-------
CO
tj
M".
O
S
E-
fjj
y
i
CO
Ep
2
fr,
0
§
1
^g
8
c-i
3
C-i
w
* 3
o
H * .
fl*
i
ee
ffl
Oj
°
= 5
•M U
ot» tn
i o-
§
u
> "c
a, -H
CO —
u £ —
H U *
O 01 —
W »H
i s-8
QUO
3C >« 3 V.
O. OQ — (
O OC O> 3
CO 5 * «
o* "e
< 1
H *-*
a. Z
Ul
VI
Ci) X U
§3 a
1 s a i
5 3 15 -
|
O
3
* m (N CM (N
r* * r^ m r*
*n iH «H r-»
.*-! (N 00 04 O
»H f* «n tn «o
o o o o o
«"» *e o n* o
flO « (S ^ 00
o o o o o
in *1 ™ " ^
r^ w o c** n
A CM CM r-l
m o% 9 -^ in
o CD «r ^ ^
r-4 x 2 O
- i
•o
r-t
• "O "O Tl T3
O
r* i
i " « n n
«•
1 232
19 b W W
«H >, U O O
W ki
J< « Q. «
30 fl) — ^ j-| J< X At
ft C O W 4-> C C C
3 o u o uitiqtg
-•^ c 2. -HHHH
•••* O £
si r i xiii
20
-------
CO
N
CO
2
s
t ^
w
s
M
u
w
CO
£3
&
Q
S.
^*lj
S
D
CO
en
W
<
DC
1
Q
1
Ed
S
co
<
03
w
§
&4
£
S
FOULED
.S «
llf
s -.
W*"^
CO®
&«
w
Q^ 1O
Q CO **
<9
x~v
•*— ' o
^ CO
3" 10
O
O
O
ii
1
OJ ** CM CO CO tD
£
»i • -J *J ' ^H
^ * *4 -H ^ ^
1 I
l-( 1— t T- t O ^H
- -J J J ei S-
S S $
§ § ^i ». rt 0
t* 1O CO IQ r* v
d d d d d d
P §
2 ^
S s
S i > i i i ^i
^SSp « 35 ttfi
co2"2 S S S§
co S
° S
1 §
S o
oSrai f= n S!S
w§2§2» | g
< z
- 8
g § ^ g
*-« T!« o o S co
d ' d ' d d d d
--••----
iq oo
§ t^ < S 55 w
« M » «-< A i-t
!i
• £ •'S i°
>. 9 o » 2 £
, « 1 H H *„
1 I 1 If 11 !l
g tS g gc? g^ ^^
*H «— * «M *^ i^
f^ ^H ^ ^1 S
JO IO jj JO JO
!
S- S -r S- S-
i ;
» < d i i •
A O) OS CN IO
N C* LO U3 LQ
do odd
c3 u5 t^ IS c^i
23 £J if 2i
CO Ci CO ^* CO
I
o> od £J 2 «
1 ''•
S «-« CO CO CO'
o d odd
s
CO ;
eo *H o c* 01
CO <-< CN C4 <-<
N W ^ CW CO
T ? 11 T T
o o to o o
I: •
11 !
-------
?
3
C
•a
a
1
reRisncs i
o
oj
-j
|
Q
S
JTj
I
1
en
W
W DIFFUSER
CEIVED/NE
K)
K
co
Ed
fe
D
1
1
-£8
41 "C '
i
i;g
e
Ed
S -»
®
§1
•1.0
a®"3
>
o
o
>
OS'S
sac.
§
I
2J •* C4 TO P3 (0 (Q (p
• ... *"
ss.ft SB » SS-
22222 32S
A IO
§ S ^ s t~
^^«-J*H »H C*4' "Y S" *>
*
? s $
odd d d odd
I I
1" %
&> S
3 < < < < -i So>.
Jo; MO n o> atS«S
gSffS g S «SSS
S2 ^
3 I
U o
i 2A ss'n
fr* « «
|l ? 1
1 £ "5 §0 3 (S
, * § II |l *' ' '
2 a I 3^ «« 501
1 1 1 II . 14 II
v to to
•a !
^ *H ^H
1
S- S- -4
i
[
s
d d o
^ ^ is;
C4 *« O5,
n •* n
i
« 5s p
S '
d [
00 S S
n n m .
odd
00
I n ro'
§' a 2,
!
a" N •» '
Whitter Narrow
Tank 3 - Grid
• Grid
• Grid
22
-------
The same tests were carried out for the diffusers from the seventh site, Whittier
Narrows, excepting that diffusers were taken from two tanks. The diffusers had passed through
different cleaning regimes. The locations from which diffusers were sampled are shown in
Figure 14. Tank 3 had previously been cleaned by high pressure hosing only. The diffusers
had been in service for 15 months prior to the time of sampling. Tank 2 had earlier been
cleaned by the Milwaukee method; high pressure hosing, acid wash and high pressure hosing.
It, too, had been in service for 15 months prior to the time of sampling. Diffusers were
collected from each of the three grids in each tank to observe any change in SOTE along the
length of each tank and any corresponding change in biofilm properties. i
Results
Data for SOTE, BRV, DWP, BRV/DWP, EFR and foulant characteristics for all
test sites are reported in Table 3. The diffusers from Houghton were broken in transit limiting
analysis to BRV only. The diffusers from Monroe were not cleaned although data for a new
diffuser are included for comparative purposes. Data for plants with biofilm thicknesses greater
than 3 mm, Milwaukee, Jones Island, Monroe and Whittier are also recorded in Table 2 to
provide a direct comparison with data on biofilm properties.
Discussion
The diffusers from Monroe, Green Bay, Madison, Milwaukee,; Jones Island and
Milwaukee, South Shore, were all Sanitaire ceramic discs. Except for the ' Milwaukee, Jones
Island diffusers, they demonstrated an SOTE of 18 to 19% at 1 cfm/diffuser in the fouled state
reducing to a value of 15 to 16% at 3 cfm/diffuser (Table 3). The Jones Island diffuser was
heavily fouled registering an SOTE of 10.5 and 8.3% respectively. The Jones Island diffuser
also registered a very high BRV, COV and EFR and a low value of DWP/BRV. However, DWP
at 1 cfm was at a normal value indicating that the biofilm had not significantly affected the
diffuser head loss despite significantly affecting uniformity of bubble formation and SOTE.
