Evaluation of Algal Biofilms on Indium Tin Oxide (ITO) for Use in Biophotovoltaic Platforms Based on Photosynthetic Performance

In photosynthesis, a very small amount of the solar energy absorbed is transformed into chemical energy, while the rest is wasted as heat and fluorescence. This excess energy can be harvested through biophotovoltaic platforms to generate electrical energy. In this study, algal biofilms formed on ITO anodes were investigated for use in the algal biophotovoltaic platforms. Sixteen algal strains, comprising local isolates and two diatoms obtained from the Culture Collection of Marine Phytoplankton (CCMP), USA, were screened and eight were selected based on the growth rate, biochemical composition and photosynthesis performance using suspension cultures. Differences in biofilm formation between the eight algal strains as well as their rapid light curve (RLC) generated using a pulse amplitude modulation (PAM) fluorometer, were examined. The RLC provides detailed information on the saturation characteristics of electron transport and overall photosynthetic performance of the algae. Four algal strains, belonging to the Cyanophyta (Cyanobacteria) Synechococcus elongatus (UMACC 105), Spirulina platensis. (UMACC 159) and the Chlorophyta Chlorella vulgaris (UMACC 051), and Chlorella sp. (UMACC 313) were finally selected for investigation using biophotovoltaic platforms. Based on power output per Chl-a content, the algae can be ranked as follows: Synechococcus elongatus (UMACC 105) (6.38×10−5 Wm−2/µgChl-a)>Chlorella vulgaris UMACC 051 (2.24×10−5 Wm−2/µgChl-a)>Chlorella sp.(UMACC 313) (1.43×10−5 Wm−2/µgChl-a)>Spirulina platensis (UMACC 159) (4.90×10−6 Wm−2/µgChl-a). Our study showed that local algal strains have potential for use in biophotovoltaic platforms due to their high photosynthetic performance, ability to produce biofilm and generation of electrical power.


Introduction
Algae are amongst the most efficient photosynthetic organisms with fast growth rates, diverse products and tolerance to extreme environments. Diatoms, green algae and cyanobacteria (also referred to as the blue-green algae, Cyanophyta) are the major primary producers in the aquatic ecosystem, contributing to carbon dioxide removal, photo-oxygenation and also serving as sources of valuable biochemicals [1]. The biomass productivity of microalgae was estimated to be 50 times higher than switchgrass, which is the fastest growing terrestial plant [2]. With the increased interest in alternative energy sources, algae are being investigated as feedstock for biodiesel, bioethanol, biohydrogen and bioelectricity production [3,4,5]. Microalgae have oil content exceeding 80% DW, grow fast and can be mass cultured using open ponds or enclosed photobioreactors [6]. The filamentous cyanobacterium Anabaena was reportedly the first to be used for hydrogen generation [7,8]. In 1997, electricity was generated using 2-hydroxy-1,4-naphthoquinone as an electron shuttle between Synechococcus sp. (UTEX 2380) and a carbon-cloth anode [9]. The current intensity increased with increasing cell concentration, reaching 320 mAcm 22 [9]. A bioreactor (microbial fuel cell) with an air cathode and a graphite-felt anode coated by a biofilm of bacteria and algae, generated electricity when irradiated. On day 10, the voltage output and current density produced by the reactor were 0.32 V and 8.6 mAcm 22 respectively [10].
