Haat Transfer Engineering, 26(1]:44-50. 2005Copyrigbt c Taylor fi Francis Inc.ISSN: 0145-763S pnnt / 1521-0537 onlineDOI: 10.1080/01457630590890166

Taylor St Francis

Biocide Dosing Strategies for BiofilmControl

D. M. GRANT and T. R. BOTTSchool of Engineering, University of Birmingham, Birmingham. U.K.

In order to reduce the em'inmmental impact of biocide for the contnti of hiofilm formation in cooling water citvuits,"environmentally friendly" hiocides have been developed: lunvevcr. they are generally tnore e.xpen.m'e than traditionalchemicals. It is imperative, thetefore. that the minimum quantity ofbiocidv is employed so that costs are kept to a tninimum.To achieve this objective, optimum dosing strategies are required. Using a pilot plant in conjunction with a monoculture ofPseudomonas fluorecsens as the biofouling bacterium, tests were carried out imng a proprietary biocide to invc.stifiate theeffects ofdo.\e concentration, duration, and frequency of dosing and fluid mechanics im bittfilm control. With fourfifwen-mimaeapplications per day at a peak concentration of 16.8 mR/l. it was not possible to Inhibit hiofitm development. Control waseffected, however, by doubling the peak concentration usinf^ a short dosing period. As would be expected, concentration wasshown to be a critical factor for cotitroL A boicide concentration below that for growth inhibition .seemed to enhance biofilmformation. Increasing the frequency of dosing is only effective if the concentration employed is biofihn growth-inhibiting.

Because of the well-known eiTecls of heat exchanger foul-ing, it is necessary to control biotilm formation in coolinj; wa-ter circuits. The usual approach is lo apply a biocide lo kill themicro-organisms. For many years, the preferred biocide ha.s beenchlorine because of its effectiveness and relatively low cosl. Be-cause the water in many cooling water circuits is laken fromnatural sources, such as lakes or rivers, and discharged back tothe source, the presence ofthe biocide represents an environmen-tal hazard. Chlorine, for instance, reacts with organic matter toproduce cancer-forming substances, and the chlorine can enterihe ftHid chain.

In order to overcome these potential problems for the environ-ment, "environmentally friendly" biocides have been developed,but their cosl is high. Il is therefore imperative that techniquesto maximize their effectiveness are employed.

This paper reports work using a proprietary biocide to illus-trate how optimum biocide dosing strategies may be developed.

MATERIALS AND METHODS

The laboratory pilol plaiii used in the work is based on a"feed and bleed" concept and operates for long periods of time.Figure I is a simple schematic sketch ofthe equipment.

Address correspondence to Dr. T. R. Bon. School of Engineering. Universityof Birmingham. Birmingham BIS 2TT. U.K. E-mail: bolltr@ bham.ac.uk

A monoculture of Fsettclimumas fiuore.scens is grown in afennenier. and a predetermined flow of water laden with micro-organisms is pumped into the mi.\ing vessel. The Fseiulonumasspecies is known to produce biolilnis in aqueous systems, andin these experiments, the species is used as a representative con-taminant found in cooling water circuits. The concentration inthe circulating waler from the mixing vessel is lO" cells /ml. Aknown rate of nutrient addition and filtered tap water, from whichtraces of chlorine have been removed by an activated carbon fil-ter, is also fed to the mixing vessel. The tiller system removesall particulate matter down to 1 fx. so il may be assumed allpotential contaminates, including micro-organisms, have beenremoved. Tests with no added micro-organisms demonstratedthat no biofilms were formed, contirming that the filter systemwas effective. The nutrient solution is sterili/.ed in an autoclaveal 120 C for one hour before being fed to the mixing vessel.Automatic alkali addition in response to a pH probe maintainsthe pH at 7. Pseudonumas fiuore.scens is aerobic, so the mixingvessel is sparged with tillered air to ensure that the circulatingwater is saturated with oxygen. The simulated cooling water ispassed through vertically mounted glass tubes, which act as thelest surfaces. The tubes are Im in length with internal diametersof 9 anti 23.1 mm.