Upon cleaning, all the diffusers were restored to almost new conditions with SOTE, DWP and
BRV values close to that in clean water.
The diffusers from Whittier Narrows were all Norton ceramic domes. They
demonstrated an SOTE of 14 to 16% at 1 cfm/diffuser in the fouled state, reducing to a value of
10 to 13% at 2.7 cfm/diffuser (Table 3). All the diffusers had a thick growth ;of biofilm (greater
than 3 mm). The tank 3 diffusers, which had historically been cleaned by hosing with water,
demonstrated a head loss effect that was greater than in tank 2 perhaps indicating that the
acid cleaning procedures followed in tank 2 had delayed a head loss increase as compared to
tank 3. ,
Upon cleaning the Whittier Narrows fouled diffusers in the laboratory, SOTEs
were not restored to clean water values. They remained the same or were slightly reduced.
BE,V and DWP were reduced to almost clean water values. This difference in behaviour (as
compared to diffusers from the other sites) may be attributed to the fact that the Whittier
Narrows diffusers were cleaned in the laboratory by hosing only, unlike the other diffusers that
were cleaned in the laboratory by the Milwaukee method.
When comparing biofilm properties with diffuser process characteristics (Table
2), the lowest SOTE and the highest BRV measured were those at Milwaukee, Jones Island
which also reported the thickest biofilm. On the other hand, the DWP at Jones Island was
relatively low indicating that the biofilm was interfering sufficiently with bubble formation to
depress SOTE but did not affect the head loss across the diffuser as measured by DWP.
Whittier Narrows also reported thick biofilms but relatively high SOTEs were measured for both
23
-------
3c
3b
3a
2
1
<
Inlet
Control Tank
(Tank 3)
3
2c
2b
2a
1
nlet
Acid-Cleaned Tank
(Tank 2 )
Figure 14. Sampling locations at the Whittier Narrows plant.
24
-------
tanks 2 and 3. However, BRVs and DWPs for tank 3 were higher than observed at the other
plants. The biofilm at Whittier Narrows appeared to have only minimal impact on SOTE but
some effect on DWP and BRV was observed. Although the biofilm thickness did not change
along the tank length, its carbohydrate content declined slightly. In contrast, SOTE increased
slightly but measures of head loss and uniformity, BRV and DWP deteriorated slightly. The
different cleaning histories of tanks 2 and 3 appeared to have had no effect on biofilm thickness
and SOTE. Higher carbohydrate values were observed for the acid-cleaned tank (2) versus the
tank hosed with water (3). Lower values of BRV and DWP were measured for diffusers from
the acid-cleaned tank (2) as compared to the tank hosed with water (3). There appeared to be
no relationship between ABR and the diffuser process characteristics nor between volatility and
biofilm thickness or carbohydrate value.
In the case of Monroe, biofilm properties were measured on the thickest portion
of biofilm that occurred on the diffuser periphery. However, measurements of diffuser process
characteristics were measured in the centre of the diffuser as required and show that the
difiEusers were operating well at high SOTE and low DWP and BRV. Unfortunately, the biofilm
in the centre of the diffuser was not as thick as that at the periphery. Rather, it was very thin
and as a result, did not impair SOTE nor affect DWP and BRV. A comparison of biofilm
properties and diffuser process characteristics for the Monroe diffusers is probably not
appropriate in this case in view of the differences in biofilm thickness at the centre and the
periphery of the diffusers.
Observations of biofilms on operational fine pore diffusers| have shown that
biofilm formation is sometimes relatively uniform (Figure 15) but sometimes markedly irregular
(Figure 16), even at adjacent sites on the same four-lunger. It is possible that localized
differences in shear forces and periodic sloughing of parts of the biofilm produce variations in
biofilm thickness and distribution. This behaviour will require careful sampling in subsequent
investigations. Comparisbn of diffusers that appeared to be equivalent showed that only limited
trends appeared to exist between biofilm characteristics and diffuser process effects. Biofilm
thickness and carbohydrate content are two quantitative estimates of biofilm. accretion and they
can be easily measured. There appeared to be two effects that the biofilm 'exerted on diffuser
process characteristics. The biofilm can reduce SOTE but have minimal effect on DWP
(Milwaukee, Jones Island) and it can have minimal impact on SOTE but increase DWP
(Whittier Narrows).
CONCLUSIONS - PRELIMINARY FIELD STUDIES
p.»n»»m.i.. in .»• nM- !-•• i Mi n -• n i PIWi . . - •TiTiTghTB im • i* n .^•^iCTaaTT^r i» -T.n»i.Tf - i f FM, j
Microbial fouling of fine pore diffusers occurred irregularly at the seven wastewater
treatment plants examined. The thickness of a biofilm and its distribution over a
diffuser appeared to vary randomly.
Biofilms appeared to reduce oxygen transfer efficiency but have minimal effect on
dynamic wet pressure for diffusers at some locations. However at other locations,
biofilms had minimal impact on oxygen transfer efficiency but appeared to increase
dynamic wet pressure.
Scanning electron microscopic examination of biofilms with thicknesses greater than 1
mm showed that the biofilms were largely composed of linear bacterial cells enmeshed in
their own exopolysaccharide matrices. They did not appear to be intimately connected to
all areas of a diffuser surface and occasionally large spaces occurred between the biofilm
and diffuser surface allowing the accumulation of air bubbles in pockets. These biofilms
were traversed by structured air passages that terminated in large (approximately 1.5
mm) round apertures at the biofilm surface.
i
10
25
-------
Figure 15. Diffusers taken from a four-lunger after 15 months' service at the Houghtonplant showing
relatively homogenous accretion of biofilm material.
26
-------
Figure 16. Diffusers taken from an adjacent position in the same four-lunger at Houghtpn (Figure 15)
showing the irregular accretion of a microbial biofilm. \
27
-------
The highly structured biofilms are thought to interfere with the passage of an air bubble
such that the size of the bubble released from a fouled diffuser surface is probably not
the same as that from a clean and unfouled diffuser surface. i
The phenol-sulphuric acid procedure for measuring carbohydrate content provided a
useful measure of biofilm mass.