Biological components have since then been introduced into fuel cells (FCs), giving rise to microbial fuel cells (MFCs). MFCs were designed to operate as a new generation of solar cells called biological photovoltaic devices (Biophotovoltaic, BPV) [11]. Both the MFCs and BPV share similar functions; where MFCs generate electricity from the metabolic process of living microbes, whereas BPV produce electricity from light energy via the light harvesting apparatus of photosynthetic organisms [12]. Previous BPV studies have utilized various exogenous soluble mediator compounds to facilitate electron transfer such as 5 mM ferricyanide as mediator between the biological materials and the anode [11]. In their paper, a new BPV device was fabricated with several advantages over the previous experiments such as multiple microchannels to facilitate multiple simultaneous experiments, removal of the need for external energy supply and a new design to cater for investigation with various types of biological materials instead of solely a single type. Results indicated a direct relationship between effects of cell density, electron mediator concentration and light intensity with the efficiency of the BPV device. The biological materials used in this study were intact Synechocystis cells and thylakoid membranes isolated from the cells, which generated total power output of 4.71 and 9.28 nWmmol Chl 21 respectively. In another study, McCormick et al. [13] replaced the use of exogenous mediators with the development of biofilms. Several algae were grown directly on an Indium tin oxide-Polyethylene terephthalate (ITO-PET) anode on a sandwich type or an open air design. This work demonstrates the ability to produce simple, portable BPV devices without the need of an artificial electron mediator. The Synechococcus biofilm in this work produced a peak power output at 1.03610 22 Wm 22 under 10 Wm 22 of white light [13]. Biofilm development offers several advantages by increased power output due to direct contact between cell and electrode and reduced internal potential losses. In 1684, Antonie van Leeuwenhoek [14] reported the existence of animalcules on teeth, representing the first scientific report on biofilms. Microorganisms are able to attach to surfaces and form a hydrated polymeric matrix termed ''extra-cellular polymeric substances (EPS)'' that hold the biofilm together [15]. In an EPS matrix, there are interstitial water channels, which separate bacterial cells from each other, thereby allowing the transportation of nutrients, oxygen and genes [16]. Biofilms composed of microorganisms attached to surfaces, form a hydrated polymeric matrix consisting of polysaccharides, protein and nucleic acids [15]. There is a growing interest in the study of artificial phototrophic biofilms. In biofuel production, the cultivation of algae as biofilms reduces costs due to reduction of water volume and power required for pumping. This increases biomass and also avoids costly harvesting and dewatering technologies [17]. Traditional methods of monitoring changes in biomass of microphytobenthic biofilms are destructive, for example, pigment sampling by using syringe coring method. This method involves pulling the syringe plunger up while pushing the syringe to the sediment to create counter pressure [18]. Fluorescence techniques have been employed to measure photosynthetic performance of biofilms in sediments without interfering with the sediment surface [19]. Radiant energy absorbed by chlorophyll can undergo one of three fates: (i) used for photosynthesis (ii) dissipated as heat or (iii) re-emitted as chlorophyll fluorescence [20]. Hence, by measuring the yield of chlorophyll fluorescence, information about the efficiency of photochemistry and heat dissipation can be generated using a pulse amplitude modulation (PAM) fluorometer (Diving PAM, Walz, Germany) [21]. Photosynthesis parameters such as electron transport rate (ETR) can be calculated to measure the efficiency of photochemistry of Photosystem II (PSII) [22].
In a recent paper, Luimstra et al. [23] produced a cost effective microbial fuel cell design that can be used on algae and cyanobacteria. Several strains of benthic cyanobacteria were screened and displayed electrogenic qualities suitable for microbial fuel cell purpose. The paper also describes a particularly good green alga, P. pseudovolvox that demonstrated greater electrogenic activity compared to the others. However, they also highlighted that the mechanisms used in the microorganisms to donate electrons have to be determined at this moment of time. Ng et al. [24] reported that Chlorella sp. (UMACC 313) and Spirulina platensis (UMACC 159) are able to form biofilms on ITO anode. Chlorella and Spirulina formed biofilms with coverage and maximum relative electron transport rate (rETRmax) of 99.46% and 140.796 mmol electrons m 22 s 21 and 80.70% and 153.507 mmol electrons m 22 s 21 respectively. Results indicate the potential for generating electrical energy from these microalgae using biophotovoltaic platforms.
The objectives of the present study are to (i) establish libraries of 16 strains of algae for selection of most suitable strains for application in biophotovoltaic platforms; (ii) investigate biofilm formation of selected algae on Indium tin oxide (ITO) and glass and determine the photosynthetic efficiency of the biofilms; (iii) Produce and test biofilms on Indium tin oxide (ITO) in a biophotovoltaic (BPV) device for electric power generation. Glass and ITO were chosen as substrates for biofilm formation to compare the adhesion of algae cells to these surfaces. Glass is hydrophilic with high surface energies and ITO is hydrophobic with low surface energy respectively [25]. ITO was an early favorite for hole injection cathode with good transparency and conductivity [26]. ITO was also successfully used to cultivate healthy biofilms for MFCs application [13].

Ethics statement
Not relevant. Only microalgae were used in this study.

