for good operation. He found that clogging
tendencies were reduced with reduction of applied sewage to
the filters and that the best results were obtained when the
filters were allowed to rest for three to four days after receiv-
ing an application of sewage. From 1868 to 1871, he worked
with various types of filtering materials the results of which
have since formed the basis of the modern development of
biochemical sewage purification.
The Massachusetts State Board of Health, by the Act of
1886, established the Lawrence Experiment Station, and began
a movement to try out the methods of Sir Edward Frankland.
Six years later, in 1892, their first results appeared in the Massa-
chusetts State Board of Health Annual Report (13). The early
studies of the Lawrence Experiment Station showed that oxi-
dation within the filter is due to biological action as well as
Intermittent sand filtration may be defined as the intermit-
tent application of settled or unsettled sewage to inclosed level
beds of sand provided with underdrains to receive and discharge
the filter effluent. The purification of sewage by this process
is accomplished by both physical and biochemical means. Physi-
cally, the upper inch of sand serves as a strainer in retaining
the coarser particles and the succeeding depth mechanically
supports a thin gelatinous coating of active microorganisms on
Figures in parentheses refer to Bibliography, page 27.
the sand grains which physically adsorbs colloidal and sus-
pensoidal particles. Biochemically, the adsorbed colloids and
suspensoids are oxidized to more stable forms by enzymes pro-
duced by the flora of active aerobic microorganisms inhabiting
The essential feature of the process is the balance of the
oxygen supply and the quantity of organic matter applied. The
even application of the sewage to the filters at intervals
(usually once in twenty-four hours) draws air behind it through
the interstices of the sand bed which is used by the aerobic
microorganisms in oxidizing the previously adsorbed putrescible
Intermittent sand filters may be classified under two heads,
natural and artificial. Where sufficient areas of sand of the
proper physical characteristics are available at the plant site,
natural sand filters may be prepared by: (1) clearing the area
of trees and undergrowth and stripping the top soil; (2) leveling
the area and subdividing it into beds of proper size by throwing
up the top soil to form partition embankments; (3) under-
draining the beds with open-joint tile pipe; and, (4) construct-
ing the distributing system. The shape and arrangement of
the various filters are determined by the various topographic
conditions encountered. The size of the beds is adjusted to the
unit sewage loading and the physical characteristics of the sand.
It must be possible to have one bed out of service at certain
times for cleaning or repairs without excessively overloading
the remaining beds.
In some sections of Florida the soil conditions are not such
that natural filters may be utilized. In these sections sand must
be transported from surrounding areas and artificial filters con-
structed along the principles described in this bulletin.
Formerly, intermittent sand filtration constituted the entire
sewage treatment process. Raw sewage was applied to the
filters at intervals of one to four days. Modern application of
intermittent sand filters is, however, like the activated sludge
and trickling filter processes, a form of secondary treatment.
The incoming sewage is screened and settled for one to two hours
before being applied to the filters. The primary settling may
take place in the conventional type settling tank, septic tank
or Imhoff tank. Presedimentation of the sewage prior to filtra-
tion greatly increases the quantity of sewage that may be ap-
plied to the filters and reduces the number of bed cleaning
Where a high quality effluent is required, other secondary
treatment processes, such as the trickling filter or activated
sludge processes, may precede the sand filters which serve as
a polishing or finishing treatment. The effluent from the com-
plete treatment closely resembles drinking water in clarity and
odor and may remain stable by the methylene blue test in-
Experience has demonstrated the fact that properly designed
and maintained intermittent sand filters will produce a higher
degree of purification than any other sewage treatment process
with a minimum of skilled supervision. Years of successful
operation of many intermittent sand filters in Massachusetts
have clearly demonstrated their permanence and efficiency.
When operated at reasonable rates in connection with a satis-
factory settling tank, sand filters should reduce the total organic
matter, as expressed by the bio-chemical oxygen demand
(B. O. D.) determination, by more than 95 per cent, and the
removal of bacteria under the same conditions should exceed
95 per cent. (1) Phelps, (19) in discussing sand filter experi-
ences at the Lawrence Experiment Station, reports that they
"had an excellent record for continuous and reliable performance
and, under favorable conditions of operation, could be relied upon
to remove consistently 98 or 99 per cent of the organic matter
and about the same proportion of the total bacteria." He fur-
ther relates that "any sewage treatment plant calls for con-
scientious and skilled operation at all times, but it seems to be
true that the farther we have come from the old-fashioned
accident-proof and nearly fool-proof sand filter, the more diffi-
cult has become the task of successful routine operation. If
the operation of sand filter may be likened to rowing a boat,
then the routine control of an activated sludge plant is more
like flying an airplane."
The above facts, along with the economy of installation, ap-
pear to make this type of treatment the most feasible for small
communities where sufficient sand of the right quality is located
at the plant site or where large filter areas are not required.