It is probably true to say that the water in every cooling watercircuit is unique, as each system will draw on water availablein the vicinity of the plant. The different sources will displaya considerable range of composition depending on the local

44

BIOCIDE DOSING STRATEGIES FOR BIOFILM CONTROL

Filtered Tap Water (,2 ^ m niter

45

Dniin

I'igurc 1

Pump

Flow.sheet ofthe biofouling pilot plant.

ecology. The simulated cooling water in the pilot plant, madeup as described, is meant to represent a eooling water of a stan-dard composition that is suitable for testing the effects of dif-ferent treatment strategies. The principal difference between itand a "real" cooling water is that it only contains one organism.whereas a natural source of water will contain a whole spectrumof species. The point of using a monoculture is that the waterwill be of consistent quality in all respects so that comparisonsof different dosing strategies will he valid.

In the experiments reported in this paper, two water ve-locities were chosen to represent industrial water flows of 0.5and 1.3 m/s. The Reynolds number was around 12.000 be-cause it has been shown that at this value, the biolilm growthrate is a maximum 11]. The water temperature was approxi-mately 2O'C. A proprietary hiocide containing 2.2 Dibromo-3-nitrilopropionamide was used in the experiments at differ-ent concentrations with different dosing regimes to determinetheir effect on hiofilm control. The residence time in the pilotplant was sixty minutes; variable residence times could affectthe quality of the simulated cooling water. The use of a fixedresidence time, therefore, is to ensure that a consistent qualityof simulated cooling water is maintained throughout Ihe testsso that reliable comparisons of the dosing strategies may bemade.

The pilot plant is equipped with a non-intiuslve infrared mon-itor [2j. The ahsorbance of the infrared radiation by the biofilmre.siding on the surface of the glass test section is a measure ofhiofilm accumulation at the point of application of the infraredprobe. A comparison of the absorbance recorded under differ-ent operating conditions, biocide concentration and applicationstrategy, provides a means of assessing the effectiveness of theparticular technique.

In strategies where the dosing is not continuous and the ap-plication of biocide is over a relatively short time (e.g., fifteenminutes), the concentration of the biocide rises to a peak valueand then falls away to a level approaching 0 mg/l. A model ofbiocide concentration variation with time has heen presented

t3|. A typical biocide concentration/time profile is presented inFigure 2.

RESULTS

Continuous Dosing

Continuous dosing of the biocide at concentrations of HK)and 50 mg/l were found to control biofilm formation (see Fig-ure 3). A reduction of biocide concentration to 20 mg/l showedthat there was some biofilm development at this concentrationdetected by the infrared monitor: nevertheless, there seemedsome inhibition of growth at this biocide concentration as thebiofilm could not be seen with the naked eye. A further reduc-tion of biocide concentration to 15 mg/l resulted in an increasedrale of biofilm growth, particularly noticeable at the lower watervelocity of 0.5 m/s. The growth rate of the biofilm at this ve-locity was similar to the control (no biocide present), indicating

120

100 200

Time, minutesFigure 2 BltKide conceiiiraiion protile.

300

heattransfer engineering vol.26 no.l 2005

46 D. M. GRANT AND T R. BOTT

0-4

02

0.0

Conirol l.3m/s

(\ininil O.J m/i

SO ppm

20'pfm

0 5 10 15

Time, days

Figure 3 Effect of continuous dosing of biocide from a slaning conceniraiiunof 100 mg/l.

that the biocide was exerting very little effect. At the higherwater velocity of 1.3 m/s, biofilm growth appeared to be undercontrol. Work with a similar biocide demonstrated that eontinu-ously dosing at a concentration of 15 mg/I. effectively preventedbiofilm formation [4). Table 1 provides data on the total biocideused daily in the continuous dosing experiments at the differentconcentrations employed. If these concentrations were used ona full scale cooling water system, they would represent largequantities and associated costs.

