APPLIED AND ENVIRONMENTAL MICROBIOLOGY, June i986, p. 1186-11980099-2240/86/061186-13$02.00/0Copyright C) 1986, American Society for Microbiology

Countermeasures to Microbiofotling in Simulated Ocean ThermalEnergy Conversion Heat Exchangers with Surface and Deep Ocean

Waters in HawaiiLESLIE RALPH BERGER* AND JOYCE A. BERGER

Department of Microbiology, University of Hawaii at Manoa, Honolulu, Hawaii 96822

Received 4 October 1985/Accepted 7 March 1986

Countermeasures to biofouling in simulated ocean thermal energy conversion heat exchangers have beenstudied in single-pass flow systemis, hsing cold deep and warm surface ocean waters off the island of Hawaii.Manual brushing of the loops after free fouling periods removed most of the biofouling material. However, overa 2-year period a tenacious film formed. Daily free passage of sponge rubber balls through the tubing onlyremoved the loose surface biofouling layer and was inadequate as a countermeasure in both titanium andaluminum alloy tubes. Chlorination at 0.05, 0.07, and 0.10 mg liter-' for 1 h day'1 lowered biofouling rates.Only at 0.10 mg liter-' was chlorine adequate over a 1-year period to keep film formation and heat transferresistance from rising above the maximum tolerated values. Lower chlotination regimens led to the buildup ofuneven or patchy ifims which produced increased flow turbulence. The result was loWver heat transfer resistancevalues which did not correlate with the amount of biofouling. Strfaces which were let foill and then treated Withintermittent or continuous chlorination at 0.10 mg of chlorine or less per liter were only partially of unevenlycleaned, although heat transfer measurements did not indicate that fact. It took cdntihubus chlorination at 0.25mg liter-' to bring the heat transfer resistance to zero and eliminate the fouling layer. Biofouling in deep coidseawater was much slower than in the warm surface waters. Tubing in one stainless-steel loop had a barelydetectable fouling layer after 1 year in flow. With aluminum alloys sufficient corrosion and biofouling materialaccumulated to require that some fouling coutermeasure be used in long-term operation of an ocean thermalenergy conversion plant.

The ocean thermal energy conversion (OTEC) systemconnects a large heat source of warm surface ocean waterand a potentially unlimited volume of deep cold ocean waterwith a heat engine to drive a conventional electricity-producing turbine generator. In a closed-cycle OTEC plantheat exchangers (HXs) extract the thermal energy fromwarm ocean waters. A small fraction of that energy isconverted to electrical power and waste heat is rejectedthrough a second HX to cold water pumped from the oceandepth. Solar energy absorbed by the ocean surface providesthe heat source.The efficiency of the OTEC system is inherently low. The

theoretical maximum based on the temperature differencebetween the water masses is only 6 to 7%. Thermal losses,the power requirements to pump large volumes of seawaterand working fluid, power losses in turbines, generators, bymicrobiofouling (the attachment arid growth of microorga-nisms and their products on the HX surfaces), and corrosionof these HX surfaces by seawater are critical problems to beovercome ifOTEC systems are to become a practical reality.All of these factors may reduce the net efficiency to as littleas 1 or 2%.

Unlike conventional power plants, however, the OTECsystem uses an inexhaustible and virtually cost-free fuelwhich offsets the inherently low thermnodynah-iic efficiencyand will make it a competitive alternative to energy gener-

ated from increasingly costly fossil and nuclear fuels.Many microorganisms adhere to surfaces, often with

specificity and considerable firmrness. Chemical receptors onthe surfaces of the cells or their appendages and in theirextracellular secretions may bind the cell to other surfaces

* Corresponding author.

and structures. In bacteria, in particular, exopolysaccha-rides (2) are secreted by the cells to form masses of fiberswhich have been termed "glycocalyx" (3). The bacterial cellsurfaces and the adhering polymers are in turn suitable locion which other bacteria and particulate materials may attachor on which inorganic precipitates may forrn. Such accumu-lations may be termed biofilms. A film only ca. 25 to 50 ,umthick may further reduce by 40 to 50% the heat transferefficiency of the HXs, rendering the entire OTEC systemeconomically impractical.

This paper reports studies of microbiofouling and counter-measures to control or retard it. Long-term experiments arereported using open loops through which either surfacewarm water or deep cold sea-water flows year-round andcontinuously. The experiments have run for several years.They are part of an integrated program to determine thematerials and conditions required for long-term OTEC oper-ation.

MATERIALS AND METHODS

Site and experimental setups. The experimental site was atthe Seacoast Test Facility at Ke-ahole Point on the westerntip of the island of Hawaii. The surface (warm) water rangesbetween 24 and 280C and it is collected from about 100 moffshore. There, a large gyre of current brings in open oceanwater. The shoreline is arid and mainly lava rock withrelatively little vegetation. Most of the year, there is littlerainfall and terrestrial runoff. The bottom drops off rapidly.The warm water is pumped through two 0.3-m-diameterpolyvinyl chloride pipes to large header tanks from which theexperiments are fed by gravity.The cold water intake is located at a depth of -580 m, -20

m above the sloping ocean floor some 1,400 m from the

1186

Vol. 51, No. 6

on October 13, 2017 by guest

http://aem.asm

.org/D

ownloaded from

COUNTERMEASURES TO MICROBIOFOULING 1187

TABLE 1. Operating conditions of the simulated HXs (open loops)Loop no. Material Operating mode

Warm surface seawater1 Titanium (2) Cycles of free-fouling and hand brushing2A Titanium (2) Soft Amertap balls, 12 passes h-', 1 h day-', brush cleaned when re-

quired2B Aluminum (3004) Hard Amertap balls, 12 passes h-', 1 h day-', brush cleaned when re-

quired3 Titanium (2) Chlorinated at 0.05 mg liter-' for 1 h day-'5 Aluminum (3003) clad with Cycles of free-fouling and hand brushing

aluminum6 Titanium (2) Cycles of free-fouling and continuous chlorination at 0.05 mg liter-' or

as indicated7 Aluminum (3004) clad with Chlorinated at 0.10 mg liter-' for 1 h day-'

aluminum10 Stainless steel (AL6X) Chlorinated at 0.07 mg liter-' for I h day-'11 Aluminum (5052) Cycles of free-fouling and chlorination at 0.10 mg liter-' for 1 h day-'