The trapped air pocket and air bubble release procedures provided a measure of the air
holding capacity of biofilms with thicknesses greater than 3 mm.
The enumeration of living bacteria in the biofilms was not possible by conventional
methods due to the leathery nature of the biofilms.
FIELD STUDY - VERIFICATION
Pitxzedures
On June 13, 1988, tank #1 at the Monroe, Wisconsin wastewater treatment
plant was drained for cleaning (Figure 17). Tank 1 contained Sanitaire ceramic disc diffusers.
Before cleaning, four representative diffusers were taken from each of four grids at locations
shown in Figure 18. They were identical except that the diffusers at location 1 had a specific
permeability of 50 units while the remainder had a specific permeability of 26 units. Two of
these diffusers were retained for microbiological analysis. The diffusers were photographed
(Figure 19) to record visual observations of the evenness of their biofilm accretions (Figure 20).
The biofilm thickness of one of each of the diffusers was measured at 7 randomly chosen points
on the diffuser surface using a simple penetrometer. The biofilm of each diffuser was scraped
from two measured 4 cm" sections and transported to Calgary, one for analysis by the phenol-
H2SO< method and the second sample was prepared for SEM by the methods outlined in
Appendix A. Four other diffusers were transported to Milwaukee where DWP, BRV, EFR, and
SOTE were determined "as received" and after cleaning by 4 methods:
(i) Hosing at 65 psi.
(ii) Hosing, bleach treatment and hosing.
(iii) Hosing, acid treatment and hosing.
(iv) Hosing, acid treatment, brushing, and hosing.
Results
x
SEM photographs are presented in Figures 21 and 22. Biofilm thickness and
carbohydrate content values are given in Table 4. Data for diffuser process characteristics are
recorded in Table 5.
Discussion . :
i
The biofilm thickness and the biofilm carbohydrate content were more variable
on the diffusers at location 1 (Table 4) and visual observations indicated heterogeneity between
individual diffusers at this location. Bacterial biofilms were consistently thick at location 2
(Figure 15) and these values decreased slightly and stabilized as the process fluid moved past
locations 3 and 4, towards the outlet. Apart from the heterogeneity seen in location 1, near the
inlet, the biofilm thickness and biofilm carbohydrate values were similar on pairs of diffusers
from each location. On all diffusers sampled, there was good agreement between the measured
biofilm thickness and the carbohydrate content per cm" of surface area (Table 4).
28
-------
Figure 17. Drained aeration tank at the Monroe plant.
29
-------
IANK
IANK 2
TANK1
25
S.PERM
GRID
3.2
-<-
25
S.PERM
-<-
GRID
3.1
*
I
25
S.PERM
GRID
3.3
25
S.PERM.
GRID
3.4
i r
I
I
25
S.PERM.
GRID
2.3
25
S. PERM.
GRID
2.4
1 1
\
25
S.PERM
-**-
GRID
2.2
-<-
40
S.PERM.
^«-
GRID
2.1
f
i
1.
\_^
i
25
S.PERM.
GRID
1.2
••
location 2
-*-
80
S.PERM.
location 1
• •
-»*•
GRID
1.1
I
25
S.PERM.
GRID
13
• •
locations
I
25
S.PERM.
GRID
1.4
location 4
• •
i r
'
A
TOCLARIFIER
•^
INFLUENT
i, , eg
Figure 18. Sampling locations for diffusers at the Monroe plant.
30
-------
Figure 19. Detail of the fouled diffusers in drained aeration tank at the Monroe plant
31
-------
PFV-. :"y,--~'' -•?»•' "• TB*V '
Figure 20. Distribution of biofilms on diffusers taken from the Monroe plant
32
-------
Figure 21. SEM of the surface of the biofilm on a Sanitaire ceramic disc diffuser with detail of a large
aperture. Bar indicates 0.5 mm.
33
-------
Figure 22. SEM of the surface of the biofilm on a Sanitaire ceramic disc diffuser with detail of a large
aperture. Bar indicates 0.5 mm.
34
-------
a 01
u 3
o
"
H U
O a>
tfl r-l
£; S
I
si
t-« O
b. ca
O OC
t-H <
CO U
I
cd tti L)
U O M
5 3 g
5 ,
3
35
-------
co
g
L^
CO
2
2
>5
W
^/
^
CO
D
pL|
Q
W
0
&
*z
i
s
is
^<
1
CO
TABLE 5.
«
co
D
fc.
3
a
§
c
>
6
a
CO
<
CS
1
tl.
5
a
D
g
'S.8 .
P*3
J
fs"**
g
EO
IsN
®
II
0<0
> CO
Q9
•Sia
~^t>
|SI
I
> __
ca-o
1
-*-^-*-* -ti''*^'^ •^•-^ ^f
WINCNIN C4C4NN CX« c5
"J 1 >o cqoso; c- 1-1
S22' S^2' Jui' *°
SSn. S?35. § S
^^»~*^ »-*I^^H 1-4 ' *
SSS. §gg. ' g §
odd odd d' d
Ci-*t*D ^t0^1"! to ^j-
^od^' crsi^oi' ^3' n
co <* w 01^01 (o ci
coioe-' t^oic^' t-^1 c^
CNtDC^ MOO— •* O
CJ^iM WCN— ^, ^i
ooo' odd' d d
«--cqi> -^--jo c« o
^^Oiso* tOoOrt* ^*' t~-
a*! "1 "I "1
I6* 1** i6* I6*
fcg a.8 ft-a sg
"•-" ""ej "".-i "-"N
*? =? •S? -2?