At Ihe higher water velocity of 1.3 m/s., the infrared ab-sorbance of the biofilm wa.s lower than that observed with awater velocity of 0.5 m/s. It ha.s been demonstrated that at lowvelocities, ma.s.s tran.sfer plays an important role in biofilm devel-opment; growth increases with increasing mass transfer of nutri-ents as velocityis raised up to about 1 m/s [5]. At velocities above1 m/s. the effect of the increased shear on the biofilm is greaterthan the effect ofthe associated increased mass transfer of nutri-ents. As a result, the thickness ofthe biofiim is reduced. Theseresults contimi the findings of previously published work [6|.

Shock Treatment

Shock treatment involved subjecting the system to a relativelyshort burst of biocide over a period of fifteen minutes.

The first tests used a concentration of 100 mg/l of hiocide,as this concentration was found to he effective on a continuousbasis, but the results indicated that shock dosing at this con-centration did not prevent biofilm growtb (Figure 41. After twodays, biohlni could be observed un the test surfaces al the lowerwaler velocity of 0.5 m/s. For the first six days, biofilm growth

Table I IVital daily use of biocide at different continuously dosed

concentraiions

Dose concenlralion mg/lDuilv total mn

KM)

28.aoo30

!4.40t)20

.5.760

4 6

Time. da\s

Figure 4 Eflect of daily shock dosing UK) mg/l biocide.

was similar to the control (no biocide present), that is. duringthe lag phase and the beginning of the exponential phase. De-velopment began after two days and five days at water velocitiesof 0.5 and 1.3 m/s, respectively. Beyond the lag phase, there didappear to be some inhibition of biofilm growth, although it hasto be reported that in ihe parts of the pilot plant (e.g., the mixingvessel and pipe work, where the velocity was particularly low),significant biofilm could he seen.

The conclusion of this experiment was that a shock dose of100 tng/l of biocide was not an effective ireatmeni for biofilmcontrol at the water velocities tested, although tfiere was an in-dication of inhihition of bioHlm growth.

Since this level of shock treatment was not effective, it wasdecided to go to the other extreme and give a massive shockdose, based on the tolal amount of biocide used per 24 hoursfor continuous treatment, as presented in Table 1. On this basis,28800 mg of biocide was pumped into the system in a periodof fifteen minutes, resulting in a peak biocide concentration of2133 mg/l. The results are given in Figure 5, demonstrating that

08

06

04

02

00

O 0.5 m/sA Comiul r.im/sD Control U.5 m's

• • i

heattransfer engineering

0 2 4 6 6 10

Time, days

Figure 5 Effect of daily shock dosing a total of 28,800 mgolbiociiJe.

vol. 26 no. 1 2005

BIOCIDE DOSING STRATEGIES FOR BIOFILM CONTROL 47

10

O.B

0.6

0.4

0.2

00

• 1.5 m; sO (1,5 m/s

A Diinitil l.!l mi'i

O Conlrol D.S nVs

• C l -

0 82 4 6Time, days

Figure 6 Etfecl of pulse do.sing biocide 4 x 15 minute.s/day al a peak con-centration of 106 mg/l.

this shock treatment was a failure. As can be seen from Figure 5,biofilm could be observed on the fourtb day after commencingthe biocide treatment. A sudden growth of thick biofilm wasobserved on the lest surfaces at bolh water flow velocities (0.5and 1.3 m/s). There was no difference in the time taken forthe onset of biotilm growth. In contrast to normal experience,biofilm growlh al the higher velocity regime was grealer thanat a water velocity of 0.5 m/s. A possible explanation of thisobservation is that following the "wash out" of Ihe biocide fromthe system, the higher mass transfer afforded by the higher watervelocity provided micro-organisms and nutrients at a greater ratethan al the lower velocity. The experiment was nol repealed.

The general conclusion of these experiments Is that shortshock treatmeni times on a relatively infrequent basis, even withbigh biocide concenlrations. is not effective in the control ofbiofilm formation.