Algae cultures
Fourteen local tropical algal strains from the University of Malaya Algae Culture Collection (UMACC) [27] and two strains from the Culture Collection of Marine Phytoplankton (CCMP), USA were screened for growth rate, biochemical composition and photosynthetic performance to be used for compiling the algal libraries of potential candidates for BPV platforms. The Cyanobacteria are treated as blue-green algae and will be referred to as Cyanophyta in this paper. All cultures were grown in Bold's Basal Medium for Chlorophyta, the green algae [28], Prov Medium for marine algae [27], Kosaric Medium for Cyanophyta, the bluegreen algae [27] and f/2 Medium [29] for marine diatoms (Table 1). An inoculum size of 20%, standardized at an optical density at of 0.2 at 620 nm (OD 620 nm ) from exponential phase cultures was used. The cultures were grown in 1 L conical flasks in an incubator shaker (120 rpm) at 25uC, with irradiances of 30 mmol photons m 22 s 21 on a 12:12 light dark cycle. Each microalga was grown in triplicate with a total volume of 500 ml. Growth was monitored based on OD 620 nm ; which has a high correlation (r 2 = 0.9) with chlorophyll a (Chl-a) [30]. In the present study, OD 620 nm was strongly correlated to Chl-a content (r 2 ranged from 0.9102 to 0.9877) for the 16 strains (2 Cyanophytes, 2 diatoms, 14 Chlorophytes) used.

Algal libraries
Growth and biochemical profiling. The specific growth rate (m, day 21 ) for all cultures were based OD 620 nm and calculated using the following formula: where N 2 is OD 620 nm at t 2; N 1 is OD 620 nm at t 1 , and t 2 and t 1 are time periods within the exponential phase [31].
The algal samples were harvested at the end of the experiment (stationary phase) by Millipore filtration using glass fibre filter paper (Whatman GF/C, 0.45 mm) for determination of dry weight and extraction of biochemicals. For dry weight (DW) determina-tion, a known volume of the culture was filtered onto an ovendried pre-weighed glass fibre filter, which was then dried in an oven at 100uC for 24 h. The DW was calculated as follows: DW(mg:L À1 )W eight of filter with algae (mg) À Weight of blank filter (mg) Volume of culture (L) The biomass at stationary phase was determined as dry weight (B DW ) as well as calculated from the Chl-a content 667 (B CHL ), assuming that Chl-a makes up 1.5% of the cell biomass [32,33]. Protein content of cells was determined by the dye-binding method after extraction in 0.5 N NaOH [34]. Carbohydrates extracted from the cells in 2N HCL were determined using the phenol-sulphuric acid method [35]. Lipids were extracted in MeOH-CHCl 3 -H 2 O (2:1:0.8) and determined by gravimetric method [36]. These biochemicals were expressed as %B CHL . This may avoid errors with the diatoms that contain silicate frustules but since the content of Chl-a may vary with taxa, these values can only be considered as estimates for comparison between strains.
Pulse amplitude modulation (PAM) fluorometer measurement of 16 algal strains. Photosynthetic parameters were measured fluorometrically using a Diving-PAM (Walz, Germany) [37,21,38]. Data provided by the PAM based on fluorescence is useful for assessing the performance of a photosynthetic microbial fuel cell. Inglesby et al. [39] showed that in a study using Arthrospira maxima, the use of in situ fluorescence detection allowed for a direct correlation between photosynthetic activity and current density. Rapid light curves (RLC) were obtained under software control (Wincontrol, Walz). Red light emitting diodes (LEDs) provided the actinic light used in the RLC at the level of 0, 307, 426, 627, 846, 1267, 1829, 2657 and 4264 mmol photons m 22 s 21 . The cultures of each strain were dark-adapted for 15 minutes before exposure to each light level for 10 seconds. Maximum quantum efficiency (F v /F m ), a parameter to indicate the physiological state of phytoplankton was used to indicate if the cells were stressed by the exposure to light: F v =F m~( F m -F 0 )=F m where F m is the maximum fluorescence and F 0 is the minimum fluorescence resulting in the variable fluorescence F v . The maximum photosynthetic