£3 £5 5S ^i
36
-------
•a
o
1
1
1
33
U
W
C/5
fe
Q
W
o
o
^
tt<
o
1
r~^
00
in
S
p
I
5
g
llg
&
3 ,-s
*o S
III
0
1
Ed
-I
O
p
1
Ess
Q
i"~
© ^~"
g
Cd
g- 10
|«S|
9
^
'"^o
& « ,£
Q® °
S t~ _
o "3° "§
o
o
«"£•
cao
bo £
.5
CB o
s!
c
o
•s
00 — . OS C- « O
S 2 22 d 2
•* t^ t~ to fl (,
S i§ S3 S • S
W OS CO M Tf t^
3 JN 2 " 2 **
aqn^fco t-;t-io^-i •*« oi«
2222 22°2 2°' do
^^SS SS^S g|2 S§
*~* •-* w —i i-i
S5S2 SSSct gS 2g
*H ^ ^i ^ ^^H^CJ cJr4 -^d
Cicqc;-^ ^o co c> co •'tco cs1*
t-lOcOC*- OoicOOS ™ *^* ^ M* K Q Q Q
EEKW KK'KS aw sw
-•§ :J «| rf|
^s ^s la la
J.J, j, • j, • j, •
II si li li
HO 5-U JHCJ tics
i
X
•"• o "^*
JjlJ
ill?
ii«t
03 <£ Cd
- * « S^
37
-------
Visual observation of all of the diffusers in a particular grid showed
heterogeneity at location 1 and increasing homogeneity at locations 2 to 4 (Figure 19). The
bicifilms at all locations were soft but they were sufficiently coherent that 4 cm2 sections could
be cut and lifted out without breakage. Few trapped air pockets could be seen and none were
seen that were larger than 1.5 mm in diameter. Visual examination of the biofilm surfaces
showed a large number of crater-like holes ranging in size from 1 to 10 mm. SEM showed
these crater-like holes very clearly (Figures 21 and 22) and measurements indicated that they
were as large as 1.5 mm. Because of the smaller areas that can be observed by SEM, other
larger holes that were observed visually were not seen by SEM.
The microbiological indices of biofilm formation, biofilm thickness and biofilm
carbohydrate per cm2, appear to be well chosen because results are consistent between methods
and because they confirm visual observations. The indices were successfully applied for biofilms
greater than 2.0 mm. These data indicate that biofilms formed unevenly on the high
permeability diffusers at location 1 near the inlet of Tank 1 of the Monroe plant, but that they
formed evenly on the less permeable diffusers at location 2 and decreased in thickness towards
the outlet (locations 3 and 4). The observation of biofilm heterogeneity at location 1 may be
linked to the higher organic loading and non-uniform mixing, effects typically found at the inlet
end of plug flow aeration tanks rather than to differences in diffuser permeability.
The morphological data confirmed previous observations that the size of
apparent bubble releasing structures on the surface of biofilms on the fouled diffusers ranged
from 0.5 to 10 mm, in wet and SEM samples. Further evidence of air passages within these
fouling layers (e.g. trapped air bubbles) indicated that the biofilms on these air-releasing
surfaces are traversed by a network of air passages and apertures that increase the bubble size
produced by fouled diffusers. :
The diffusers from Monroe were Sanitaire ceramic diffusers and had been in
sei-vice for two years. Interim measurements of SOTE, DWP and BRV had been made of these
diffusers during this period and are reported by Redmon (1989). After two years, the diffusers
demonstrated an SOTE of 15 to 19% at 1 cfm/diffuser depending upon location. There did not
appear to be a change in SOTE along the length of tank 1. Rather, the differences in SOTE
appeared to be random nor did they appear to be related to differences in diffuser permeability.
Moderate organic fouling was observed for all diffusers as shown by the mass of
foulant that ranged from 10.8 to 21.9 mg/cm2 with percent volatilities ranging from 54.6 to 68.7.
Although the variability appeared to occur at random, when expressed as mean values as shown
in Table 4, the foulant mass declined along the tank length similar to biofilm thickness and film
carbohydrate values. However, mean volatility increased along the tank length indicating
pei-haps that measurement of volatility includes components in the biofilm other than the
organic matter. The biofilm did affect the uniformity of bubble formation with moderate values
of BRV and DWP/BRV being recorded (9.6 to 18.1 in. and 0.46 to 0.60 respectively). When
expressed as mean values as shown in Table 4, there was a steady trend of increasing BRV
values along the tank length despite the declining biofilm thickness and film carbohydrate
values. The average values of percent acid soluble content also shows a decline along the tank
length. When coupled with a corresponding increase in percent volatility along the tank length
the trends perhaps indicate that the nature of the biofilm is changing along the tank length. It
would suggest that the biofilm contains more inorganics at the head of the tank, an area which
is heavily influenced by the properties of the influent. The inorganics may contribute to a more
open structure thereby minimizing the increase in BRV at the head of the tank. Along the
tank, the biofilm changes in nature to that of reduced biomass content (as expressed by biofilm
thickness and foulant mass), but more importantly, reduced inorganic content which may affect
the openness of the structure and deteriorate the uniformity of bubble release as expressed by
the increasing value of BRV. Moderate values of COV and EFR were also: observed, although
these were reduced to clean water equivalent values at the end of the second pass.
38
-------
On the other hand, values of DWP at 0.75 cfm were relatively low with values
close to clean water equivalents, indicating that the biofilm had not significantly affected the
diffuser head loss despite affecting uniformity of bubble formation as measured by BRV and
BRV/DWP.
Upon cleaning in the laboratory, the diffuser SOTEs were restored to values
ranging from 16.3 to 20.9% at Icfm/diffuser, an increase of approximately 10%. Inadequate data
were collected to demonstrate statistically significant differences but the data show little
difference in SOTE improvement between the cleaning processes used. Values of BRV, COV,
DWP and BRV/DWP were restored to clean water equivalent values. It appears that after two
years' service which included a cleaning in situ by the Milwaukee method after one year, the
biofilm had minimal impact on SOTE but a moderate impact on diffuser head loss as expressed
by DWP. ;
CONCLUSIONS - VERIFICATION STUDY '
. • i
The morphological characteristics of biofilms formed on the diffusers from the Monroe
activated sludge plant agree well with those seen in other plants examined in the
preliminary study. They lend support to a physical mechanism by which bacterial
biofilms can alter the size of bubbles released by diffusers as described in Figure 13.
The utility of measuring biofilm thickness with a penetrometer and the chemical
measurement of the bacterial component of biofilms by the phenol -H2SO4 method as used
in the preliminary stuHy has been confirmed in this phase of work.
