Biodegradation of artificial monolayers applied to water storages to reduce evaporative loss

The application of artificial surface films to retard evaporative loss was first developed as a technology for water conservation in the 1950s and 1960s (Barnes 2008). These amphiphilic molecules (containing a hydrophobic chain and a hydrophilic head) must spontaneously spread over the water to produce a monomolecular film with a surface pressure in excess of 15 mN m⁻¹ (a condensed monolayer), to effectively retard evaporation. The technology was not widely adopted due to highly variable performance in field trials, attributed mainly to wind disruption. Other natural processes operating at the air–water interface (the microlayer, Norkrans 1980), including photodegradation have not been considered. The synergy between microbial degradation and photodegradation is responsible for ‘cleansing’ anthropogenic compounds from the microlayer (Brinkmann et al. 2003).

The microlayer is an ecologically distinct region concentrating a diverse range of amphiphilic organic compounds structurally similar to long-chain fatty alcohol monolayer compounds (Brinkmann et al. 2003). Natural microlayer films do not retard evaporative loss, as the presence of double bonds, side-branches, and aromatic rings impede the formation of a condensed monolayer (Pogorzelski & Kogut 2001). The saturated, aliphatic chains of the fatty alcohols hexadecanol (C16OH) and octadecanol (C18OH), and the glycol ether (C18E1) monolayer compounds form densely packed films, inducing a surface pressure well in excess of 15 mN m⁻¹ (Barnes 2008).

Photodegradation is the abiotic fragmentation of molecules absorbing photons of light energy (Vahatalo 2009). In freshwater ecosystems, the most common absorber molecules are aromatic organic compounds containing unsaturated bonds that absorb ultraviolet light (chromophores; Sulzberger & Durisch-Kaiser 2009). Direct photodegradation occurs when dissolved organic matter is photochemically transformed after absorbing ultraviolet radiation. Indirect photodegradation occurs when molecules that lack chromophores are transformed by reactions with photochemically generated, chemically reactive species (Vahatalo 2009). The compounds C16, C18 and C18E1 do not contain chromophores, but may be susceptible to indirect photodegradation.

In our study, the susceptibility of the fatty alcohols C16OH and C18OH (chain length of 16 and 18 carbons respectively), and a glycol monoalkyl ether (C18E1) monolayer compound to photo- and microbial degradation were compared. Photodegradation assays were conducted on brown and clear water samples, exposed to cloud-free, subtropical sunlight. The bacterium Acinetobacter was isolated for the microbial degradation bioassays, as it assimilates a range of organic carbon substrates (Uthoff et al. 2005) and is a common inhabitant of polluted water (Lemke & Leff 1999). Phenol was used as the sole carbon source in selective media as phenols are often present in industrial wastewater, and Acinetobacter species are known to degrade them (Pittaway & van den Ancker 2010).

Microlayer and subsurface water sampled from a brown water lagoon (Narda Lagoon) and activated sludge from a municipal wastewater treatment plant were used to select phenol-degrading microbes (Pittaway & van den Ancker 2010). Pure cultures isolated by streaking inocula onto trypticase-soy agar were biochemically characterized using the Gram stain and light microscopic observation, and the API 20NE system (bioMerieux^®).

Pure cultures were re-inoculated into a mineral salts phenol (MP500) medium and incubated to an optical density of between 0.6 to 0.7 optical density units (ODUs; Gerhardt et al. 1994). Dilution plate counts using peptone water and trypticase-soy agar were incubated at 25 °C to quantify the viable cell population. MP500 cultures were used as inoculants to compare the susceptibility of the monolayer compounds to bacterial degradation.

Five millilitres of stock solutions of Fluka analytical grade C16OH, C18OH or C18E1 dissolved in acetone were added to 16 sterilised flasks and evaporated overnight on an orbital shaker, to coat the lower internal surface to a height of 1.5 cm. Twenty millilitres of MS broth (Bouchez et al. 1995) was added to produce a final monolayer concentration of 1 mM as the sole source of organic carbon for microbial growth. Twelve flasks were inoculated with 2 mL of an MP500 Acinetobacter medium and incubated at 25 °C on an orbital shaker at 120 rpm for 2, 3 and 4 days.