Cold deep ocean water13 Stainless steel (AL6X) Free-fouling16 Alclad (3003) Free-fouling17 Aluminum (5052) Free-fouling

shoreline. Details of these inlet lines have been fully de-scribed (J. Larsen-Basse, in Y. Mori and W.-J. Yang, ed.,ASME/JSME Thermal Engineering Joint Conf. Proc.2:285-289, 1982). The cold water is pumped directly into thelaboratory building where it is diverted into the various testloops.Each test loop is fitted with a heat transfer monitor (HTM)

plus sequential strings of corrosion and microbiofouling testcoupons of the same test metal. Test coupons vary from 100to 250 mm in length. They are 25.4 mm in outside diameter.Aluminum tubes are 22 Inm in inside diameter; titanium andstainless-steel tubes are 24 mm in inside diameter. Flowvelocities were maintained at 1.8 m s-1.Table 1 summarizes the operating conditions of the simu-

lated HXs discussed in this paper.

14

12

Io_

8 _

31

E

un

Lr0

6 _

4

2 _

o

21R750 850 950

Heat transfer measurements. The HTMs were modifiedversions (T. M. Kuzay, G. N. Granneman, and A. P. Gavin,in W. L. Owens ed., ASME HTD 12:39-46, 1980) of thedesign by Fetkovich (J. G. Fetkovich, G. N. Granneman,L. M. Mahalingam, and D. L. Meier, Argonne NationalLaboratory Rep. OTEC/BCM-002, Argonne, Ill., p.237-380, 1978) in which direct heating of the water is usedtransiently to determine the heat transfer resistance (HTR)of the test tubing.HTR values are cited in the text in Rf units, which are in

meters square degree Celsius per watt x 105. The maximumHTR value tolerated in most of these experiments beforeantifouling measures were taken was approximately 9 Rfunits.Three types of biofouling countermeasures were used: (i)

280 -j'Eu240 M

200Z

120

801Lo

40

0

1050 1150 1250 o.

TIME IN FLOW (days)FIG. 1. HTR and biofouling levels in titanium loop 1. The loop surfaces were allowed to free-foul to about Rf 9 and then were cleaned with

10 manual passes of a bristle brush while seawater flowed slowly. Flow was resumed at 1.8 m s-' and the cycle was repeated. Only a portionof the experiment is shown.

LOOP 1, Titanium

VOL . 5 1, 1986

on October 13, 2017 by guest

http://aem.asm

.org/D

ownloaded from

1188 BERGER AND BERGER

TABLE 2. Percentage of total film removed by extraction for 15min in loop 1l

OC + TN Film DWDays In flow (%) %

81 96 95323 85 88610 58 60

a Total film removed was estimated by extrapolation of the kinetic datacurves to the point at which no further film would be removed on continuedextraction.

EXTRACTION TIME (min)

FIG. 2. Amount of biofouling removed from loop 1 plottedagainst extraction time. Biofouling samples were in the flow fordifferent periods of time. Solid symbols are DW; open symbols areOC and TN.

manual brushing of the tubes, in which a bristle "test tube"brush was passed through each arm of a loop five times (10strokes) for each "brush cleaning" while water flowedslowly; (ii) cleaning with sponge rubber balls (Amertap Co.,Parkwest, Woodbury, N.Y.), in which one sponge rubberball with a diameter of 25.5 or 26.5 mm (depending on thetubing used) was released every 5 min for 1 h once per 24-hperiod; and (iii) chlorination to prevent or retardmicrofouling or to remove already formed films, in whichhypochlorite was generated electrochemically in situ imme-diately upstream of the HTM. Residual chlorine was deter-mined by amperometric titration from samples taken imme-diately downstream of the HTM or at the end of theexperimental loop 10 to 15 m downstream. Chlorination wasdone either for 1 h day-' or continuously following a periodof free-fouling. In either case, chlorination was, done atnominal 0.05, 0.07, or 0.10 mg liter-1 except as noted.Sample preparation for scanning electron microscopy

(SEM). Biofouling samples were removed from loops and

treated as follows. A piece of tubing about 2.5 cm was cut offfrom the sample (coupon), dipped in filtered seawater, andthen fixed in 4% glutaraldehyde in filtered seawater, at pH7.2 for 1 to 2 h at 15°C in the dark. The sections were dippedin filtered distilled water and then for 5 mmn each in 5%ethanol-distilled water (twice), 25% ethanol, 50% ethanol,and 75% ethanol. Iced samples were flown to Honolulu intightly closed jars in 75% ethanol. The dehydration processwas continued in 95% ethanol (twice), absolute ethanol(twice), and xylene (twice). Samples were then air dried andstored in a desiccator until they were cut and mounted onSEM stubs. They were then coated with carbon and goldbefore examination in the SEM (Cambridge Stereoscanmodel S4-10 or ISI model DS 130).For dry weight (DW), organic carbon (OC), and total

nitrogen (TN) analyses, the other section of the tubingsample was rinsed in filtered distilled water, drained, baggedin polyethylene, and frozen in dry ice. The frozen sampleswere flown to Honolulu and kept frozen until analyzed.

Extraction procedures. Sections of the frozen sampleswere cut in 2.6- to 3.3-cm lengths. One end of the cut piecewas covered with Parafilm (American Can Co., Greenwich,Conn.) and a silastic cap. A 0.5-ml amount of acid-washedBallotini-like microglass beads (Scotch Lite; 3M Co., St.Paul, Minn.) and 2 ml of glass double-distilled water wereadded. The open end was closed as above. The sample was

16

14

:3N,

E

u0

U.)