Measurements of biofilm thickness and carbohydrate content do not appear to be directly
related to measurements of SOTE, DWP and BRV. However, the microbiological and
process characteristics are complementary and support the trends and observations made,
particularly when supported by foulant properties such as mass per unit area, volatility
and acid soluble content. Taken collectively, the data contribute to an improved
understanding of the plant diffuser operation. j
For the Monroe plant, biofilms were of moderate thickness and were found in a
homogeneous manner along the tank length except for some variability at the head end
of the tank. The biofilms appeared to have only a slight impact on SOTE whereas they
had a progressively deteriorating influence on BRV along the tank length. Diffuser head
loss (DWP) was not affected by the biofilm. . .. ;
LABORATORY SCALE SIMULATION OF
MICROBIAL FOULING AND CLEANING PROCEDURES
A laboratory scale system in which small ceramic discs were exposed to
wastewater and progressively fouled was developed to determine whether the formation of a
bacterial biofilm on an air releasing surface had the capacity to increase the size of air bubbles
released from that microbially fouled surface.
Procedures ;
Five small (1.25 cm diameter x 1.3 cm) fine bubble diffusers were mounted on a
manifold and placed in a tank 72 cm x 30 cm x 40 cm (Figure 23). The diffusers were made
from a porous stone medium used in the manufacture of Sanitaire ceramic disc diffusers. The
tank was filled with mixed liquor obtained from the Bonneybrook Wastewater Treatment plant
in Calgary, Alberta. Sludge was settled in the clarifier section and recycled to the aeration
39
-------
1
c/5
I
t
I
•a
o
CO
I
en
cs
40
-------
tank. 500 g of powdered skim milk was added daily in 50 L of feed giving an hydraulic
retention time of 0.58 days. Utility air at a pressure of 10 psi was regulated to provide a flow
of 0.03 cfm at each diffuser. This value was chosen because it provided the most even
distribution of bubbles across the diffuser surface. :
Three experiments were completed and twelve diffusers examined. In the first,
the air flow rate was not controlled and progressively decreased as a result of the back pressure
induced by the biofilm. At each sampling interval, the air flow rate and bubble size were
measured at this decreased air flow and again after the air flow rate was returned to the
original setpoint of 0.03 cfm. In the second and third experiments, new diffusers were installed
and the air flow rate was calibrated weekly to maintain a constant air flow of 0.03 cfm for the
duration of each experiment.
Biofilm development was monitored visually. Periodically, the diffiiser manifold
was transferred to a clear water tank and air bubble diameters were measured using a clear
plastic cylinder (30 cm x 12 cm x 3 cm). The cylinder was inverted, submerged and held over
each diffuser allowing bubbles to enter the open end. The bubbles were photographed and the
bubble size was measured against a grid on the back wall of the cylinder (Figure 24). In the
first experiment, bubble size was measured initially with no biofilm, at intervals during the
development of the biofilm, and again when the biofilm completely covered the diffuser surface
(after 6 to 8 weeks) at which time the biofilms were then characterized by the phenol-sulphuric
acid test for carbohydrates. The diffusers were then cleaned by soaking in 5% bleach for 24 hr
and bubble size was measured again for comparison with initial bubble sizes. '
In the second experiment, the effects on bubble size after cleaning with bleach
and acid were examined. Following the growth of a mature biofilm, one diffuser was left to
soak in 14% HC1 for 24 hr while a second was soaked in 5% bleach for 24 hr. A third was
placed in a tank of clean water, 100 ml of 5% bleach was poured into the air line which was
reconnected to the diffuser and the bleach was allowed to remain in contact with the air side of
the diffuser for 24 hr. The carbohydrate content of all three cleaned diffusers was measured as
well as that of a control diffuser which had not been cleaned. A fifth diffuser was retained for
measurement of colony forming units and SEM analysis. After cleaning, bubble sizes were
measured as in the first experiment, along with carbohydrate content, for comparison with
initial values for clean diffusers.
!
A mature biofilm was grown in the third experiment and then exposed to either
14% HC1 or 5% bleach that was delivered in liquid form to the air side of the diffusers via the
air line. Bubble size and carbohydrate content were not determined in this third experiment.
To compare the effects of 14% HC1 and 5% bleach cleaning on the artificially
generated biofilm with those on a naturally occurring biofilm, a fouled Sanitaire ceramic disc
from the Green Bay, WI. sewage treatment plant was obtained. The diffuser was broken into
approximately 4 cm2 pieces. Different pieces were placed in 5% bleach, in 14% acid or both in
sequence to see how each would affect the biofilm. The biofilm that remained was then
characterized using the carbohydrate assay and by measuring the biofilm solids content from a
4 cm2 area. Atomized bleach or acid could not be placed in the air line to the'. diffuser to attack
the large diffuser from the air side, as was done on the small diffusers.
Results
Bleiach Cleaning
The average bubble size produced by the unfouled fine bubble diffusers, at the
beginning of the experiment (9 October), ranged between 2.54 and 3.40 mm with standard
deviations ranging from 0.68 to 0.96 mm (Table 6). Direct observations through the submersible
41
-------
Figure 24. Bubble size determination. Each division in the grid equals 1 mm.
42
-------
TABLE 6. SUMMARY
Date
9 Oct 87
30 Oct 87
22 Nov 87
•
6 Dec 87
After bleach
cleaning
Carbohydrate
av. dia.1
st. dev.
no. of dets.
cfm
av.dla.
st. dev.
no. of dets.
cfm
av. dia.
st. dev.
no. of dets.
cfm
av. dia.
st. dev.
no. of dets.
cfm
av. dia.
st. dev.
no. of dets.
av. dia.
st. dev.
no. of dets.
cfm
av. dia.
st. dev.
no. of dets.
av. dia.
st. dev.
no. of dets.
cfm
av. dia.
st. dev.
no. of dets.
av. dia.
st. dev.
no. of dets.
cfm
av. dia.
st. dev.
no. of dets.
av. dia.
st. dev.
no. of dets.