OPERATION AND MAINTENANCE
A predetermined dose of settled sewage is generally applied
to the filter beds from dosing tanks equipped with rotating
siphons set to discharge in a definite order and connected to
their respective beds. The filter influent may be distributed
through a system of wooden troughs, or if the bed is relatively
sma'l, it may discharge onto a concrete apron designed to pre-
vent erosion of the filter surface and decrease the velocity of
the incoming sewage.
When a new bed is first placed in service, the operator will
notice that most of the sewage will flow through the sand near
the inlets. After a few weeks of operation, however, some
solids will be retained on the surface nearest the inlets and
the distribution will become increasingly more uniform. For
best operation each section of the filter surface should treat
the same quantity of sewage.
The distribution system in well designed plants will rapidly
and completely flood each filter to a depth of two to three inches
and maintain this depth until the complete charge has been
added. This assures that the entire bed will be evently worked
The depth of each dose or the quantity of sewage applied
and the period between dosings vary with operating conditions,
i.e., sand size, bed depth, sewage strength, and other variables.
The time required for disappearance of a dose of sewage from
the beds will vary somewhat with the hydraulics of the sand
and the time the bed has been in service with a particular sew-
age. It will disappear very rapidly from the surface of a clean
bed and more slowly as the surface becomes clogged with re-
tained suspended matter. The bed should be completely drained
for several hours before the succeeding dose of sewage is applied
in order to re-oxygenate.
The upper /. to '/ inch of the bed surface, serving as a
strainer, will remove most of the sewage solids. The retained
solids will gradually form a mat which eventually reduces the
rate of sewage passage to the extent that the "resting period,"
or period of re-oxygenation, will be dangerously shortened be-
tween dosing cycles. Before the filter surface becomes imper-
vious, it is necessary to take the filter out of service for clean-
ing. This is best accomplished by allowing the mat. of organic
matter to thoroughly dry, harden and curl up to aid in its re-
moval without disturbing the cleaner sand underneath. In small
plants the hardened organic mat may be removed by raking
lightly with an ordinary iron garden rake, but taking care that
the prongs of the rake do not dig deeply enough into the sand
to form small furrows which tend to cause sub-surface clogging
with subsequent doses and necessitate the removal of more sand
with the next cleaning. This will insure that only a thin layer
of sand will be removed and discharged with each cleaning.
Filter surfaces usually require cleaning when sewage remains
pooled at the surface four to five hours after dosing.
The frequency with which filter beds require cleaning is a
function of the quantity of suspended matter in the filter in-
fluent. The greater the efficiency of the presettling units, the
greater will be the time interval between cleaning. Finely
divided silt washed from street surfaces in time of storm is very
likely to cause serious surface clogging.
In spite of careful surface cleaning the upper few inches
of sand will store up organic matter with time. Microorganisms
will slowly oxidize most of this material, but a certain amount
of unoxidizable, dark humus-like material will remain which
will eventually require removal. This is accomplished by scrap-
ing the dark-colored mass of sand and humus from the surface
until the lighter colored sand is reached.
After the depth of a filter has been significantly reduced
by repeated cleaning, it becomes necessary to build the bed
back to its former depth with new clean sand of the same size
and preferably from the same source as the original sand. Be-
fore resurfacing the bed, it is good practice to make certain
that all of the clogged previous surface has been removed. In
resurfacing, it is advisable to remove more sand than would be
done under ordinary cleaning operations to lessen the chances
of internal clogging.
When filter beds are overtaxed or overloaded for an appre-
ciable period of time, as indicated by a lack of stability of the
effluent, it often becomes necessary to allow varying intervals
of time to lapse before dosing operations are continued. The
rest period will allow the bed to dry out, overcome the effects
of capillarity and recharge the lower portions of the bed with
air. This restores the original aerobic condition and the bed
will gradually resume its normal efficiency. The rest period
may vary from one to several weeks depending upon the degree
to which the bed was overloaded.
The filters should at all times be kept free of grass or other
vegetation and care should be exercised to prevent dirt from
being washed onto the filter from the surrounding embankments.
This latter may be lessened by lining the banks with stone or
broken concrete, or by planting the entire bank with a suitable
Ponding or wet filter areas are never allowed in good operat-
ing practice. When portions of a bed remain wet continuously
or when there is any pooling of sewage on the bed, algae and
fungi will appear. They will clog the sand, prevent the beds
from drying, and greatly increase the difficulty and cost of elean-
ng. Keefer (12) recommends the use of chloride of lime to
control the fungi and copper sulfate to eliminate the algae.
He further recommends that where pools of sewage remain on
the surface of filters for several hours after being dosed, that
the units be taken out of service and allowed to dry. The surface
mat of sewage solids should then be removed.