efficiency was determined from the initial slope (a) of the RLC. The relative electron transport rate (rETR) was calculated by multiplying the irradiance by quantum yield measured at the end of each light interval. The RLC consists of eight consecutive ten-second intervals of actinic light with increasing intensity. The photoadaptive index (E k ) is obtained from the curve fitting model [40]. The interception point of the alpha (a) value with the maximum photosynthetic rate (rETRmax) is defined as: E k~r ETRmax=a. The Non-Photochemical Quenching (NPQ) reflects the ability of a cell to dissipate excess light energy harvested during photosynthe- sis as heat and is used as an indicator of photoprotection. NPQ is calculated as (F m 9). NPQ~(F m -F m 0 )=F m 0 . F m is the maximum fluorescence yield during the saturating flash and F m 9 is the maximum fluorescence in the light-adapted state during the saturating flash. All statistical analyses were performed using the Statistica 8 program. Eight strains were selected according to photosynthetic performance based on F v /F m , rETRmax, alpha, E k and NPQ. Different algal types and different habitats are also considered as factors that influence the selection of strain. and the diatom Coscinodiscus wailesii (CCMP 2513) were used for the biofilm studies. 100 ml of exponential phase cultures of OD 620 nm = 0.5 were used. Each culture was placed into a 200 ml autoclaved glass staining jar. ITO coated glass slides (purchased from KINTEC, Hong Kong) and glass slides measuring 20620 mm were placed in the staining jar with the microalgae culture and transferred into an incubator at 24uC illuminated by cool white fluorescent lamps (30 mmol m 22 s 21 ) on a 12:12 hour light-dark cycle to allow for the algae biofilms to form on the slides. This experiment was conducted in triplicates. The biofilm growth was monitored by photographing the slide surface with a Sony Cyber-Shot DSC-WX30 Camera every three days until the slides were completely covered by the biofilm. The surface area coverage, SAC (%) of the biofilm captured in the photograph was calculated using ImageJ software [41]. At the end of the experiment (day 15), the biofilm thickness of each slide was measured using Elcometer 3230 Wet Film Wheels [42]. The wheel was held using a finger and thumb by its centre and the wheel was placed on the wet film ensuring that it was perpendicular to the algal film. The wheel was rolled across the algal film through an angle of 180u and then removed from the surface. The thickness of the biofilms was recorded based on the scale on the side of the wheel. The biofilms were removed by washing using jets of distilled water from a pipette, into a sterile beaker for extracting biomass for determination of Chl-a content. The microalgae cells were then harvested by millipore filtration using filter paper (Whatman GF/C, 0.45 mm) and the Chl-a of the eight strains were extracted using acetone [31].The absorption of the extract was measured at 665 nm, 645 nm and 630 nm and the Chl-a content calculated using the formula below: Pulse amplitude modulation (PAM) fluorometer measurement of biofilms Photosynthetic parameters were measured fluorometrically using a Diving-PAM (Walz, Germany) as described above [21,38]. RLC were obtained under software control (Wincontrol, Walz). Red light emitting diodes (LEDs) provided the actinic light used in the RLC at the level of 0, 127, 205, 307, 426, 627, 846, 1267 and 1829 mmol photons m 22 s 21 . The biofilm of each ITO slide on day 15 was dark adapted for 15 minutes prior to the exposure to each light level for 10 seconds.

BPV set up and electrical measurement
The BPV devices used in these studies was provided by our collaborators from the University of Cambridge. The closed, single-chamber BPV consisted of a 50650 mm platinum-coated glass cathode placed in parallel with 35635 mm ITO coated glass with biofilm grown on the surface (10 mm apart) in a clear Perspex chamber sealed with polydimethylsiloxane (PDMS) and then filled with medium (Figure 1 a-d). The body of the open-air, singlechamber BPV was constructed of clear Perspex [11,13].