(Jig/era1)
OF BUBBLE DATA FOR BLEACH CLE
8
2.54
0.68
82
0.03
2.89
0.66
46
0.01
2.81
0.83
37
0.03
2.6 S
0.56
25
0.01
6.5 L1
1.29
4
2.48 S
0.81
32
0.03
6.1 L
1.47
5
2.87 S
0.79
23
0.002
5.33 L
0.58
3
2.64 S
0.92
41
0.03
5.0 L
1.17
6
2.64
0.50
60
NA
C
3.03
0.95
43
0.03
3.47
0.81
42
0.005
3.25
0.77
50
0.03
4.8
0.42
10
0.002
4.12
0.74
47
0.03
4.54
0.50
12
0.002
4.38
1.53
36
0.03
2.95
0.68
69
31900
0
3.05
0.86
49
0.03
3.21
0.69
58
0.02
3.27
1.01
41
0.03
2.58
0.45
19
0.01
5.0 L
0.89
6
2.85
0.35
0.03
4.65
1.12
6
4.31
,0.72
13
0.003
4.05
1.26
48
0.03
2.91
0.60
87
38200
E
3.40
0.83
46
0.03
3.90
0.49
35
0.01
3.77
0.71
46
0.03
S 5.18
0.96
11
0.005
S 4.76
0.69
32
0.03
L
NA'
(no biofilm)
3.33
1.08
55
0.03
NA
F
3.19
0.96
39
0.03
3.41
0.54
56
0.02
3.35
0.94
42
0.03
4.69
0.94
21
0.01
4.07
0.71
55
0.03
5.75 ,
0.65
8
0.003
4.21
1.63
33
0.03
3.51
0.71
68
21200
Average diameter in ran
S refers to small diameter bubble series
L refers to large diameter bubble series
Not available
43
-------
plasitic tube showed no obvious heterogeneity of bubble sizes from different areas of these small
diffusers.
After 3 weeks of fouling (30 October), fairly uniform biofilms had developed on
all five diffusers (Figure 25) and the average bubble size had increased in each case, at the
reduced air flow rate produced by progressive bacterial fouling and at an air flow rate setpoint
of 0.03 cfm. Large differences in. air flow rate (see diffuser C on 30 October, Table 6) did hot
radically affect measured bubble size.
After 6 weeks (22 November), all five of the diffusers had developed thick
biofilms but those of diffusers B and D were very irregular in thickness (Figure 26). Bubble
size had increased in the more evenly fouled diffusers (C, E and F) at both the decreased air
flow rate produced by fouling and the air flow rate setpoint. Close examination of diffusers B
and D through the submersible plastic cylinder showed that some areas were fouled by thick
biofilms while other areas of the diffuser surface were almost clean. The heavily fouled areas
produced smaller numbers of very large bubbles while the cleaner areas produced large numbers
of much smaller bubbles. For this reason, averages of the large and small bubble sizes were
not calculated but were reported separately (Table 6). '
After 8 weeks (6 December), thick biofilms had developed on all five diffusers
but the biofilm on diffuser E was accidentally lost during transfer to the clear water tank. The
thick biofilms on diffusers C, D and F had produced increases in bubble size at both the
reduced air flow rate and the air flow rate setpoint. The biofilm on diffuser B was seen to
release large bubbles from its fouled area and small bubbles from its almost clean areas, when
observed directly using the submersible plastic cylinder. Therefore, bubble sizes were again
reported in large and small size categories (Table 6). At the end of the experiment, the mature
biofilms from diffusers B, C, D and F only were removed, assessed for carbohydrate content and
bleaich cleaned.
In every instance in which bubble size was measured at the reference air flow
rate of 0.03 cfm, there was a progressive increase in bubble size with increasing biofilm
accretion, and a return to near initial values when these fine bubble diffusers were cleaned
(Figure 27). The data from diffuser E indicate that the spontaneous removal of a fouling
biofilm returns bubble size to very near the initial unfouled value. The direct observation of
unevenly fouled diffusers (B and D) indicate that heavily fouled areas of these; diffusers produce
larger bubbles than almost clean.areas of the same diffusers observed at the same time and at
the same air flow rate values.
t
Bleach and Acid Cleaning \
The initial, diameter of the bubbles from the unfouled diffusers (7 March),
ranged between 2.82 and 3.09 mm with standard deviations ranging from 0.49 to 0.69 mm
(Table 7). After 3 weeks (29 March), a biofilm had started to form on all five diffusers but no
air channels were visible except for one that had formed on diffuser 3.
After 6 weeks (20 April), all five diffusers had thick biofilms (approx. 13 mm)
and all had visible air channels (Figure 28). The average bubble size with biofilms present had
increased and ranged between 3.77 and 4.62 mm with standard deviations ranging from 1.05 to
2.15 mm. The smallest average bubble size from diffuser #2 (3.77 mm) was also observed as
having two small air bubble streams coming from one area of the diffuser, with no biofilm
present which could account for a relatively smaller increase in average bubble size.
On 25 April, the diffusers were cleaned. Diffuser #5 was left to soak in 14%
HC1 for 24 hours, diffuser #4 was soaked in 5% bleach for 24 hours and the air side of diffuser
#3 was soaked with 5% bleach for 24 hours.
44
-------
Figure 25. Biofilm development after 3 weeks (30 October).
45
-------
Figure 26. Biofilm development after 6 weeks (22 November).
46
-------
CO O Q UJ U.
1
en
o
o
D
S
I
in-
(ujuj) saiqqng jo
47
-------
TABLE 7. SUMMARY OF BUBBLE DATA FOR ACID AND BLEACH CLEANING
Date
07 March
29 March
20 April
25 April
Cl waning
88 av.
St.
no.
88 av.
St.
no .
88 av.
St.
no.
88 av.
St.
no.
dia.1
dev.
of dets.
dia.1 '
dev.
of dets.
dia.1
dev.
of dets.
dia.1
dev.
of dets.
1
2.92
0.52
51
3.06
0.73
37
4.62
1.27
37
3.07
0.78
35
Procedure
2
2.82
0.69
42
2.97
0.60
29
3.77
1.05
35
2.94
0.64
31
control
3
3.09
0.60
46
3.30
1.18
42
4.59
1.21
24
3.14
0.87
32
100 ml 5%
bleach in line
4
2.90
0.57
2L
'3.01
0.69
45;
4.47
1.56
38;
2.97
0.5Q
23
5% bleach
soak
5
2.98
0.49
30
3.12
0-74 . ,
31 :
4 .27
2.15
22 ••
3.59
1.05
40 I
14% KCi
soak
Carbohydrate (pg/cm*)
37,200
1,920
680 i
15,200
cfu'a/diffuoar
3.70x10*
Average diameter in mm
48
-------
Figure 28. Biofilm development after 6 weeks (20 April), showing air channels in biofilms.