The temperature effect on the operation of a sand Alter is an
important factor. Due to the fact that low temperatures inhibit
the microbiological activity which bring about purification within
the bed, the eciency of sewage liters in northern climates is
greatly reduced during the winter months. Winter operation
in cold climates requires considerable surface care in plowing
the beds into ridges and furrows to prevent the beds from freez-
ing too deeply and the application of large and irregular doses
of sewage to keep the bed thawed out. Under the mild year-
round climate in Florida, it is doubtful that winter operational
practices will ever become necessary in any part of the state.
Also, filter efficiencies should be relatively constant throughout
the year over the greater part of the state.
The controlling factors for successful operation of intermit-
tent sand filters are recognized as including: (1) The character-
istics of the sand in the filter, i.e., its homogeneity, size and
uniformity, cleanness, freeness from cementing materials, and
shape of sand grains; (2) depth of sand bed; (8) chaacteristic
of the sewage; (4) uniformity of distribution; (5) efficiency of
underdrains; (6) climatic conditions under which operations
will occur; and, (7) the care which the filter will receive during
DESIGN AND OPERATIONAL EXPERIENCES
There are, at present, insufficient data on the behavior of
intermittent sand liters under Florida conditions to make
specific design or operational recommendations. Prior to making
these recommendations, it is imperative that more definite in-
formation regarding the relationships of the variables involved,
especially the relationship between the organic load applied and
the physical characteristics of the filter be obtained under
Florida's year-round mild and humid climate. These factors
can be determined only after a careful study of many different
intermittent sand filters constructed from native sands. Until
these data are available, the general recommendations of re-
sponsible engineers appearing below may serve as a guide to
the judgment of the designing engineer.
Sand Characteristics: The selection of a natural sewage
filter plant site is based primarily upon the physical properties
of the sand found in situ. The characteristics of the sand em-
ployed in the construction of either natural or artificial sewage
filters are of prime importance since they affect the quantity
of sewage that may satisfactorily be treated by the particular
filter, the rate of filtration, the permanence of the filter, and the
frequency with which the bed must be cleaned. The properties
of a satisfactory sand, as suggested by various authors, are
individually considered in the following paragraphs.
Homogeneity: The sand, as found in situ or placed in arti-
ficial filters, should be free from strata or veins of material of
varying degrees of fineness (1, 17). Many internal clogging
difficulties experienced have been due to stratification of filter
material finer or coarser than the average. One of the prime
requisites for permanence of the filter is homogeneity.
Uniformity and Size: Hazen (8) was the first to develop a
method for grading filter sand according to size and uniformity.
His suggested measurements, the effective size and uniformity
coefficient, are still widely employed to describe the particle size
distribution of filter sand. They are determined from mechani-
cal sieve analyses of representative samples of sand using stand-
ard eight inch diameter sieves with known or calibrated mesh
openings. An accurately weighed sample of sand is passed
through a nest of six or more of these sieves, having progres-
sively smaller sieve openings, by means of a suitable mechanical
shaking machine. The quantity retained on each sieve is care-
fully weighed after fifteen minutes of shaking. The percent-
age of the total sample retained on each sieve is calculated, and
from these data, the percentage passing each sieve size is com-
puted. A sieve analysis curve is then plotted on logarithmic
probability paper using the logarithmic scale for the diameter
of the largest grains which will pass the sieve of the stated
size, i.e., the calibrated sieve size, and the probability scale for
the per cent passing the sieves. Typical sieve analysis curves
of three natural Florida sands are shown in Fig. 2. The 10 to
60 per cent size are then read from this curve. The effective size
is the size of the particle of sand than which 10 per cent by
weight is smaller. The uniformity coefficient is the ratio of the
effective size to the size of the particle than which 60 per cent
Uniformity Coefficient: The uniformity coefficient is an ex-
pression of the size range within which fifty per cent of the
sand particles lie. The limiting values for the uniformity co-
efficient have been placed by some authors at as low as 2.5 (3)
and 3.0 (20) through the intermediate value of 4.0 (14) and to
the high value of 5.0 (1, 12, 17, 18). The uniformity coefficient
is one measure of the porosity of the sand; the nearer this value
approaches unity the greater becomes the porosity and the more
desirable the sand.
Effective Size of the Sand: The size of the filter sand greatly
affects the quantity of sewage that may be treated by a particu-
lar filter, the quality of the effluent, the rate of filtration and
the tendency of the filter to become clogged. The sand should
be neither too coarse nor too fine. If the sand is too coarse it
permits too rapid passage through the bed, insufficient time
of contact with the active biological gel, and deep penetration of
fine solids which may lead to internal clogging. On the other
hand, fine sand limits the volume of sewage too greatly and de-
creases aeration by too long detention of the sewage and capil-
lary saturation of the sand. Boyce (8) and Metcalf and Eddy
(15) suggest that not more than one per cent of the sand be
finer than 0.13 mm.