Monolayer compounds were recovered by adding 20 mL of chloroform to 4 inoculated, incubated and to 4 un-inoculated flasks, placed on magnetic stirrers for 20 minutes and poured into separation funnels. One microlitre of the chloroform phase was injected into a Shimadzu GCMS-QP 2010 Plus Gas Chromatography – Mass Spectrometer with an Rtx-5 column (30 m × 0.25 mm, 0.25 μm thickness, column oven initial temperature 130 °C, increasing to 200 °C at 20 °C/min, then to 300 °C at 10 °C/min over 13.5 minutes) using ultra-high purity helium (1.95 mL/min flow, linear velocity 52.6 cm/s) run with a split ratio of 1:50. The injection port temperature was 270 °C, and the detector 250 °C. A dilution series of 1.0, 0.5, 0.3, 0.2 and 0.1 mM of each monolayer compound dissolved in chloroform was used as a concentration standard. A two-way analysis of variance (SigmaPlot 12, Systat Software Inc.) was used to compare the susceptibility of the three monolayer compounds to microbial degradation.

The aqueous phase of the Narda Lagoon Acinetobacter cultures in the separation funnels was used to prepare a cell extract. The samples were centrifuged for 5 minutes at 3,000 rpm and ten mL of the supernatant was frozen prior to analysis. Five mL of the thawed sample was mixed for 2 minutes with 5 mL of a phenol:chloroform:isoamyl alcohol solution (25:24:1) and centrifuged at 3,000 rpm for 5 minutes (Ausubel et al. 1995). A 0.8 mL sample of the supernatant containing DNA and RNA was placed in a microcentrifuge tube and mixed for 2 minutes with 1.2 mL of isopropanol before centrifugation at 3,000 rpm for 30 seconds to pelletise the DNA. Two 1 mL aliquots of 80% ethanol were pipetted into the centrifuge tube to precipitate the DNA. The centrifuge tubes were inverted to allow the DNA pellet to air-dry for 30 minutes.

The DNA was re-suspended by adding 50 ml of Tris EDTA buffer to each tube (Ausubel et al. 1995). The tubes were mixed and left overnight in a refrigerator. The DNA suspension was vortexed, and centrifuged for 10 seconds at 3,000 rpm. Three μL of the purified DNA solution was placed into an Implen Nanophotometer cuvette and the absorbance of the sample was recorded using wavelengths of 230, 260 and 280 nm. The DNA concentration was calculated as 1.0 absorbance units at 260 nm equals 50 μg of double stranded DNA (Brown 1990). A two-way analysis of variance (SigmaPlot 12, Systat Software Inc.) was used to compare the ability of the Narda Lagoon Acinetobacter isolate to utilize the three monolayer compounds as a microbial substrate.

A Langmuir trough was used to measure the surface pressure of the three monolayer compounds applied to the surface of 700 mL of water at a rate up to five times the amount required to achieve a monomolecular layer (Herzig et al. 2011). The compounds C16OH and C18OH (Fluka) were urea-clatharated and recrystallized three times in hexane. Milli-Q water with 18.2 MΩ cm resistivity was used for the subphase in the clean water experiments.

Microlayer and subsurface water samples were collected from Narda Lagoon (Pittaway & van den Ancker 2010). The microlayer sample was applied drop-wise to the surface of the subsurface water in the Langmuir Trough using a pipette at a volume that approximated the surface to volume ratio of the in-situ microlayer. The initial surface pressure was measured using the plate detachment procedure and an evaporimeter (Herzig et al. 2011). Five milligrams of finely ground C16OH, C18OH or C18E1 monolayer solids was applied to the surface of Narda Lagoon or the Milli-Q water samples. After reaching equilibrium spreading pressure, the surface pressure and evaporation resistance was monitored over 3 to 5 days.