4-C_

a

12 _

10 _

6 _

4

2

0 _600 700 800 900 1000 1100 12C

900 CTEu

800 N

700z

600 e_

5004

400 ;;I

300 0-2200 >-

100 0

-J

0o° 170

TIME IN FLOW (days)FIG. 3. HTR and biofouling levels in Alclad (type 3003) loop 5. The tube surfaces were allowed to free-foul to about Rf 9 and then were

cleaned with 10 passes of a bristle brush while seawater flowed slowly. Flow was resumed at 1.8 m s-' and the cycle was repeated. Only a

portion of the experimental data is shown.

0LOOP 5, Alclad 3003 A

A AA # A A 4 A A A AA

APPL. ENVIRON. MICROBIOL.

rliEu"Icn3

z

u

L0

1--xU

.J'31F

cx0

8 _

on October 13, 2017 by guest

http://aem.asm

.org/D

ownloaded from

COUNTERMEASURES TO MICROBIOFOULING 1189

IFIG. 4. SEM pictures of loop 5. (A) A complex film exists on the tube surfaces at Rf 9.3 before manual brushing (bar = 2 i±m). (B) Note

areas free of film after brushing to Rf 2.3 (bar = 2 pLm).

placed in a model Vi2 vibrating ballistic disintegrator (RHOScientific Co., Commack, N.Y.) and shaken at approxi-mately 100 Hz for 5 min. After allowing the glass beads tosettle, the supernatant fluid was removed. Two milliliters ofglass double-distilled water was added to the sample, and thefilm removal procedure was repeated twice. The aqueoussuspensions were pooled.For kinetic studies, fouled samples were progressively

ballistically extracted with eight sequential 5-min treat-ments. The extracts were analyzed individually.

Aliquots (1 ml) of the extracts were hydrolyzed in 4 N HClat 121°C for 1 h. The hydrolysates were then dried over P205and NaOH in vacuo at 55°C overnight. For each sample twovials of hydrolysate were used. Each hydrolysate was takenup in 0.1 ml of glass double-distilled water and pooled foranalysis. A 20-,ul portion of pooled hydrolysate was added to

14

122

10 -

4

0

800 850

weighed tinfoil cups (two cups per sample) and dried invacuo over P205 and NaOH at 55°C. A second aliquot wasadded to the tin cup and the sample was then dried overnightas above. Cups were closed and then weighed on a microbal-ance (model 21; Cahn Instrument Co., Cerritos, Calif). Thelength of tubing used for the extraction was measured tocalculate the area of the inner surface. DWs (less carbon-ates) are reported in micrograms per square centimeter. OCand TN were determined on these samples in an elementalanalyzer (Carlo Erba model 1106; Haak-Buchler Instru-ments, Saddle Brook, N.J.). Values are reported in micro-grams per square centimeter.

RESULTSMechanical cleaning experiments. (i) Brushing as a fouling

countermeasure. (a) Loop I (titanium, free-fouling and peri-

900 950 1000 105C

280EU

240 N

j

200z

160 %

.4120 Oa

80 LD

40 3

0 0

11$-LL

0TIME IN FLOW (days)

FIG. 5. HTR and biofouling levels in titanium loop 2A. Twelve sponge rubber balls were passed through the sytem over 1 h, once dailyto control fouling. When HTR reached about Rf 5, the tubes were cleaned with 10 passes of a bristle brush while water flowed slowly. Theflow was then restarted at 1.8 m s-1, and the use of sponge rubber balls was resumed. Only a portion of the experimental data is shown. Theasterisk indicates that four extractions were used instead of three.

LOOP 2A, Titanium

O*8 _

**

6 _A

2 _

e

VOL. 51, 1986

on October 13, 2017 by guest

http://aem.asm

.org/D

ownloaded from

1190 BERGER AND BERGER

odically brushed). Figure 1 shows part of the data from loop1. The experiment ran for -3.5 years, although biofoulingcoupons were installed and replaced at various times. Earlysamples were undated. The loop was allowed to free-fouluntil the HTR reached Rf -9 (which took between 28 and 42days). It was then brushed to bring the HTR to near 0.

Microscopically, films were complex with filamentousorganisms of various sizes and colonies of cocci, rods, andother forms all imbedded in a glycocalyx matrix. Macroscop-ically, the films appeared evenly dispersed and quite thin.The DW never exceeded 150 ,g cm-2 and the OC plus TNaveraged -29 jxg cm-2, although the HTR was approxi-mately at Rf 9.The recalcitrance of older films to removal was tested

kinetically. Figure 2 shows the kinetics of film removal asfunctions of the time of extraction and the actual days inflow.

Table 2 shows the percentage of the total film removed byextraction for 15 min.The data show that long periods of time in the flow and

frequent brushings produced an increasing film tenacity.(b) Loop 5 (Alclad 3003, free-fouling and periodically

brushed). Figure 3 summarizes some data from loop 5. Theexperiment had been running for -3.5 years. Biofoulingsamples were installed in the loop 376 and 667 days after thestart of the experiment and additional replacement sampleswere added between days 700 and 900. None of thesesamples were labeled, although their positions in the loopwere noted.Tubes were let free-foul from 26 to 43 days until the HTR

reached Rf 9. They were then brushed back to Rf 1.5 to 2.5.The period between brushings was influenced by the

condition of the bristle brushes and seasonal factors, inparticular, water temperature. Film DW and OC-plus-TNanalyses and SEM photographs (see Fig. 3 and 4B) demon-strated that there was always a fouling film residue afterbrushing. The bacterial flora also changed with time. Formsevident in Fig. 4A include prosthecate and spiral or colonialhat-shaped bacteria. The latter have not been previouslydescribed to our knowledge.