There is considerable difference in opinion among various
authors in what constitutes the optimum effective size range.
The range most frequently suggested is 0.2 mm. to 0.5 mm,
(1, 10, 17, 20). Much narrower ranges are specified by some
authors (3, 7, 12, 14, 15, 18), however, all fall within the above
Cleanness: A siliceous sand should be specified generally
with a minimum of calcareous or argillaceous matter. The total
organic matter in the sand should be less than one per cent and
the total acid soluble matter should not exceed three per cent.
The sand should be free from clay, loam, soft limestone or other
material which may be disintegrated by the sewage liquid or
which may have a tendency to cement the particles of sand
(1, 17, 18). The American Society of Civil Engineers Commit-
tee (1) states that calcareous or argillaceous coatings on sand
grains have been found objectionable. The National Sand and
Gravel Association (16) is more lenient in its specifications con-
cerning acid soluble material. They state that not more than
five per cent shall go into solution in hydrochloric acid after
twenty-four hours digestion.
Shape of Sand Grains and Porosity of Filter: There is gen-
erally good agreement among engineers that the grains of sand
employed in the construction of sewage filters should be round
or oval as opposed to sharp and angular grains recommended
for water filtration (1, 17). The few engineers that recommend
spicular or angular shaped sand grains have obviously confused
intermittent sewage filtration with water filtration. Essentially
sewage filters are not filtration units in the true sense of the
term, but are absorption and oxidizing devices dependent upon
microbiological activity and distribution of air throughout the
bed to bring about purification.
The porosity of the filter, or the portion of the filter receiv-
ing air, is a function of the quality of the effluent. This factor
is determined by the size and shape of the sand grains, the
more spherical the grains the greater the per cent voids through-
out the bed and the more air incorporated within the filter be-
tween doses. The frictional resistance to flow is also lessened
by this factor.
It has been pointed out (1) that crushed flint or quartz
gravel should not be used in the construction of artificial filters
since the packing effect of such angular material may be too
great. This would reduce the porosity to that of a much finer
material and, therefore, correspondingly reduce the quantity
of sewage that may be treated. It may be restated that the
nearer the uniformity coefficient approaches unity, and the
more spherical the shape of the grains, the greater is the
porosity of the sand bed and the more desirable becomes the
sand for sewage filtration.
Depth of Sand Beds and Underdrainage: Experiments con-
ducted by the Lawrence Experiment Station (14) show there
is no advantage in having a depth of sand greater than three
feet. The one and two foot depths included in their work gave
a remarkable amount of purification when the depth of sand
is considered. Generally, sand depths of thirty to forty-two
inches have been specified (7, 10, 12, 15, 17, 18, 20). More than
half of the total purification accomplished by a sand filter takes
place within the top nine to twelve inches of the sand bed, the
purification becomes progressively less for each succeeding foot.
This is clearly brought out by the data in Tables II and IV.
In the design of sand filters, as with all sewage filters, the
principle factors governing the depth are available head and
topography of the site. Metcalf and Eddy (15) state that
"although very little added improvement is obtained by making
sand filters deep, it has been found advantageous practically to
have 3 or 4 feet of sand above the underdrains in order to pre-
vent sewage from breaking through and reaching the collectors
in an inadequately oxidized state." The greater depth tends
to produce more constant results and permits the removal of
more sand in the course of the scraping or cleaning operations
before it becomes necessary to rebuild the bed. Where fine sand
is used, additional allowances in depth must be made to offset
the capillary rise of water. Sand filters should never be made
so deep that they are not properly ventilated. "If anaerobic
conditions are established in the bottom layers the effluent is
deteriorated in quality and, in the absence of dissolved oxygen,
iron is taken into solution and growths of iron bacteria (Creno-
thrix polyspora) may clog the drains" (15).
The size, slope and spacing of underdrains are usually de-
termined by the sand size, depth and shape of the bed, and the
quantity of liquid to be handled by them. The natural sand beds
of the east are usually four to six feet deep and are underdrained
at intervals of from 20 to 100 feet. Generally, the greater the
depth and coarser the sand, the wider may be the spacing. The
artificial beds are, however, ordinarily not more than 30 inches
deep and are underdrained at much closer intervals. Spacing
of 10 feet between underdrains is common and 5-foot spacing
is not unusual (1).
Metcalf and Eddy (15), Imhoff and Fair (10) and Keefer (7)
state that underdrain spacing of 40 feet with moderate sand
and 30 feet with fine sand has been found satisfactory. The
Ohio Department of Health (17, 18) and the American Society
of Civil Engineers Committee on Filtering Materials (1) specify
that underdrain systems be laid in gravel at intervals of not
more than 15 feet. Boyce (3) recommends, however, that the
entire sand bed be underlaid with gravel and that the under-
drain spacing be reduced from 15 feet to 5 or 6 feet. It would
appear that a better balance between efficiency and economy
would lie closer to 15-foot spacings than to the extremes of
5- and 40-foot spacings.