Pulse Dosing

In effect, pulse dosing may be taken as shock treatment on amore frequent appHcaiion. Since a continuous dose of 20 mg/Iofbiocide appeared lo restrict biofilm growth, it was decided touse the total biocide used over a 24-hour period and dose thisquantity in four fifteen-minute pulses evenly spaced throughout24 hours. Tbe total dose in 24 hours given in Table 1 is 5760mg, whicb gave a peak biocide concentration of 106.19 mg/l.As can be seen from Figure 6. this pulse dosing completelyprevented biofilm growlh. and the monitor indicated almost zeroaccumulation of biofilm on the test surfaces.

Treatment with 5760 nig/day in eight fifteen-minute dosesgave a peak concentration of 53.1 mg/l for eacb dose. Tbe dos-ing regime, as can be seen from Figure 7, prevented bioHimgrowlh at bolh velocities tested. As wilh ihe previous experi-ment, the infrared absorbance monitor recorded zero at the lestsurfaces. Biotilm deposits were seen in other parts of tbe equip-

1.0

0.8

H 0.6

II 0.4

0.2

0.0

•oA

O

l.'m.s0.5 m/s

Coniml 1.3 m/sConlrnI 0.5 m/s

r

p//

If

JA 6

Time, davs

10

Figure 7 iiffeci of pulse dosing biocide 8 x 1 5 minutes/day ai a peak con-centration of 53.1 mg/l.

ment, wbere Ihe water velocity was low. It could be staled thatbolh these pulsed dosing strategies were successful and, in com-parison wilh the continuous dosing treatmeni wilh the 20 mg/lbiocide concentration, yield better control.

Taylor [7], using Ibe same biocide but with 4 x 30-minutedoses in 24 hours at a peak concentration of 15 mg/I, demon-strated that biotilm formation was prevented at a waler veloci-ties of 1.27 and 0.86 m/s (i.e., similar to the present study witha Reynolds number of 14000). !t was calculated that under tbistreatmeni, the lolal biocide consumption would be 914.9 mg/day,and this figure was used as the basis for treatment using fifteen-minuie pulses. The peak biocide concentrations associated withthe three regimes are given in Table 2.

Figure 8 shows the results of using eight fifteen-minute dosesin 24 hours, wilh the peak concentration of 8.43 mg/l. Therewas a measure of control for about five days after which biofilmdevelopment began. It would appear tbat this treatment regimesimply extended the lag phase. The exponential phase startedalter the sixth day al both waler velocities, and by the eighthday, biofilm was greater at tbe higher water velocity of 1.3 m/s.

The results from a treatment regime of four fifteen-minutedoses every 24 hours are shown on Figure 9. The growth wasaccelerated for both water velocities compared to the control (nobiocide present), tbe lag phase being reduced by approximatelyhalf.

The effect of halving the dosing period from thirty to fifteenminutes, even though the total amount ofbiocide was the same,seemed lo stimulate growlh al Ihe water velocilies studied, to asurprisingly greater effect at the higher velocity.

Tiible 2 Biocide peak concentration equivalent to 914.9 mg/day

Dicing regimePeak cone, mg/l 8.43

4 x 1 5 mill16.8

2 x 1 5 min33,6

heattransfer engineering vol.26 no . l 2005

48

l..lms0.5 m/s

0.30

0.25

0.20

0.15

0.10

0.05

0.00

0 2 4 6 8 10

Time, days

Figure 8 Effect of pulse dosing biwide 8 x 1 5 minutes/day al a peak con-

centration of 8.4;i mg/l.

The results ofthe last test with two fifteen-minute treatmentsin 24 hours with a peak biocide concentration of 33.6 tng/l,was not successful in the prevention of biofilm formation, asshown on Figure 10. Biofilm was detected after two days, andgrowth remained in the lug phase for a further five days beforethe exponential pha.se was initiated. The extended lag phase maybe noted for each ofthe water velocities. On the eleventh day ofthe experiment, biolilm accumulation was again greater in thetesl section, with water flowing at ! .3 m/s.