Before brushing, the inner tube surfaces appeared smoothand evenly coated with a layer which varied from off-whiteto pale yellow. SEMs showed that the tube surfaces werecovered with complex films similar in appearance to thosefound in (titanium) loop 1. After brushing the surface ap-

100190807060

_ 50

< 40

n: 302 2LL

20

FIG. 6. SEM showing biofouling on the surface of titanium loop2A through which sponge rubber balls had been passed as a foulingcountermeasure. Note the combed appearance of the surface andthe smaller diversity of microorganisms than in Fig. 4. Loosesurface material is removed by the passage of the balls. Bar = 2 ,um.

peared streaked in the direction of the brushing. The DW ofthe extracted films just before brushing varied from 286 to460 ,ug cm-2 which depended on the age of the sample, i.e.,the number of free-fouling/brushing cycles and the efficiencyof the brushing procedure itself. The OC-plus-TN valuesaveraged 24 ,ug cm-2, values which were slightly lower thanthose in loop 1. HTR values before brushing varied from Rf9 to 11.

(ii) Sponge rubber balls as a fouling countermeasure. (a)Loop 2A, which ran for 34 months, using Amertap spongerubber balls and titanium tubing. Only 6.5 months of thisexperiment is shown in Fig. 5. The HTR increased slowlycompared with untreated free-fouling surfaces (Fig. 1).When the HTR approached Rf 5, the loop was brushed. HTR

EXTRACTION TIME (min)FIG. 7. Total biofilm remaining versus extraction time in free-fouled (loop 1) and sponge rubber ball-treated (loop 2A) titanium tubing.

Symbols: Loop 1, squares; loop 2A, circles; film DW, closed symbols; OC plus TN, open symbols.

APPL. ENVIRON. MICROBIOL.

on October 13, 2017 by guest

http://aem.asm

.org/D

ownloaded from

COUNTERMEASURES TO MICROBIOFOULING 1191

k0

350° PEum300

Q tf 250 -Iu

I - ~~~~~~200 0.f1i/C\o4

-150

0-~100 w

50 =co

25 50 75 00 25 10H15(25 50 75 100 125 150 175 200 ILi

TIME IN FLOW (days)FIG. 8. HTR and biofouling levels in aluminum (type 3004) loop 2B. A sponge rubber ball was passed through the system 12 times over

1 h, once a day to control fouling. The system was cleaned with 10 passes of a bristle brush when Rf reached about 9. The use of the sponge

balls was then resumed.

measurements, biofilm analyses, and both macroscopic andmicroscopic examination of coupon surfaces all showed thatthe cleaning by the sponge rubber balls was not complete.SEM and macrophotographs indicate that the scrubbingaction was irregular giving rise to a combed-like appearanceon the biofouling layer. The microflora was much lessdiverse than that found on free-fouling surfaces (Fig. 6).Manual brushing was required periodically to bring the HTRback to near zero. Following each such brushing the filmreformed at a greater rate than the previous cycle despite thedaily sponge rubber ball treatment. For example, in one

cycle towards the end of the experiment, it took 192 days toreach Rf 9. In the next cycle, it took 128 days. Thisrepresents a 1.8-fold fouling rate increase.Except for the first 10 min of film extraction from loop 1,

first-order kinetics, which are characteristic of ballistic treat-ments of homogeneous layers, occurred in samples fromloops 1 and 2A. This is indicated in Fig. 7; the lines in thesemilog plot are both straight and parallel after 10 min oftreatment.

(b) Loop 2B (aluminum type 3004 treated with sponge

rubber balls with occasional brushing). In June 1984, loop2A was dismantled, acid cleaned, and refitted with aluminumtype 3004 parts. Flow was restarted and the sponge rubberball treatment was resumed (Fig. 8). In the first cycle, thetime to foul to Rf9 was -73 days. After brushing the Rf againrose to 9 in -20 days, with the third and fourth cycles takingonly 17 and 18 days, respectively.Manual brushing of the aluminum surface was not as

effective as with the titanium loops. The data show that noneof the measured parameters returned near to their startingvalues. The experiment was stopped after 5.5 months, whenit was evident that this use of the sponge rubber balls was notan effective countermeasure on the aluminum alloy.SEMs showed that on the fouled surface of aluminum type

3004, bacteria are both attached to the surface and imbeddedin the aluminum hydroxide layer (Fig. 9).Chemical cleaning experiments. (i) Loop 3 (Ti, chlorinated

for 1 h day-' at 0.05 mg liter-'). Loop 3 was started in early1981. Biosamples were added at various times beginning inlate 1982. Rigorous record keeping of these samples was

begun in 1983. HTR during the first 1,300 days ranged fromRf 0.3 to 3.3 (Fig. 10), indicating that chlorination at a

nominal 0.50 mg liter-1 for 1 h day-' is sufficient to maintaingood heat transfer. However, the OC, TN, and film DWanalyses of the biocoupons showed that fouling was occur-ring. In fact, these parameters were considerably higher at Rf3.3 than those found in loop 1 at an HTR of Rf 8.7. DWvalues were 486.4 and 97.21 pLg cm 2, and OC-plus-TNvalues were 78.54 and 19.51 ,ug cm-2 in loops 3 and 1,respectively. Both SEM and the visual appearance of thesurfaces showed uniform biofilms in loop 1 which completelycovered the tube surfaces. By contrast, those of loop 3 hadlight, patchy, and clear areas alternating with extremelyheavy patchy films with no uncovered areas (Fig. 11A andB). The conical structure is probably a plate from a

dinoflagellate. Some unusual bacterial forms are shown in

FIG. 9. SEM showing biofouling on the surface and imbedded inthe aluminum hydroxide corrosion layer of aluminum (type 3004)loop 2B. Sponge rubber balls were passed through the system as a

fouling countermeasure. The film shows a sparse distribution of thediverse microflora after 59 days in flow. Bar = 5 ,um.

16

14

:3

C\ME

u

Ln0

In

12

10

8

6

4

2

VOL. 51, 1986

on October 13, 2017 by guest

http://aem.asm

.org/D

ownloaded from

16

14

:3

N

E

0

Lf)1-

121

101

81

61(4.-

4

2

O--800 900 1000 1100 1200

600

500

400

300

200

100

I0

1300

E0N

z+

U

40'

I-~

w:3

1-JLL

0TIME IN FLOW (days)

FIG. 10. HTR and biofouling levels in titanium loop 3. The biofouling countermeasure was chlorination at 0.05 mg liter-' for 1 h day-'.Data are shown after the first 800 days of operation. The biofouling samples were in the flow for various periods of time: A, 740 days; B, 358

days; C, 373 days.