Sand filter underdrains are usually laid on uniform slopes
with 6 inches in 100 feet as a minimum (15) and with open
joints approximately %-inch apart. Vitrified salt-glazed sewer
pipe, of 4 inches minimum diameter has been generally recom-
mended as satisfactory pipe (1, 7, 10, 15, 17). Cement pipe has
not proved durable due to attack by acids formed in the
Underdrains (10) are laid in trenches surrounded by a 3-inch
layer of coarse stone (1 to 21/ inches in size) followed by a
second 3-inch layer of finer gravel (1/ to 1 inch) and a third
layer of coarse sand (less than /4 inch in size) in contact with
the filter sand proper.
In order to insure the proper aeration of the bed after dosing,
underdrains are designed to carry away the sewage at the rate of
percolation (10). This may be gaged approximately from
Hazen's formulation of the flow of water through sand. The fol-
lowing figures show the general magnitude of the flow involved:
Effective Size Maximum Rate of Percolation
mm. M.G.M. per Acre
0.2 20- 50 30- 60
0.3 50-100 70-150
0.4 100-200 120-250
0.5 150-300 200-400
As a rule, the lower rates are more common than the higher
For the construction of artificial filters, more exacting re-
lationship between sand size and depth, and underdrain spacing
is badly needed. In spite of the fact that the majority of the
purification occurs in the upper few inches of the bed, a bal-
ance must be struck between the depth of bed and underdrain
spacing to insure both uniform flow and economy of installa-
tion. Where the filters are too shallow or the underdrains
spaced too far apart, there is a tendency to overload those por-
tions immediately adjacent to the underdrain and greatly reduce
the effectiveness of the bed.
Rate of Distribution on Filter Bed: Imhoff and Fair (10),
Metcalf and Eddy (15) and Barbour (2) recommend that suffi-
cient sewage be run onto the bed to cover it to a depth of 1 to 4
inches. The higher value is preferable in order to load large
beds uniformly. The dosage should be regulated so that this
depth of flooding is obtained in from 7 to 20 minutes, or at a
rate of application of 0.2 c.f.s. per 1,000 sq. ft. of bed area.
Dosing tanks should store the full dose of a bed, and siphons,
220 A; 117.8
180 | 96.4
160 d 85.7
140 & 75.0
120 o 64.2
Limit for Raw Sewae
SLiAit ~fe at Se ttA s* ge
I Fs. ,mw.
10.7 i___1 i I I I I I I I I I -"
20 60 100 140 180 220 260 800 340 380
Average Net Dosing Rate, Thousands of Gals./Acre/Day.
Fig. 1.-Suggested Safe Loads for Intermittent Sand Filters.
(Reprinted by permission from Am. Soc. C. E. Manual No. 13).
or other regulating devices, should be dimensioned so as to dis-
charge, under minimum head at about twice the maximum ex-
pected rate of inflow.
Sewage Loading: The quantity of sewage that may be satis-
factorily applied to an intermittent sand filter depends upon
the physical properties of the sand bed, degree of pretreatment
and strength (in terms of B.O.D.) of sewage, and the desired
quality of the effluent. There are, apparently, no definite data
available in the literature by which optimum sewage doses may
be quantitatively predetermined. More exacting data are needed
giving the relationship between the organic load applied and
the physical characteristics of the sand in order to rate any
particular filter for any particular set of conditions. The data
given below will serve to show trends, approximate doses and
approximate qualities of the effluents under given conditions.
The conclusions of G. A. Johnson (11) from the Columbus
(Ohio) experiments show a relationship between safe loading
and quantity of suspended matter in the filter influent. He con-
ducted a series of tests on 21 sand filters 0.001 acre in area
with 3-foot depths of Lake Erie sand of effective size 0.20 to
0.29 mm. using sewage which had received various preliminary
treatment. The results of these tests show that the approxi-
mate rates of treatment (see Table I) may be maintained with
the indicated quantity of suspended matter in the filter influent.
These conclusions were later confirmed by Clark and Gage
of the Lawrence Experiment Station who reported (4) that
"Sand filter rates can be increased only in proportion to the
amount of organic matter removed by sedimentation, .."
TABLE I.-Dosing Rates of Sewage Containing Varying Quantities
of Suspended Matter.
Suspended Matter in Influent, ppm. Dosing Rate in Gal.-Acre-Day
As a guide to judgment on the part of the designing engi-
neer, the A. S. C. E. Committee (1) has suggested the loadings
shown in Fig. I. These loadings were based on the study of
available data and are considered conservative.