A piston pipette applied ethanol solutions of C16OH, C18OH or C18E1 (equivalent to three times the monolayer application rate; Barnes 2008), to the surface of 50 mL of distilled water or to filtered (0.45 μm glass fibre) Narda Lagoon microlayer water in plastic petri dishes (85 mm diameter, 13 mm depth). One series of petri dishes was irradiated with cloud-free sunlight (light). A second series was placed in a light-excluded box (dark), adjacent to the irradiated series. Water temperature and solar radiation intensity were measured using Type K thermocouples (RS-Components) and an Apogee SP-110 Pyranometer.

Seven petri dishes per treatment were destructively sampled every 20 minutes for the first hour, then hourly over eight hours. Five millilitres of hexane was added to the water to recover any residual monolayer compounds. One microlitre of the hexane phase was injected into a Shimadzu GCMS-QP 2010 Plus mass spectrometer as described above, to quantify the residual monolayer concentration. Linear regressions for the natural log transformed exponential decay curves were compared using slope ratio analysis (Finney 1978). Half-lives expressed as MJ m⁻² were calculated using t_1/2 MJ m⁻² = (t log 2)/log ([C₀]/[C_t]) where [C₀] is monolayer concentration at time 0 and [C_t] at time t. Equivalent half-lives after exposure to solar radiation on a sunny summer day (long-term average in Southeast Queensland 30 MJ m⁻², Bureau of Meteorology 2012) were calculated as t_1/2 hrs = (t_1/2 MJ m⁻²)/30 × 24 hours.

All phenol-degrading isolates were Gram negative short rods, oxidase and indole negative, with limited fermentative ability, producing smooth white to creamy brown colonies on solid media, consistent with the genus Acinetobacter (Doughari et al. 2011). All isolates assimilated phenol, glucose, potassium gluconate, capric acid, malate, trisodium citrate, and phenylacetic acid. The ability to metabolise aliphatic alcohols, unbranched hydrocarbons, and relatively recalcitrant aromatic compounds is a diagnostic feature of the genus (Towner 1992). The starting concentration of the Acinetobacter inoculant in the monolayer mineral salts media was 9 × 10⁶ cfu per mL.

Both Acinetobacter isolates degraded all three monolayer compounds to varying degrees. The Narda Lagoon isolate degraded C16OH most rapidly (P < 0.001 for isolate, incubation time and interaction, n = 24), with more than 50% of the compound lost within 2 days (Figure 1). The C18E1 compound was less readily degraded (isolate P = 0.093, incubation P < 0.001, interaction P < 0.001, n = 24), with most of the monolayer recovered after 4 days of incubation.

Biosurfactant produced by the bacteria aggregated microbial cells, compromising the validity of most standard bacterial quantification methods. Biosurfactant was evident as a viscous material adhering to the sides of the inoculated flasks only, forming a layer between the chloroform and water phases in separation funnels. DNA extraction and purification was selected as the bacterial quantification method, as the initial chloroform extraction procedure was the same as the solvent extraction procedure used for monolayer compound recovery.

The shorter chain C16OH monolayer was most readily utilized by the Acinetobacter culture (P < 0.001 for monolayer, incubation time and interaction, n = 45), with the rise in DNA evident within 2 days (Figure 1). Even after 4 days, the DNA concentration in the C18E1 cultures was low. The average ‘DNA’ value for C18E1 un-incubated, un-inoculated flasks (484 ng μL⁻¹) was much higher than for C16OH and C18OH, indicating the samples were contaminated with protein and/or phenol (Brown 1990). Subtracting the correction factor of 484 ng μL⁻¹ from DNA recovery values indicated that the microbial utilization of C18E1 may have been as low as 34 ng μL⁻¹, 128 ng μL⁻¹, and 227 ng μL⁻¹ for 2, 3 and 4 days of incubation respectively.

On clean water all three monolayers induced surface pressures above 30 mNm⁻¹ (Figure 2(a)), and increased evaporative resistance by between 2 and 6 s cm⁻¹ (Figure 2(c)) for at least 3 days. On Narda Lagoon water, the surface pressure (Figure 2(b)) and evaporative resistance (Figure 2(d)) dropped within 2 days, with the rate of decline greatest for the C18E1 monolayer (within 24 hours).