Biofilms of the two strains of Cyanophytes, the Synechoccus elongatus (UMACC 105) and Spirulina platensis (UMACC 159), and the two Chlorophytes Chlorella vulgaris (UMACC 051), and Chlorella sp. (UMACC 313) grown on ITO, were placed in the devices and the experiment conducted in triplicates. Crocodile clips and copper wire served as connection between anode and cathode to the external circuit. Prior to operation, the chambers were filled with fresh medium (Bold's Basal medium and Kosaric medium) and maintained at 25uC with irradiance of 30 mmol photons m 22 s 21 for the duration of the experiments. Current outputs were measured using a multimeter (Agilent U1251B). Polarization curves were generated for each strains by applying different resistance (

Algae libraries
Growth and biochemical profiling. Table 2

Growth and photosynthetic efficiency of biofilms
The biofilms were grown on glass and ITO slides in triplicates. Table 3 shows the surface area coverage (% SAC) of biofilms produced by the eight strains on two different substrates (glass and ITO) on day 3, 6, 9, 12 and 15. The % SAC was monitored every three days. Visible growth of the biofilms was already observed on day 3 except for Coscinodiscus wailesii (CCMP 2513) which did not show any biofilm growth on both the glass and ITO slides until day 12 and registered the lowest % SAC among the eight strains. After three days of incubation, the Cyanophytes Synechococcus elongatus (UMACC 105) and Spirulina platensis (UMACC 159) had formed appreciable biofilms on both glass and ITO. Of the Chlorophytes, the Chlorella (UMACC 313) from POME pond and the freshwater Chlorella (UMACC 051) had high % SAC as well. All strains reached their maximum % SAC after 15 days of inoculation. Chlorella sp. (UMACC 313) showed fastest biofilm coverage, having achieved around 80% SAC on day   Table 4 gives details of the biofilms and the PAM data on day 15 of the study. The fluorescence characteristics of the biofilms from the eight strains exposed to a similar range of irradiance were

BPV set up and electrical measurements
A study was carried out to correlate the %SAC of biofilm on ITO with the amount of photocurrent being generated as well as the overall performance of the device. Figure 4

Discussion
The preliminary screening of 16 strains, comprising 14 local strains belonging to the blue-green and green algae, and two diatoms from the CCMP, allowed the selection of eight out of the 16 strains that were characterized in terms of growth rate, biochemical composition (protein, lipid and carbohydrate contents) and photosynthetic efficiency, for further studies on biofilm formation. The 16 strains were cultured till stationary phase to observe the batch culture characteristics. At stationary phase, the lipid contents are expected to be higher than at exponential phase of culture, as lipids are accumulated upon reaching stationary phase, as opposed to protein [30]. This was clearly observed with most of the strains except for the Cyanophyta (Synechoccus elongatus UMACC 105) and two Chlorophytes (Chlorella vulgaris UMACC 051 and Chlorococcum UMACC 207), where protein was highest at day 15. In general, the Chlorophytes contained higher lipid than the other strains, with Chlorella sp. (UMACC 256) having the highest lipid (63.64%DW) contents on day 15.
Eight out of the 16 strains, namely two Cyanophytes, five Chlorophytes and one diatom, were selected for the biofilm studies. The Cyanophytes were selected because blue-green algae produce abundant mucilage and are expected to form biofilms easily. The blue-green algae has higher potential to form a hydrated polymeric matrix termed extracellular polymeric substances (EPS) which encourages the biofilm formation [15]. Also from the preliminary screening, the Synechococcus elongatus (UMACC 105) and Spirulina platensis (UMACC 159) had high rETRmax values, indicating efficient photosynthesis. Chlorella vulgaris (UMACC 001) has been used for many studies in our laboratory and has shown ability to grow in wastewaters [43,44,45] and has potential for biofuel production [30]. In 2007, Wong and co-workers [46] compared the tolerance of Antarctic, tropical and temperate microalgae to ultraviolet radiation (UVR) stress. When the Chlorella vulgaris (UMACC 001) was exposed to ambient light, the specific growth rate was 0.16 d 21 but decreased to 0.04 d 21 when exposed to ultraviolet radiation (UVR) treatment. Vejesri et al. [30] reported that Chlorella vulgaris (UMACC 001) was a potential feedstock for biodiesel production due to high specific growth rate (m = 0.42 d 21 ) and high saturated fatty acid (SFA) content (68.2%DW). Chu and co-workers [43] reported that immobilized cultures of Chlorella vulgaris (UMACC 001) in alginate removed 48.9% of color from textile wastewater. Chlorella vulgaris (UMACC 001) showed high NPQ and strong photoprotection with higher capacity to cope with high irradiance. Chlorococcum (UMACC 207) had high rETRm as well as high protein (30.29%DW) content. Chlorella sp. (UMACC 313) and Chlorella vulgaris (UMACC 051) isolated from the aerobic pond for palm oil mill effluent (POME) treatment were selected because they formed biofilms on the base and sides of the culture flasks, and had very high lipid content. One marine alga Chorella sp. (UMACC 256) was included because it had the highest rETRmax value among the three marine strains. It was also shown that this strain had high specific growth rate (0.75 d 21 ) and high saturated fatty acid (SFA) content (53.8%DW) in our previous study [30]. One strain from the CCMP culture collection was also selected for comparison with the UMACC cultures. The strain, Coscinodiscus wailesii (CCMP 2513) was a centric diatom with high rETRmax (166.32 mmol electrons m 22 s 21 ) and higher lipid content of the two diatoms. In addition to the photosynthetic parameter of rETRmax, consideration was also given to the biofuel potential of the strains, in terms of lipid content.