A series of tests were carried out by Grady (6) at North
Carolina with small 8-inch diameter sand filters and using septic
tank presettled sewage. The results of these tests, given in
Table II, show per cent reduction in 5-day B.O.D. for various
sand sizes and depths, at loadings of 125,000 and 200,000 gals.
per acre per day. These results indicate that much higher doses
may be successfully applied than those recommended by the
A. S. C. E. Committee on Filtering Materials.
TABLE II.-Per Cent Reduction in 5-Day B.O.D. with Various
Sand Size and Depth.
Depth Effective Size mm. Effective Size mm.
Inches 0.20 0.30 0.33 0.50 0.20 0.30 0.33 0.5
125,000 Gals. per Acre per Day I 200,000 Gals. perAcre per Day
9 95.8 90.4 85.9 73.5 91.8 85.5 76.2 67.5
12 97.9 91.5 88.8 74.1 95.8 87.8 77.2 71.0
18 97.2 97.2 89.8 82.2 96.4 89.0 84.4 70.4
24 99.0 98.1 95.5 86.0 98.2 94.0 90.0 74.8
30 98.5 98.8 96.7 82.8 98.0 96.1 92.6 80.0
PROPERTIES OF NATURAL FLORIDA SANDS
The large quantities of sand required for sewage filters, and
the prohibitive cost of transporting the material long distances,
led to an investigation of the natural sand deposits over the
state in an effort to locate sufficient quantities of suitable sand
within a short distance of any community desiring to install
intermittent sand filters. While many communities and institu-
tions will find sufficient sand of good quality in situ to construct
natural filters, others may be compelled to build artificial filters.
In the former case, the depth of the filter, within reason, is
not important economically. Where it is necessary, however,
to haul all or a part of the sand, or to handle the material other
than to distribute it evenly over the bed, the cost of installation
becomes increasingly great with each additional inch depth. It
would appear advisable then to build artificial filters only to the
depth actually required to treat the sewage to the desired degree.
As explained above, there is an economic balance between under-
drain distance and sand depths for each particular sand size for
Representative samples of natural sand deposits over the
state were collected and classified according to their effective
size, uniformity coefficient, porosity, and solubility in acid. The
results on some of these sands, along with their locations, are
given in Table III. All sands collected were light in color and
rounded to sub-angular in shape. The loss on ignition was less
than one per cent. This list is by no means complete, but may
indicate to sanitary engineers the type of sand they have avail-
able in or near their community. Additions to this list will
become available as more samples are obtained.
TABLE lll.-Physical Properties of Some Florida Sands.
Nearest Town and County I S 3 ;