49
-------
The upper two-thirds of the biofilm on diffuser #5 had been detached and was
found floating in the tank. One small section of the cleaned diffuser had no 'biofilm but in the
remaining areas the air passages seemed to be partly occluded by a condensed biofilm residue
which may have accounted for the failure of the bubble size to return to initial clean values
(Table 7). The carbohydrate content of diffuser #5, after cleaning, was 15200 ng/cm2.
Diffuser #4 was soaked in 5% bleach for 24 hours. After 30 minutes,
approximately one quarter of the biofilm had been removed and some large pieces of biofilm
were found floating at the top of the tank. After 4 hours, two-thirds of the biofilm had been
removed. After 24 hours, approximately 95% of the biofilm was gone, the average bubble size
had decreased to values almost the same as initially measured, and the carbohydrate content of
the cleaned surface was 680 ug/cm2 (Table 7).
Diffuser #3 received 100 mis of 5% bleach from the air side.. ; Almost all of the
bleach went through immediately with a small amount remaining in-line. After 4 hours, the
biofilm thickness had declined and within 24 hours approximately 75% of the biofilm had been
removed and the average bubble size was approximately the same as measured with the clean
diffuser. The remaining 25% of the biofilm was still attached at the outer edge of the diffuser
that, had not been reached by the bleach. This made the carbohydrate content reading of 1920
ug/cm2 higher than it should have been in terms of biofilm fouling (Table 7), The cfu data
showed that there were many (3.70 x 109) living bacterial cells per cm2 of the fouled surface and
the SEM showed a thick biofilm with visible air passages (Figure 29).
Air Side Acid and Bleach Cleaning ;
In the third experiment only two diffusers were used, because the other three
had sloughed or lost their biofilm, one when removing the apparatus from the tank. With the
two remaining biofilms, diffuser #3 was cleaned with -100 ml of 5% bleach and diffuser #4 with
100 ml of 14% HC1. Approximately 90 ml of both cleaning agents passed through the diffuser
immediately with the remaining 10 ml of bleach or acid being carried through the diffusers over
a period of 2 to 3 hours. After 4 hours, the biofilm on diffuser #3 was visibly Deduced and after
7 hours, it had almost completely disappeared, with a carbohydrate reading of only 250 ug/cm2.
After 24 hours, diffuser #4 was still 75% covered with biofilm and yielded a surface
carbohydrate value of 8500 ug/cm2. ,
Cleaning of Full Size Diffuser
Cleaning with bleach and acid was carried out on pieces of a naturally fouled
ceramic disc. The control piece which had not been cleaned, registered a carbohydrate value of
3200 ug/cm2 (Table 8), and had a biofilm dry weight of 64 mg/cm2. One section of the diffuser
was placed in 5% bleach for 4 hours and another for 24 hours. After soaking in 5% bleach for
4 hours, a piece of the diffuser registered a carbohydrate value of 680 ug/cm2 and a biofilm dry
weight of 20 mg/cm2, but after 24 hours the carbohydrate and biofilm dry weight values were
both zero. Some non-measurable but visible residue remained on the bleach-cleaned diffuser but
after being placed into 14% HC1 for one hour it was visibly clean. Soaking pieces of the
diffuser in 14% HC1 for 24 hours resulted in a carbohydrate value of 1920 ug/cm2 and a biofilm
dry weight of 43 mg/cm2. The remaining biofilm although very soft was still attached to the
diffuser. The diffuser was then placed in 5% bleach for one hour and the remaining biofilm was
removed with carbohydrate and biofilm dry weight values reduced to zero.
50
-------
TABLE 8. CLEANING OF FULL SIZE DIFFUSER
Cleaning Procedure
Control 5% Bleach 5% Bleach Cone. HC1
4 hr 24 hr 24 hr
Cone. HC1-24 hr
+ 5% bieach-1 hr
Carbohydrate 3200
tUg/crn2)
680
0.0
1920
;0.0
Biofilm
Dry Weight
(mg/cm2)
20
.0.0
43
:0.0
51
-------
Discussion
It is recognized that this laboratory scale test system cannot duplicate
conditions in a full scale activated sludge plant equipped with fine bubble diffusers. The depth
of the laboratory tank is too shallow to simulate large tank conditions and the nutrient factors
affecting biofilm formation are probably different from those obtained in most regions of
activated sludge tanks.
The objective of these laboratory experiments was to determine whether the
development of a fouling biofilm on an air-releasing surface could change bubble size. This was
considered to be very important in any analysis of the effects of biofilm formation on bubble size
in activated sludge plants and it can be addressed only in a simple system in which the bubbles
produced by individual diffusers can be measured separately and related to the degree of fouling
of the diffuser in question. In these experiments, the diffusers could be returned to their initial
clean states by a variety of cleaning methods so that each diffuser was, in a sense, its own
control.
The gradual development of bacterial biofilms on the small diffusers, in this
scaled-down tank model, significantly changed the size of the air bubbles released from their
surfaces. The size of the released bubbles increased progressively as the bacterial biofilm
developed, and returned to initial values when the biofilms were removed from the diffuser
surfaces. In unevenly fouled diffusers, large bubbles were released from surface areas covered
by thick biofilms while much smaller bubbles were released from clean surface areas. These
data show that the development of bacterial biofilms on an air-releasing surface, in a laboratory
scale model system, consistently causes significant increases in the size of the air bubbles
released from these surfaces. Thus, in a reproducible laboratory-scale model, the principle that
bacterial biofilm development can increase the size of bubbles released from a fine bubble
diffuser has been established. ;
The biofilms developed in this laboratory-scale model system (Figure 29)
resemble those found on fine bubble diffusers (Figure 21) in activated sludge tanks, in that they
are thick, largely bacterial accretions, of cells and exopolysaccharides that are traversed by
branching air channels. These laboratory data suggest that the decrease in SOTE that is seen
in actual operating activated sludge plants may be related to increases in bubble size resulting
from the development of the thick bacterial biofilms, honeycombed with air passages, that have
been observed on operating fine bubble diffusers in this investigation at several locations.