FIG. 11. SEMs. (A) Biofouling in titanium loop 3, which was chlorinated at 0.05 mg liter' for 1 h day-'. HTR was Rf 2.7. The patchyquality of the film results in abnormally low HTR values due to turbulence (bar = 250 ,um). (B) Magnification of a section of the biofilm in

(A). Note the diversity of the microorganisms in the fouling layer (bar = 2 ,um).

LOOP 3, Titanium0

A A

A0A A

8 C

0 v

& 0 AA A vp-

on October 13, 2017 by guest

http://aem.asm

.org/D

ownloaded from

COUNTERMEASURES TO MICROBIOFOULING 1193

ARM1-6 1-3 1-1

1 HTM E INLET

2-1 2-5 2-122 l_ 1

3-9 3-5 3-13

4-1 4-114 OUTLET

FIG. 12. Diagrammatic representation of loop 3, which is typicalof the other loops. The electrolytic chlorinator (EC) was not presentin all loops. The HTM was made of the same alloy as the samplecoupons. Other parts of the loop were of polyvinyl chloride plasticpipe. Sample coupons were connected to each other by nylon orsilastic couplings. The numbering system for the test samples isgiven. Each arm was approximately 3.5 m long. The numbers of testcoupons in each arm and loop varied.

Fig. llB. The chlorination regimen slows film formation, butis does not prevent it.

Figure 12 is a diagram of the setup for loop 3 and otherloops showing the general positioning and the numberingsystem for biocoupons as used in this study.

(ii) Biofouling and position of samples in the loop. There isevidence to indicate that with chlorination at low levels (e.g.0.05 mg liter-') the hypochlorite generated was not uni-formly in the bulk water as it flowed through the HTMs.Thus the HTM and those coupons closest to it were probably"cleaned" with more than 0.05 mg of hypochlorite liter-.Based on the appearance of the biocoupons, uniform mixingappears to have occurred by 1 m downstream of thechlorinator. Chlorination caused small pieces of the foulinglayer to slough off periodically. The resulting surface rough-ness produced added turbulence which in the HTM results inlowered values of HTR. The level of residual chlorine wasmeasured on several occasions to be 15 to 20% lower at thefar end of the loop than just downstream of the HTM. Overa period of -3 years a 1- to 2-mm-thick biofouling layerresulted.

Figure 13 shows the effect of the time in flow on biofilmdevelopment. Only the data points for the sample closest tothe HTM and the chlorinator (in the 1-1 position) are not onthe curve. Additional data shown in Table 3 support the fact

NE

m

LnNOa

z

J

Ln)

I-

z

LII

TABLE 3. Effect on biofouling of position of samples in loop 3a

Position Days in OC + TN DWin flow flow (pLg cm-2) (pg cm-2)1-1 349 28.84 125.51-3 >740 115.4 545.31-6 169 6.75 34.712-1 >740 160.0 1,0022-7 >740 135.0 637.93-1 220 32.61 128.63-2 >740 291.0 1,5523-12 >740 435.4 2,750a HTR was Rf 2.9 when the loop was dismantled.

that, except near the HTM, the amount of biofouling is notdependent on the position of the sample in the loop.The lumpy and patchy coating on the HTM surface at the

end of the 1,300-day experiment was not indicated by thelow HTR values.

(iii) Loop 10, (stainless steel [AL6X], chlorinated for 1 hday-' at 0.70 mg liter-'). Loop 10 was started in 1982.Twenty biocoupons were added sometime during that year.Over the 372 days of the experiment, HTR varied between Rf0.6 and 2. The data indicate that chlorination at 0.07 mgliter-1 is sufficient to maintain good heat transfer. OC, TN,and DW analyses of the films showed that fouling occurredmore slowly than that in the free-fouling cycle of loop 1. Thebiofouling parameters in loop 10, at Rf 1.8, howfver, werehigher than those found in loop 1 at Rf 8.7. DWs were 128.5and 97.21 ,ug cm2 and those for OC plus TN were 24.5 and19.51 ,ug cm-2 in loops 10 and 1, respectively.

(iv) Loop 7 (Alclad [3004], chlorinated at 0.10 mg liter-l for1 h day-'). Loop 7 was run for about 15 months. The HTRgradually rose, reaching a maximum of Rf 2.8. Figure 14summarizes data for 350 days of the experiment. The DW ofthe films varied between 196 and 432 ,ug cm2, while the OCplus TN was between 3.5 and 7.0 pug cm-2. After a fewmonths of flow, the surfaces of the tubes were sparselycovered with a limited morphological range of bacteriaimbedded in and on the inorganic/polysaccharide film matrix(Fig. 1SA). Four months later the film had become morecomplex (Fig. 15B), but the amounts of OC and TN had onlyfluctuated slightly.

(v) Loop 6 (Ti, free-fouling, then cleaned by continuouschlorination at 0.050 mg liter-'). In the first experiment in

600

U500 0

400* (1-5)U

300

200 _(3-1) * (1-1)

100 -

a (1-6)n ---

f00 200 300 400 500 600 700 800

ACTUAL TIME IN FLOW (days)

FIG. 13. Loop 3. Film DW and OC plus TN in the fouling layer are plotted versus the time coupons were in the flowing system. Positionsof the samples in the loop are indicated in parentheses. HTR varied between Rf 2.9 and 3.3.