In the light of these re.'iults, based on the use of 914.9 mgbiocide per day where no control was achieved, it was decidedto repeat two of the tests, but using double the daily consumption(i.e., 1829.8 mg). Because of the particularly poor result of usingonly two pulses per day in the earlier experiments, the testswere based on four and eight pulses each 24 hours. The peakconcentrations are given in Table 3.

D. M. GRANT AND T, R, BOTT

0.4

10

OB

0.6

< 0.4

0 2

• 1.3 m/sO 0.5 m/sA Conlrol 1,3 m'sD Control 0.5 m/s

0.00 2 4 6 8 10

Time, days

Figure 9 Effect of pulse dosing biocide 4 x 1 5 minutes/day ai a peak con-

centration of 16.S mg/l.

0.00 2 4 6 8 10 12

Time, days

Figure tO Effect of pulse do.sing biocide 2 x 1 5 minutes/day at <i peak con-

centration of 33.6 mg/l.

Figure 11 provides the results of the test for eight fifteen-minute pulses ofbiocide at a peak concentration of 16.8 mg/l,demonstrating that bioflim growth is controlled under this re-gime. The itifrared monitor recorded near zero absorbance forboth velocities.

The results for four fifteen-minute pulses ofbiocide at a peakconcentration of 33.6 mg/l are given on Figure 12. Growth wasnot visible to the naked eye. but the monitor detected a small ac-cumulation of biofilm. Fora period of fourteen days, this biofilmgrowth control was not so effective under this regime when com-pared to the eight fifteen-minute pulse regime of Figure 8 withthe higher peak biocide concentration.

DISCUSSION

The results ofthe experiments suggest that dosing an in,suf-ficient amount of biocide may cause an increase in the biofilmaccumulation. The studies showed greater biofihn growth in thepresence of the biocide at a higher water velocity. The enhancedgrowth only seemed to occur in the exponential phase of biofilmgrowth.

It is difficult to appreciate why growth is stimulated in thepresence of the biocide. Extracellular polymer material thatforms the "substance" of biofilms can act as an exchange resin. Itmay adsorb biocide. thereby preventing the biocide from reach-ing the actual cells in the bioHIm [8]. At the higher water velocityof 1.3 m/s, it is likely that relatively large amounts of biocidereach the biofilm during the lag phase, compared with the situ-ation at the lower water velocity of 0.5 m/s, due to the greatermass transfer at the higher turbulence level associated with the

Table .1 BitK'ide peak concentration equivalent to i 830 mg per day

consumplion

Dtising regime

Peak tone. mi;/lX X 15 min16.H

4 x 1 5 min

3.1, fi

heat transfer engineering vol.26 no.l 2005

BIOCIDE DOSrNG STRATEGIES FOR BIOHLM CONTROL 49

• 1.5 m/s

O 0,5 m/s

A Coiilml l..1m/s

D ( oniml 0,S m/a

04

03

02

0.1

000 5 10 15

Tinne, days

Figure II EITectof pulse dosing biocide S x IS minutes/day al a peak con-centraiion of 16.8 mg/l.

higher water velocity. Under the conditions of the experimentsthere may have been insufficient biocide at the water/biofilminterface to kill the surface coloni/xrs. Exposure to the biocidemay have stimulated a degree of resistance to the biocide. pro-moting the production of extracellular polymer that would mopup the biocide as it reached the biofilm. At the lower water ve-locity {0.5 m/s), the mass transfer of the hiocide would be lessso thai the opportunity for the development ofbiocide resistancewould be reduced.

Many investigators and plant operators have noted a rapidresumption of biofouling following biocide application, partic-ularly in respect of chlorine addition | 3 | . Pujo [4] also observedthis phenomenon, which she termed "regrowth" when the ap-plication of chemical biocides ceased. This may aiso be partlyresponsible for the rapid regrowth seen in this study. Characklis[3] suggested that regrowth may be due to the following;

04

0 3

-e 0.2

0 1

0,0

• I 3 III ^

O 0,5 m\A Coninil l.3nt/sD Cunlnil 0.3 m/s

0 5 10 15

Time, days

Figure 12 Effect of pulse dosing biocide 4 x 15 minutes/day at a peiOc con-centration of 33,6 mg/l.