These data on diffuser cleaning support the developing understanding of the
nature of the biofilms that can be established on air-releasing surfaces. Air pressures that
operate on these biofilms are directed towards the surrounding fluid, tending to remove accreted
bacteria from the fouled surface. Consequently, the biofilm can be retained !only if it develops
carbohydrate "tethers" that hold it on to the fouled surface. Bleach and acid were both effective
in cleaning fouled diffusers, if they are applied to the process water side in a soaking or hosing
mode, but bleach was especially effective if introduced from the air side and allowed to
hydrolyze the biofilm-retaining carbohydrate tethers. However, a practical method of delivering
bleach to the air side of fine bubble diffusers in operating activated sludge plants is needed to
implement in process cleaning of these systems. !
Conclusions
1. The progressive development of a bacterial biofilm on the surface of a fine
bubble diffuser can increase the size of bubbles released from that fouled surface.
52
-------
Figure 29. SEM of the surface of a biofilm on the laboratory diffuser showing detail of a large aper-
ture. Bar indicates 0.5 mm.
53
-------
2. 5% bleach is more effective in removing the biofilm than 14% HC1 on both
artificially induced and. naturally occurring biofilms and a combination of bleach followed by
acid will return the stone to its original condition.
3. Bleach cleaning from the air side is effective in removing the diffuser biofilm
in a laboratory-scale model system.
REFERENCES
Cossterton, J.W. (1980). Techniques for the study of adsorption of microorganisms to surfaces.
Absorption of Microorganisms to Surfaces. K.C. Marshall and G. Bitten, Eds., Wiley Interscience
50-70.
Coisterton, J.W. ei al (1987). Bacterial films in nature and disease. Ann. Rev. Microbiol. 41
435-464. fcjuu^ -i,
Gessey, G.G., W.T. Richardson, H.G. Yoemans, R.T. Irvin and J.W. Costerton (1977).
Microscopic examination of natural sessile bacterial populations from an alpine stream. Can. J.
Microbiol.. 23, 1733-1736. !
Redmon, D.T. and L. Ewing (1989). Report on the effect of pore size on oxygen transfer
capabilities, fouling tendencies, and cleaning amenability of ceramic diffusers at Monroe, WI.
Prepared for American Society of Civil Engineers. Ewing Engineering, Milwaukee, WI. •
Southwood, T.R.E. (1978). Ecological Methods (Second Ed.), Chapman-Hall Publishers, New
York.
Watkins, L. and J.W. Costerton (1984). Growth and biocide resistance of bacterial biofilms in
industrial systems. Chemical Times and Trends, October, 1984, 35-40.
54
-------
APPENDIX A
DETAILS OF CHARACTERIZATION PROCEDURES
1. Light Microscopy
To observe aqueous channels, the fouled diffuser was broken with a cold chisel and a
fragment was placed, biofilm uppermost, on the stage of a Zeiss dissection microscope
and examined between 7x - 30x magnification with ambient light as epi-illumination.
The size of trapped air pockets was observed by cutting 3 to 8 segments through the
biofilm, at right angles to the fouled surface, and mounting these segments on their sides
between a glass slide and a glass coverslip. A variable number of; air pockets seen
within these biofilm segments and air refractive pockets within them were measured by
reference to an optical grating. If air pockets were numerous, a smaller number of
segments were examined. !
2. Scanning Electron Microscopy ;
The diffusers were washed with distilled water to remove components of the bulk fluids
and were then fractured with a cold chisel to obtain fragments with surface areas of
approximately 2 cm2. The biofilms were then preserved by fixation for 2 hours in 5%
glutaraldehyde (a tanning agent used, on animal skins) in 100 nM cacodylate buffer at
20°C. Following two washed in cacodylate buffer to remove the fixative but not the
biofilm, the specimens were shipped to Calgary where they were prepared for SEM by
dehydration in ethanol and coating with platinum in a Balzers 300 freeze etching
apparatus. These biofilm specimens were then examined in a JEOL 450S scanning
electron microscope with a 45 degree tilt stage so that they could be rotated for viewing
from a series of favorable angles.
3. Biofilm Thickness ,
The pointed probe of a universal penetrometer was placed on the top of the biofilm and
the gauge was zeroed. The rod was then released and pushed down, if necessary, until it
reached the surface of the diffuser. The biofilm thickness was then recorded in mm.
4. Enumeration of Live Bacterial Cells
A 4 cm2 area of the diffuser was aseptically scraped with a sterile scalpel blade into a
phosphate buffer solution (pbs) and serial dilutions are made to 10s. A 1 ml sample is
taken from each dilution and plated in duplicate on Brain Heart Infusion (BHI) plates.
Plates with significant numbers (between 30 and 300 colonies) are counted.
55
-------
To calculate the number of bacteria per cm", the bacterial count is, multiplied by the
reciprocal of the dilutions and divided by the initial area used. i
5. Carbohydrate Assay
Measured against glucose standard curve of 200 ug/ml according to Dubois, M. ej al (1).
A 4 cm2 area' on the diffiiser surface is aseptically scraped with a sterile scalpel blade
into 4 ml of 2 N-HaSO4 and kept frozen for analysis in the laboratory. >
Reagents •
(1) Concentrated sulphuric acid
(2) Phenol (5%) in distilled water ;
i
Procedure ;
(1) Frozen vials are hydrolized for 20 to 30 minutes in a' beaker of boiling water.
(2) Samples are serially diluted 1/10, U100
(3) The "standard glucose series" are prepared as follows:
concentration
of glucose
distilled
water
Assay
0.5 ml of phenol reagent is added to 1.0 ml of sample followed by 2.5 ml of concentrated
sulphuric acid and vortexed immediately. The addition of sulphuric acid must be rapid.
Slow or sloppy addition leads to poor standard curves and results. Samples are
incubated in darkness for 60 minutes and then read at 490 nm on a UV
spectrophotometer.
0
0
1.0
20
0.1
0.9
40
0.2
0.8
60
0.3
0.7
80
0.4
0.6
200, ug/ml
1.0 ml
0 ml
(1) Colorimetric method for determination of sugars and related substances. Anal
Chem. 22 (3): 350-356, 1956.
56
-------