VOL. 51, 1986

on October 13, 2017 by guest

http://aem.asm

.org/D

ownloaded from

1194 BERGER AND BERGER

16

N

C\jE

0

0

4IxUr

0 50 100 150 200 250 300 350

500

400

300

200

100

0

TIME IN FLOW (days)

FIG. 14. HTR and biofouling levels in Alclad (type 3004) loop 7. Biofouling countermeasures consisted of chlorination at 0.10 mg liter-'for 1 h day-'.

loop 6, free-fouling followed by continuous chlorination at0.05 mg liter-' invariably reduced HTR to Rf 0 or less, butnot the film OC, TN, or DW. After 15 months of cyclicoperation in this manner, when the continuous chlorinationperiod was extended, Rf values dropped below 0 and stayednegative. To test whether increased turbulence was thecause, the chlorine level was increased to 0.1 mg liter-1 andrun for an additional 50 days. HTR continued to drop butthen started to rise. Concurrently, film DW, OC, and TNdecreased. Increasing the chlorine to 0.25 mg liter-1 even-tually brought the HTR up to Rf 0, at which point examina-tion of the tube surfaces showed them to be free of allbiofilm.Loop 6 was dismantled, acid cleaned, and restarted with

new biofouling coupons. The regimen consisted of free-fouling to Rf 9 followed by cleaning with continuous chlori-nation at 0.05 mg liter-'. The first four cycles of theexperiment are shown in Fig. 16. Although each new cyclestarts near Rf 0, there is a progressive increase in the amountof base-line fouling.

(vi) Loop 11 (aluminum [5052], free-fouling, then chlori-nated at 0.10 mg liter-' for 1 h day-'). This experiment was

designed to test whether chlorination at 0.10 mg liter-' for 1h day-' was sufficient to remove preformed biofouling layerson aluminum HX surfaces. Three cycles of fouling andcleaning were done. Figure 17 shows the biofouling data forthe second and third cycles. Chlorination does not return theHTR to zero and each succeeding free-fouling cycle starts at

FIG. 15. SEMs of loop 7 which had been chlorinated daily for 1 h at 0.10 mg liter-'. (A) Fouling layer at Rf 1.8. Note the

exopolysaccharide slime (bar = 1 ,um). (B) Sample taken 4 months later at Rf 2.0. Note the diversity of organisms (bar = 1 ,um).

,4-LOOP 7, Alclad 3004 0

20

0~~8

06 00

4 ~~~~~~~~AAA A

2

O

APPL. ENVIRON. MICROBIOL.

I

I

I

on October 13, 2017 by guest

http://aem.asm

.org/D

ownloaded from

COUNTERMEASURES TO MICROBIOFOULING 1195

IN:3:

C\ME

u0

Ur)0

(4-UCY

700 c\jEu

600 .

500z

400 u

300 4oD

200 I

w100 :¢

-J11La

0TIME IN FLOW (days)

FIG. 16. Plot of first four cycles of data for free-fouling and cleaning by continuous chlorination at 0.05 mg liter-' of the cleaned andrefurbished loop 6 titanium (second experiment).

a higher Rf value. Similar increases occur for OC, TN, andfilm DW. Table 4 shows that for the second and third foulingcycles, at comparable Rf values, the biofouling parametersare not the same; an uneven removal of the fouling filmchanges the heat transfer properties of the system. Similarobservations were made for loop 6.Cold water loops (fouling with deep seawater). (i) Loop 13

(stainless steel [AL-6X]). During the 1-year period over whichbiocoupons were monitored, the HTR ranged from about Rf-1.0 to +3.3. Biofouling was not significant; SEM photo-graphs showed mainly inorganic deposits.

(ii) Loop 16 (Alclad [3003]). Biocoupons became com-pletely corroded within 1 year and were useless for furtheranalyses. Before corrosion began, however, biofouling was

16

14 _

12 _

10 k

observed in SEM photographs. A type of filamentous bac-terium was imbedded and adhered to the hydrated aluminumoxide corrosion layer. The OC and TN comprised only about2% of the total film weight and did not contribute signifi-cantly to the HTR.

(iii) Loop 17 (aluminum [5052]). Data for approximately 2years of flow are shown in Fig. 18. A second set ofbiocoupons was added to the system after -6 months ofrunning. Some of the coupons tested are assumed to be fromthe second set.HTR increased nearly linearly during the entire period.

After about 2 years the Rf reached 4.4, just 50% of thetolerable fouling level. The film mass accrued linearly duringthe first year and OC and TN also increased. The organic

LOOP 11. Aluminum 5052

As00

8 _

6 _0

4 _

2 _

u100 150 200 250 300 350

- 500- 450

400

350

_ 300

250

200

_ 150

_ 100

50

0400

EUN.

z

.

4od

I-

w06

1-JLLi3

W

_

11

TIME IN FLOW (days)FIG. 17. HTR and biofouling levels in aluminum (type 5052) loop 11. Tubes were cleaned by chlorination at 0.10 mg liter-' for 1 h day-'

after HTR reached about Rf 9. Only part of the experiment is shown.

0

3

E

U0

LOlC

i-1n4-cyaS

VOL. 51, 1986

0

on October 13, 2017 by guest

http://aem.asm

.org/D

ownloaded from

1196 BERGER AND BERGER

components of the film constituted only -3% of the total filmweight. Only filamentous bacteria were observed in the filmsfrom the cold-water aluminum loops (Fig. 19).

DISCUSSIONStudies of film formation in model laboratory systems

have shown a direct correlation between nutrient concentra-tion in the bulk phase and both the rates of film formationand film thickness so long as nutrient concentrations are notexcessive (1, 8). At nutrient levels in which heterotrophicbacteria are commonly cultured in the laboratory, adhesionby bacteria to solid surfaces is generally weak.