1. The remaining biofilm contains enough viable cells to pre-clude any lag phase in biofilm accumulation (as observed onclean surfaces)

2. The remaining biofilm imparts a relative roughness to thesurface, increasing convective transport and attachment ofcells and debris to the surface

3. Biocide may preferentially remove the extracellular materialand not the microbial celis, thus exposing the cells to thenutrients in the system on dosing cessation

4. The surviving cells rapidly produce extracellular material asa protective to the biocide

CONCLUDING REMARKS

From this work, it is clear that careful attention to the strategyofbiocide application is important. Not only does the techniqueof treatment affect the extent ofthe control ofthe biofilm growth,it also has a direct bearing on the cost of the treatment. Plant trialsare likely to provide the best opportunity to obtain an optimumdosing regime. It has to be stated, however, that prior to suchan investigation, it would be necessary to carry out preliminarywork in a pilot plant in order to reduce the risk of difficultiesoccurring during plant trials, which in themselves could prove tobe expensive. An alternative to pilot plant studies would be to usea side stream from the full scale plant itself. Frequent reappraisalofthe .strategy would be necessary to counteract changes in theecology of the cooling water facility that could develop over aperiod of time.

REFERENCES

111 Pujo, M., and Bou, T, R., EtTects of Fluid Velocities and ReynoldsNumbers on Biofilm Development in Water Systems, in Kxperi-mental Heat Transfer, Fluid Mechanics and Thermodynamics, eds.J. F, Keffer. R. K. Shah, and E. W, Ganic. pp. 1358-1362. Elsevier,1991.

|21 Tinham. R, and Bott, T. R., Biofouling Assessment Using an In-frared Monitor. International Specialised Conference on BiofilmMonitoring, pp. 53-56, Porto. Portugal. 2O()2.

13| Characklis. W. C , Microbial Fouling, in Biofilms, eds. W. C.Cbaracklis and K. C. Marshall, pp. 586-634, John Wiley and SonsInc, 1990.

|4 | Pujo. M., Effects of Hydrodynamic Conditions and Biocides onBiofilm Control. Ph,D, Thesis. University of Birmingham. U.K.,1993.

[51 Bott, T. R. Fouling of Heat Exchangers, pp. 242-243. Elsevier,1995.

|6] Bott, T. R.. and Taylor. R. J., The Effects of Velocity on BiocideUse for Biofilm Removal in Flowing Systems, in Heat Transfer—Baltimore, ed. M. S, Et Genk. pp. 322-326. 1997,

|7 | Taylor, R. J. Efficacy of Industrial Biocides Against BacterialBiofilms. PhD. Thesis, University of Birmingham. U,K.. 1995.

|H] Le Chevallier, M. W., et a!.. Inactivation Biofilm of Bacteria, App.and Enviro. MicrobioL, vol. 54, pp. 2492-2499, 1988.

heattransfer engineering vol. 26 no.l 2005

50 D. M. GRANT AND T R. BOTT

Dcon M. Grant iiaineit her Bachelor's degree in biu-clicmisirj. She went un lo obliiin a Master's degree intiiiKhcmical engineering. Her ductorate involved re-sc;irth into biivide activity liir ihe control of biofoul-ing in cuoling water .systems. Her work subsequentlyhas been in the waier ireauiient industry both in theU,K. and the U.S.

T. Reg. Ball has been involved with hcai exchangerIbuling prohlciiis for thirty years and is eon.sidereU mbe a world authority in the field. He was prcsenledwith the Brennan medal in 1990 hy the Institution a(Chemical Engineers tor his book Fouling Notebookiuid the D()n;Ud 0, Kem Award hy the .American Insti-tute of Chemical Engineers in 2(HH) tor his eontribu-lion to fundamental understanding of heat exchangerfouling. In 2(H)3. he was made a Member ofthe BritishEmpire by Her Majesty Queen Bli/abeth for his con-tribution to energy conservation.

heattransfer engineering vol.26 no.l 2005