1 Lakeland, Polk Co. ................ 0.18 1.98 45.0 3.00
1B Lakeland, Polk Co. .................. 0.22 2.0 38.0 1.46
2 Hawthorne, Alachua Co. .......... 0.16 2.1 45.8 0.30
2B Hawthorne, Alachua Co. ........ 0.17 2.1 46.2 0.35
3 Pembroke, Polk Co ............... 0.32 2.2 44.3 0.12
3B Pembroke, Polk Co. ................ 0.30 2.1 42.5 0.12
4 Pembroke, Polk Co. ............. 0.26 2.2 43.5 0.12
5 Lake Wales, Polk Co. ............... 0.22 2.5 42.8 0.07
6 Interlachen, Putnam Co. ....... 0.24 3.0 42.2 0.20
7 Interlachen, Putnam Co. .......... 0.2 2.6 43.4 0.16
7B Interlachen, Putnam Co. ......... 0.23 2.7 44.6 0.16
8 Okahumpka, Lake Co .............. 0.24 2.3 42.2 0.07
8B Okahumpka, Lake Co. ............ 0.26 2.4 39.6 0.07
9 Edgar, Putnam Co. .................. 0.29 2.0 44.3 0.23
11C Chattahoochee River,
Gadsden Co. ......................... 0.44 3.2 44.6 0.09
12 Mulberry Polk Co. .................... 0.21 2.0 45.6 0.18
13 Daytona Beach, Volus:a Co..... 0.13 1.4 46.5 1.38
14 Lake Placid, Highlands Co....... 0.17 2.1 45.9 0.03
17 St. Petersburg, Pinellas Co..... 0.14 2.3 43.0 0.23
17B Manatee River, Manatee Co..... 0.23 1.8 43.4 0.10
18 Clermont, Lake Co. ................ 0.27 1.7 46.2 0.81
19 Jupiter, Palm Beach Co............. 0.32 1.43 42.2 0.27
20 Opalocka, Dade Co. .............. 0.17 2.4 35.7 0.32
23 Biscayne Bay, Dade Co. ....... 0.18 1.9 42.4 1.52
24 Tavares, Lake Co. .................. 0.17 3.5 38.4 0.10
25 Milton, Santa Rosa Co. ............ 0.3 1.3 47.0 0.17
26 Barrineau Park, Escambia Co. 0.16 1.9 47.3 0.05
The effective size of the sand was determined by sieve
analysis using U. S. Standard Series Sieves and the method
recommended by Hazon (8). The uniformity coefficient was
SA --- ---- - -- - ^ ^ ^ -----I' -- I I------ ---
to I- --- - --I I I I- --
&0 Fig. 2.-Sieve Analysis Curves of Three Natural Florida
40 Sands Employed in the Laboratory Sand Filters.
40- -------------- -
4:-- --- --- -- -- -___ _-- -- -- ---- -
/ / "0036 m.
^ __ / - i - - -- -
^ ^ ^'oi
i I _I I I I _I I I L I I 1 1- -6
10 20 Ce 40 assi See
Per Cent Passing Sieve
go r of of fts
ealculted from sieve anayis curves by definition. Poroity
was determined by the method of Hulbert & Feben (9). The
per cent soluble was determined by the loss of weight of an
accurately weighed sample upon digestion for 24 hours in 6
N HCL All determinations were made in triplicate.
Three samples of natural Florida sand, No.'s 2B, 4 and UC,
of effective size 0.17 mm., 0.26 mm., and 0.44 mm., respectively,
were selected for laboratory tests in small intermittent sand
filters. Sieve analysis curves for these sands are shown in
Fig. 2. It is recognized that dependable data on this process
are not obtainable in the laboratory, but can only be derived
through pilot plant operation conducted under actual field con-
ditions. The results of these experiments, however, serve to
indicate trends in the treatment effected under the specified
conditions. The per cent removal of B.O.D. using different
depths of the above sand, 12", 18", 24", and 80", and under
various loadings, 73,000, 180,000, 186,000, and 225,000 gallons
per acre per day, are presented in Table IV. The temperature
was controlled at 850F. 8F. The filters were dosed daily with
University screened sewage which was collected at approxi-
mately the same time each day and presettled for two hours in
5-gallon bottles before use. During the course of these experi-
ments, the sewage strength varied considerably due to fluctu-
ations in campus population.
Laboratory filters were constructed in the following manner.
Various lengths, depending upon the depth of filters, of 100 mm.
diameter Pyrex glass tubing were sealed with a mixture of
beeswax and paraffin into underdrains composed of 110 mm.
diameter Buechner funnels containing two circles of 80-mesh
stainless steel filter cloth cut to fit the bottom of the funnels.:
The filters were secured in a vertical position in groups of four
on heavy metal stands and placed in a dark, thermostatically
controlled room (See Fig. 8). Rubber tubing extended from
the funnel outlet to the bottom of a one liter cylinder which
served as the effluent collector. The filters were placed in a true
vertical position with the aid of a plumb bob
Fach set of filters were charged with a different depth of
one carefully mixed sand sample. After the addition of sand,
care was taken that the filters were not jarred. No graded
gravel underdrain was used in these experiments.
Fig. 3.-A Group of Four Laboratory Sand Filters Showing Four Depths
of the Same Sand.
Dosing of the filters was carried out at ap-pro~ motel the
same time each day. The required volume of sewage was
siphoned from the upper three-fourth of a five gallon bottle,
after settling for two hours, into a graduated dispensing bottle.
Suspended solids in the settled sewage average 88 ppm. The
sewage applied to the filters through a tube extending from
the bottom of the dispensing bottle. The tube nozzle was kept
below the surface of liquid on the filter and slowly moved in a
circular direction around the perimeter to avoid aerating and
disturbing the filter surface.
Sand 2B and 11C were operated simultaneously for 5 months,
after which sand 2B became clogged and was discarded. The
filters containing sand 2B were then cleaned and refilled with
sand 4. Filters containing sand 11C were retained and operated
parallel to sand 4. These two groups of filters are still in oper-
ation, but the data presented in Table IV, for these two filters
in parallel, are averaged from a total of ten weeks operation.
At the date of this writing, sand 11C has shown no evidence of
clogging after 11 months, and sand 4 has not clogged in 6 months
of operation. For the last three months, both sets of filters
have been dosed at the rates of 186,000 g/a/d with 2-hour
settled sewage. Sand 2B gave good reduction in B.O.D. 4
weeks after being placed in operation. Sand 11C required about
8 weeks, and sand 4 about 8 weeks to give good reduction of
In general, from the foregoing data, there appears to be
no significant increase in efficiency beyond 18" depths for the
fine and medium sands. Daily results show, however, that with
the deeper depths the performance of the filters are more uni-
form. The efficiency of the coarser materials increases more
significantly with depth than the finer material. The fine sand,
2B, gave consistently good results at all loadings up to the time
it became overloaded. Sand 11C appears to be the best of the
three samples at all depths and loadings.