In these studies, we noted that the fouling rates increasedduring the summer and autumn months when the surfacewaters approached 30°C and the dissolved OC ranged from0.76 to 1.14 versus 0.38 to 0.76 mg liter-' during winter andspring. The dissolved OC was consistently found to be-95% of the total OC.The experimental setup at the Seacoast Test Facility is

unique in that some experiments have been maintained todate for well over 3 years, using single-pass (once-through)flow systems.The types of countermeasures which can be tested in these

systems is more limited than in recirculating laboratory,pilot, or industrial reactors. Surfactants and high concentra-tions of other reagents may not be used for economic orenvironmental reasons. Chlorination is the method mostused in industry for biofouling control. These studies pro-vide a base line for the chlorination of seawater systems,using single-pass flow. Because of the huge surface areas andvolumes and high flow velocities in OTEC HXs, counter-measures must be optimized. We have used low concentra-tions of chlorine, usually 0.05 to 0.10 mg liter-', because ofthe low organic content of the seawater used.For analysis, chlorine samples were taken downstream of

the chlorinator either about 0.3 or 5 s after its generation. Nomeasure of the actual chlorine consumed in those intervals isavailable as the water passed through the HTM or the entireloop. All chlorination treatments resulted in readily mea-sured decreases of the biofouling rates. Use of 0.05 mgliter-' for 1 h day-' was inadequate as a fouling counter-measure. The minimum chlorine concentration which mightbe used under the field conditions tested is 0.1 mg liter-1 for1 h day-'. Continuous chlorination at the same concentra-tion was not adequate to clean prefouled tubes.

Cleaning by chlorination led to changes in the smoothnessof the biofilms. These, in turn, affected the HTR measure-ments, which led to the belief that the tubes were clean. Infact, the biofouling layers caused increased turbulencewhich lowered HTR below that expected from the otheranalytical data. The same observation was made by Tostetonet al. (7) in reference to mechanical cleaning regimes and invery early fouling (6).For OTEC-type model installations, Nosetani et al. (5) and

Tipton (D. G. Tipton, Argonne National Laboratory Rep.ANL/OTEC-BCM-018, Argonne, Ill., part 2, p. 1-30, 1981)showed that abrasive-coated sponge rubber balls could beused effectively in titanium tubing to remove fouling layerswithout appreciably eroding the metal. The same type ofcountermeasure caused rapid wear of aluminum and copperalloys.Use of the plain sponge rubber balls retarded but did not

control fouling film formation. They only removed the looseradhering surface layers. The resulting biofilm underlayerswere dense and recalcitrant to removal in situ and in thelaboratory.

TABLE 4. Biofouling parameters during successive fouling andcleaning cycles in loop 11

Cycle (105 °C m2W)Rf DW OC + TNCycle 101 oCm2 W'1) (p.g cm-2) (ILg CM-2)Free fouling2nd 5.98 217.1 22.893rd 5.33 372.9 54.76

Chlorinationa2nd 1.85 119.8 15.483rd 1.83 172.6 20.93a Chlorination was limited to 1 h day-' at 0.10 mg liter-'.

We have noted seasonal and unexplained changes in thetypes of organisms which populate the biofilms. While adiverse range of microorganisms has been noted in thefouling layers, there is no visual evidence that any eucary-otic organisms colonize the loops. Entrapment of diatoms,whole or fragmentary, and portions of differentiated algaeand other organisms have been noted in SEM preparations.The absence of light and the high velocity of the water in thetubing apparently prevents their colonization. Direct mea-surement for constituents specific to eucaryotes orprocaryotes and other parameters of community structureby the techniques developed primarily by Nickels et al. (4)and by Uhlinger and White (9) and Berger et al. (L. R.Berger, W. F. McCoy, and J. A. Berger, in Proceedings ofthe OTEC Biofouling, Corrosion, and Materials Workshop,Argonne National Laboratory Publ. OTEC/BCM-002,Argonne, Ill., p. 38-55, 1979) has not been done.

In these experiments, however, surfaces were cleanedwhenever the HTR rose above Rf 9. For nonturbulent flowthis HTR was reached when the wet film thickness wasabout 50 to 60 ,um. That some coupons were retrieved afterseveral years in the warm-water loops having biofilms 20 to40 times thicker demonstrates the need to use more thanHTR alone to monitor the course of fouling in this type oftubular HX system.The OC/TN weight ratios in the fouling layers were not

significantly different on the various metal surfaces and didnot vary significantly (P = 0.001) between thick and thinfilms except on aluminum alloys.

In loop 5 the OC/TN ratios before brushing averaged 3.90± 0.23 and 5.55 ± 1.85 after brushing. This suggests thatmost cellular material is removed from the hydrated oxidelayer by brushing, leaving behind much of the base slimelayer on or in the remaining underlying corrosion layer.

In aerated seawater, aluminum oxidizes to form a hy-drated aluminum oxide layer. The corrosion process goesrapidly at first but slows as the diffusion of oxygen becomeslimited by the corrosion film itself (B. E. Liebert, in Pro-ceedings of the OTEC Biofouling, Corrosion and MaterialsWorkshop, Argonne National Laboratory Publ.OTEC/BCM-002, Argonne, Ill., p. 189-196, 1979). Indepen-dent of this process, biofouling proceeded much as it did ontitanium and stainless-steel coupons in our experiments. Therate of early biofouling was shown to occur at approximatelythe same rate on all of these alloys whether or not corrosionalso occurred [B. E. Liebert, L. R. Berger, H. J. White, J.Moore, W. E. McCoy, J. A. Berger, and J. Larsen-Basse, inG. L. Dugger, ed., Proceedings of the Sixth Ocean ThermalEnergy Conversion (OTEC) Biofouling and Corrosion Sym-posium, 2:381-390, 1980).We observed that the biofilms accreted faster on aged

APPL. ENVIRON. MICROBIOL.

on October 13, 2017 by guest

http://aem.asm

.org/D

ownloaded from

COUNTERMEASURES TO MICROBIOFOULING 1197

14 H

12 F

10-

8k

4k

2

00L0

LOOP 17, Aluminum 5052 0Cold deep-sea water

0 0

0 0

_I

I I I

100 200 300 400 500 600 701

600 cC\EUINOn

500 D3

z

400 +u

0

-

300 4co

200 u6-4

w

100 wcm

-J

0 6-

LL

0

TIME IN FLOW (days)FIG. 18. HTR and biofouling levels in cold water, aluminum (type 5052) loop 17. Two sets of biocoupons were installed in the flow about

6 months apart. Some of the samplings were from the second set.

aluminum alloys than on titanium. Unlike titanium, how-ever, when a new piece of aluminum was placed in theflowing seawater, the HTR increased from Rf 1.7 to 3.5 as

the early corrosion layer formed. Biofouling on top of thislayer is responsible for most of the subsequent increase inHTR. A loose outer film of bacteria and organic matterextends into the flow and increases the width of the stagnantwater layer; it is such a layer that relatively gentle brushingremoves.