In the biofilm studies, strains like Synechococcus elongatus (UMACC105) and Chlorella (UMACC 313) were observed to have formed appreciable biofilms after only three days. This may be based on the high production of EPS by Synechococcus elongatus (UMACC 105) and the high growth rate of Chlorella (UMACC 313). Application of PAM fluorometry in the study of algae biofilms has provided valuable information on the operation of algal BPV platforms. One of the most important information is the non-destructive generation of light-response curves of photosynthetic activity [47]. The distance between the optical fiber-optics and the sample surface was set at 2 mm [48]. Eight selected strains  ) were identified to produce the most stable biofilms on ITO. The growth and viability of these microorganisms in a working BPV device would have a significant impact on the photocurrent that would be generated [13]. The biofilm samples were dark adapted for at least 15 minutes before PAM measurement. When cultures were placed in the dark, minimum (F o ) and maximum (F m ) fluorescence both increased, as did the maximum PSII quantum efficiency (F v / F m ). After dark adaption, the reaction centres were in a relaxed stage, and ready to receive light for photosynthesis [19]. The maximum quantum yield F v /F m was obtained when all reaction centres were opened and was proportional to the fraction of reaction centres capable of converting absorbed light to photochemical energy [49]. F v /F m is often used as an indicator of photosynthetic capacity. Various studies have reported values of F v /F m ranging between 0.1 to 0.65 for natural populations of microalgae [50]. Results in the present study indicated that most samples were in a healthy condition, as the F v /F m values were all above 0.800 except for Synechococcus elongatus UMACC 105 with a F v /F m value of 0.799. Serôdio et al. [51]   Our results showed that algal biofilms have better photosynthetic performance than suspension cultures. Results showed that the algal strains used in this study have high photosynthetic efficiency indicating their potential for use in BPV platforms. In addition, high rETRmax and E k values obtained from the present study reflected an adequate adaptation to high irradiance [53], and was observed in the two Cyanophytes (Synechococcus elongatus UMACC 105) and the diatom (Coscinodiscus wailesii CCMP 2513). The NPQ values for each strain were compared and an increase in NPQ with changing light intensities was observed in all cases of the biofilms (Figure 3). The change in NPQ is expected due to photoprotection of photosystem II to avoid light induced damage [54]. In the case of the Cyanobacteria (Cyanophyta), the resulting changes in NPQ have been reported to be triggered by the light activation of orange carotenoid protein found in phycobilisomes [55]. In comparison, the NPQ in diatoms is mediated by the light-harvesting complex stress-related (LHCSR) protein and the conversion of diadinoxanthin (Ddx) to diatoxanthin (Dtx) [56]. In green algae (Chlorophyta), the controlled change in pH and its effect on the xanthophyll cycle has a predominated effect on NPQ [54]. In the present study, the Chlorophyte strains were observed to have higher ability (higher NPQ values) to photoprotect themselves than the Cyanophytes and diatoms. The Cyanophytes (blue-green algae) were observed to have the lowest value or in some cases, no NPQ, indicating that they have the lowest capacity for photoprotection. There was no clear correlation between the F v /F m values and the NPQ values of the 16 strains in the screening experiment as well as the eight strains used for the biofilm study, except where the lowest F v /F m values of the two Cyanophytes also corresponded to the lowest NPQ values (Tables 2 & 4 and Figures 2 & 3). Interestingly the low NPQ values of the two Cyanophytes and two diatoms corresponded to higher E k values, and may indicate that these strains may in fact be tolerant of higher irradiance. This however, has to be confirmed by detailed studies on the relationship between the various parameters. In this study, the four strains, Cyanophytes Synechococcus elongatus UMACC 105, Spirulina platensis UMACC 159 and Chlorophytes Chlorella UMACC 051 and UMACC 313, were selected for the BPV study based on strong ability to form biofilms; however their NPQ ranged from low to high values, to allow the correlation of NPQ to photo current production in a working BPV device. The diatoms had high rETRmax but very low NPQ, and were not selected due to poor ability to form biofilms. On analysis of data from the BPV studies, it was apparent that the Cyanophyte Synechococcus elongatus (UMACC 105), was able to give highest performance in the BPV device but it had the lowest NPQ value. The maximum power density was similar for the other Cyanophyte Spirulina sp. (UMACC159) and two Chlorophytes Chlorella (UMACC 051 and 313), which had higher NPQ. Further studies would need to be carried out to correlate the NPQ of the strains with the ability to extract charge from biofilms in a working BPV device.