Plans are being made to continue these studies on pilot plant
scale. Eight or more sand filters will be constructed, each
7.4' x 7.4', or 1/800 acre in area, to operate under the same or
varying conditions for comparison. In addition to studies of
the relationship between sand size, depth, and optimum organic
load applied, variations in methods of application and under-
drainage are also contemplated.
2B 4 11C
0.17 0.26 0.44
96 114 114
86.8 85.2 86.3
92.0 90.2 87.6
93.9 90.8 95.7
.9 I 92R 8.7..3
TABLE IV.-Per Cent Reduction in 5-Day B.O.D.
Dosing Rate, g/a/d
130,000 186,000 225,000
2B 11C 4 11C 2B 11C 4 11C 11C
Effective Size, mm.
0.17 0.44 0.26 0.44 0.17 0.44 0.26 0.44 0.26 0.44
S Average B.O.D. of Sewage Applied
89 89 126 126 192 92 1 108 1 108 262 262
Per Cent Reduction in 5-Day B.O.D.
86.4 78.6 77.5 88.5 88.1 73.4 80.7 82.6 76.9 77.6
90.9 79.3 82.0 80.0 88.0 I 81.4 80.7 84.2 83.0 83.0
92.4 82.6 77.0 88.0 89.6 82.1 76.5 90.5 78.0 82.0
93.0 85.9 70.3 93.3 93.8 83.8 i 76.6 97.8 77.2 78.0
. I .9 ..
MODEL WOODEN LATH TRICKLING FILTER
One of the research projects in progress at the Experiment Station is
a study of the Wood Trickling Filter designed to meet the needs of small
communities and institutions.
The author desires to express his appreciation to Dr. R. A.
Morgen, Assistant Director, and to Professor Earle B. Phelps,
Research Engineer, Engineering and Industrial Experiment
Station, for their helpful suggestions and criticisms. Special
credit is due Mr. Julio Noltenius Velasquez for the many Bio-
chemical Oxygen Demand determinations made in connection
with this study.
1. American Society of Civil Engineers, "Filtering Materials for Sewage
Treatment Plants," Manual of Engineering Practice, No. 13 (1935).
2. BARBOUR, F. A., "Intermittent Sand Filter," Jour. Assoc. Eng. Socs.,
47, 59 (1911).
3. BOYCE. ERNEST, "Intermittent Sand Filters for Sewage," Municipal
& County Eng., 72, 177 (1927).
4. CLARK, H. W., and GAGE, S. DEM., "A Review of Twenty-One Years'
Experiments upon the Purification of Sewage at the Lawrence
Experiment Station," 14th Annual Report of the State Board of
Health of Massachusetts, (1908).
5. FRANKLAND, SIR EDWARD, "Rivers Pollution Commission of Great
Britain, First Report," (1870).
6. GRADY, R. H., "The Treatment of Sewage Using Intermittent Sand
Filters," North Carolina Water and Sewage Works Assn., 17.
7. HARDENBERCH, W. A., "Sewage & Sewage Treatment," Scranton, Pa..
International Textbook Co., (1936).
8. HAZEN, ALLAN, "Some Physical Properties of Sands & Gravels,"
Report of the Massachusetts State Board of Health, 550 (1892).
9. HULBERT, R., and FEBEN, D., "An Improved Method for Measuring
Porosity of Sand," Jour. Am. W. W. Assoc., 26, 271 (1934).
10. IMHOFF, KARL, and FAIR. G. M., "Sewage Treatment," New York,
John Wiley & Sons, (1940).
11. JOHNSON, G. A., "Report on Sewage Purification at Columbus, Ohio,"
12. KEEFER, C. E., "Sewage-Treatment Works," New York, McGraw-Hill
Book Co., (1940).
18. Massachusetts State Board of Health, Annual Report, 1892.
14. Massachusetts State Board of Health, Annual Report, 1986.
15. METCALF, L., and EDDY, H. P., "American Sewage Practice," Vol. m,
New York, McGraw-Hill Book Co., (1985).
16. National Sand & Gravel Association, "Specifications for Filter Sand,"
The Canadian Eng., 57, 376 (1929).
17. Ohio Conference on Sewage Treatment, Report of Research Committee,
"Sewage Filters," 12th Annual Report, 97-185 (1988).
18. Ibid., Recommendation of Ohio Department of Health.
19. PHELPs, E. B., "Stream Sanitation," New York, John Wiley & Sons,
20. Texas State Department of Health, "Suggestions Regarding Prepara-
tion, Submission, and Approval of Plans for Sewerage Systems,"
Views of the Engineering and Industrial Experiment
Station Sanitary Laboratory.