Studies in electrical impedance spectroscopy (P. K.Sullivan, Ph.D. thesis, University of Hawaii at Manoa,Manoa, 1985) recently showed that a fouled titanium surfacecan be resolved into three layers: (i) the loose outer layermentioned above, (ii) a tougher older film which graduallydevelops under it and which is believed to contain mostlyinorganic material accounting for much of the film DW, and(iii) the thin TiO2 protective barrier covering the bare metal.

FIG. 19. Biofouling of aluminum (type 5052), using deep cold

ocean water for about 20 months. Mostly filaments and bacillaryforms are evident adhering to and imbedded in the hydrated alumi-num oxide-exopolysaccharide matrix. Bar = 2 im.

Biofouling was greatly retarded in the cold-water loops.With aluminum (3003) clad with aluminum, corrosion was so

extensive that it was not possible to follow biofilm forma-tion. Essentially no fouling was observed on the stainless-steel loop over the test period. However, biofouling didoccur on aluminum 5052, although it was very slow. In termsof a closed-cycle OTEC plant, some sort of fouling counter-measure would be needed for the cold-water HXs, but its usewould be infrequent.The deep-sea cold water used in these studies d'ffers

markedly from that used by Lewis (R. 0. Lewis, in Proceed-ings ofthe Eighth Ocean Energy Conference, Mar. Technol.Soc., 2:379-387, 1981), who compared fouling rates inWrightsville Beach, N.C., waters between summer andwinter months. While the fouling rates in the winter waterwere lower than those in summer, the differences were

relatively small; extensive fouling occurred year-round.In contrast, we have compared two different water masses

simultaneously over several years. The deep seawater islower in dissolved organic matter and contains fewerheterotrophic bacteria than are found in the warm surfacewaters. In 2 years of examining aluminum loops in the coldseawater, we have only observed filamentous bacteria on thethin hydrated corrosion layers.

ACKNOWLEDGMENTSSupport for these studies came from the Ocean Energy Program of

the U.S. Department of Energy through Argonne NationalLaboratory (ANL), Argonne, Ill.

We acknowledge Tom Daniels, Director of the Natural EnergyLaboratory of Hawaii (NELH), the staff of the NELH, and AjayBhargava for the measurements of HTR, C. Panchal and H. Stevensof ANL, and J. Larsen-Basse, Scientific Director of the project, forhis advice, aid, and support throughout this work. We thankespecially Barbara Lee for her meticulous attention to the manytasks which she performed for us at the test site, and we acknowl-edge Ted Walsch for the water quality data cited in this paper.

LITERATURE CITED

1. Characklis, W. G., M. G. Trulear, J. D. Bryers, and N. Zelver1982. Dynamics of biofilm processes: methods. Water Res.16:1207-1216.

11

e%

E

0

LO

C0U-.

VOL. 51, 1986

6

on October 13, 2017 by guest

http://aem.asm

.org/D

ownloaded from

1198 BERGER AND BERGER

2. Corpe, W. A. 1980. Microbial surface components involved inadsorption of microorganisms onto surfaces, p. 105-144. In G.Bitton and K. C. Marshall (ed.), Adsorption of microorganismsto surfaces. John Wiley & Sons, Inc., New York.

3. Costerton, J. W., G. G. Geesey, and K. J. Cheng. 1978. Howbacteria stick. Sci. Am. 238:86-95.

4. Nickels, J. S., R. J. Bobbie, D. F. Lott, R. F. Martz, P. H. Benson,and D. C. White. 1981. Effect of manual brush cleaning onbiomass and community structure of microfouling film formed onaluminum and titanium surfaces exposed to rapidly flowingseawater. Appl. Environ. Microbiol. 41:1442-1453.

5. Nosetani, T., S. Sato, K. Onda, J. Kashiwada, and K. Kawaguchi.1981. Effect of marine biofouling on the heat transfer perform-ance of titanium condenser tubes, p. 345-353. In E. F. C.Somerscales and J. G. Knudsen (ed.), Fouling of heat transferequipment. Hemisphere Publishing, New York.

APPL. ENVIRON. MICROBIOL.

6. Sasscer, D. S., T. 0. Morgan, C. Rivera, T. R. Tosteton, B. R.Zaidi, R. Revuelta, S. H. Imam, R. W. Axtmayer, D. DeVore, andD. L. Ballantine. 1981. OTEC biofouling, corrosion and materialsstudy from a moored platform at Punta Tuna, Puerto Rico. I.Fouling resistance. Ocean Sci. Eng. 6:499-532.

7. Tosteton, T. R., B. R. Zaidi, R. Revuelta, S. H. Imam, R. W.Axtmayer, D. DeVore, D. L. Balantine, D. S. Sasscer, T. 0.Morgan, and D. Rivera. 1982. OTEC biofouling, corrosion andmaterials study from a moored platform at Punta Tuna, PuertoRico. II. "Microbiofouling." Ocean Sci. Res. Eng. 7:21-73.

8. Trulear, M. G., and W. G. Characklis. 1982. Dynamics of biofilmprocesses. J. Water Pollut. Control Fed. 54:1288-1301.

9. Uhlinger, D. J., and D. C. White. 1983. Relationship betweenphysiological status and formation of extracellular polysaccha-ride glycocalyx in Pseudomonas atlantica. Appl. Environ. Mi-crobiol. 45:64-70.

on October 13, 2017 by guest

http://aem.asm

.org/D

ownloaded from