Synechococcus elongatus (UMACC 105) showed higher power outputs compared to other strains. This may be due to the fact that Synechococcus elongatus (UMACC 105) registered a high rETRmax values (147.50 mmol electrons m 22 s 21 ), high E k value and readily produced biofilms (84.68% SAC on ITO) which were able to accumulate a high level of charge. The Synechococcus produces unicells of small dimensions and with production of EPS, would form compact biofilms on the anode, compared to the filamentous Spirulina which like the Arthrospira used in the study by Inglesby et al. [39], formed porous biofilms. The maximum power density obtained in this study ranged from 1.12610 24 to 3.13610 24 Wm 22 . Inglesby et al. [39] reported very low power densities from 6.7 to a maximum of 24.8 mW m 22 with a similar BPV device using biofilm of the Cyanophyte Arthrospira maxima on ITO as well, considering that the theoretical maximum power density achievable when using microalgae in microbial fuel cells is 2.8 Wm 22 . The low power density obtained in the present study reflects an unoptimised system with the results aimed only for comparison between the four strains. Optimisation of the process in terms of biofilm thickness and irradiance properties may improve the power output. Optimisation of temperature and light intensity in the BPV device may enhance power output, as was shown by Inglesby et al. [39], where an increase of temperature from 25 to 35uC increased power output from 9.9 to 24.8 mW m 22 while there may be a limitation to how high the light intensity can be, due to photoinhibition. In the present study. Based on power output per Chl-a content, the four strains can be ranked as follows: Synechococcus elongatus (UMACC 105) (6. . These values were obtained by dividing the maximum power density by the total Chl-a extracted from the whole algal biofilm on the ITO slides. However no correlation (P,0.05) was observed between power output and Chl-a content. Of the four strains, Chlorella vulgaris (UMACC 051) had highest lipid (31.09% DW) and crude protein (46.70% DW) contents, indicating additional advantages over the other strains. Although it is premature to speculate on the long-term usage of algal cultures in BPV devices, on a large scale, biomass may be generated continuously in the BPV. The biomass may be harvested on a regular basis to prevent over-buildup of the biofilm, which results in reduced bioactivity at the surface of the biofilm furthest from the anode. Photosynthetically efficient strains with high lipid and protein productivities, may have an added advantage over other strains, providing a valuable biomass that may enter an alternative energy-producing system like biodiesel production. This may overcome the lower energy output from algal BPV systems compared to commercially available photovoltaics [39]. However the additional cost of continuous nutrient supply to the biofilms on the anodes and cost of removal of surplus biomass from the biofilms may necessitate innovative design and operation of the BPV devices.

Conclusions
The present study indicated that local algal strains were good candidates for utilization in BPV platforms in future. According to our screening results, all eight strains were able to form biofilms on ITO anode surfaces with good photosynthetic performance. From this, four strains were selected for the BPV studies, based on their high photosynthetic performance and ability to produce biofilms. The Chlorophytes Chlorella species (UMACC 051, UMACC 313) and the Cyanophytes, Spirulina platensis (UMACC 159) and Synechococcus elongatus (UMACC 105), demonstrated exoelectrogenic activity and showed their capacity to produce significant electrical power outputs without the requirement of additional organic fuel. More work is required to further understand the mechanisms of harnessing light energy and converting them to electricity as well as to investigate the correlation between PAM data and the BPV power output.