The slow rate of degradation of monolayer applied to solar-irradiated, distilled water was very similar for all three compounds (slope ratio analysis P > 0.999, n = 126; Figure 3). On irradiated Narda Lagoon water, the decline in concentration was rapid for the microbially resilient C18E1 monolayer (half-life reduced from 35 hours in distilled water to 3.3 hours in brown water), but not as rapid for the fatty alcohol C16OH and C18OH monolayers (reduced from 27.2 to 4.7, and 31.3 to 5.7 hours, respectively). The rank order of the susceptibility of the three monolayer compounds to photodegradation on Narda water (Figure 4) reflects the rank order of the loss in monolayer surface pressure and evaporation reduction on Narda water (Figure 2). Degradation occurred rapidly (half-life of hours), at a faster rate than microbial degradation (half-life of days).

Early studies using pelletised hexadecanol (C16OH) indicate the aquatic bacteria Pseudomonas and Flavobacterium readily degrade the monolayer, reducing the surface pressure within a day (Chang et al. 1962). In our Acinetobacter bioassays, increasing the carbon chain length (C18OH) and replacing the fatty alcohol with a glycol ether (C18E1) headgroup reduced monolayer susceptibility to microbial degradation (Figure 1). All three monolayer compounds were not susceptible to direct photodegradation (Figure 3), but even the brief exposure to sunlight during the transfer of petri dishes to the light-excluded box increased the rate of degradation of all three compounds on Narda water (difference between Narda Dark and Distilled Dark in Figure 4). Degradation of the C18E1 monolayer occurred within minutes, whereas in the microbial bioassay very little compound had degraded after 4 days of incubation (Figures 3 and 1). These results suggest the glycol ether headgroup is extremely susceptible to indirect photodegradation.

On clear water, the residence time of a condensed C18OH or C18E1 monolayer may approximate the 4 to 5 days required for cost-effective evaporative savings (Barnes 2008). However, the residence time of C18E1 on brown water may be too short for monolayer application to be feasible as a water conservation strategy. Narda Lagoon is relatively small (1 ha), receiving fresh leaf litter from riparian woodland vegetation and hardwood sawdust from a saw mill located on the bank (Pittaway & van den Ancker 2010). The average concentration of dissolved organic carbon in microlayer water over a 12 month monitoring period was 8.7 mg L⁻¹, with a specific ultraviolet light absorbance (SUVA) of 0.11 L⁻¹ mg⁻¹ cm⁻¹. The high rate of photodegradation in the microlayer may be associated with aromatic compounds leaching from the sawdust fines (a high SUVA; Weishaar et al. 2003) or with the high concentration of iron (Fe³⁺ 0.37 ± 0.04 mg L⁻¹), known to catalyse photodegradation (Howitt et al. 2008).

The natural synergy between photodegradation and microbial degradation rapidly degraded all three monolayer compounds on brown water containing high concentrations of chromophoric, dissolved organic matter. Our results confirm the glycol ether monolayer (C18E1) is more microbially resilient than the fatty alcohol monolayer compounds (C16OH and C18OH), but is much more susceptible to indirect photodegradation. The cost-effectiveness of repeat applications of the C18OH and C18E1 compounds as a water conservation strategy will be greatest on clear (oligotrophic) water storages. The susceptibility of the C18E1 compound to indirect photodegradation indicates application of this monolayer to brown water storages may not be cost-effective.

The assistance of Adele Jones (Faculty of Science, University of Southern Queensland, Toowoomba), Dr Nigel Hancock (National Centre for Engineering in Agriculture, University of Southern Queensland, Toowoomba), and Dr Geoff Barnes (University of Queensland) is gratefully acknowledged. We thank Dr Alan Lisle (University of Queensland) for the slope ratio spreadsheet, Mr John Mills (Toowoomba Regional Council) for the activated sludge sample, and Mr Raed Ahmed Mahmoud Al Juboori for the Fe³⁺ data. Funding was provided by the Urban Water Security Research Alliance, and the Co-operative Research Centre for Irrigation Futures.