From biofilm ecology to reactors: a focused review

Biofilms are complex biostructures which appear on all surfaces that are regularly in contact with water. A biofilm consists of prokaryotic cells and other microorganisms such as yeasts, fungi, and protozoa that secrete a mucilaginous protective coating in which they are encased (i.e., extracellular polymeric substances or EPS). Biofilms can form on solid or liquid surfaces as well as on soft tissue in living organisms. Biofilms are typically highly resilient constructs that resist conventional methods of disinfection. Biofilm formation is an ancient and integral component of the prokaryotic life cycle, and it is a key factor for survival in diverse environments. Biofilms are structurally complex, dynamic systems with attributes of both primordial multicellular organisms and multifaceted ecosystems. The formation of biofilms represents a protected mode of bacterial growth that allows cells to survive in hostile environments and disperse to colonize new niches (Hall-Stoodley et al. 2004).

The presence of biofilms may have a negative impact on the performance of various systems. For example, biofouling of ship hulls and membrane surfaces reduces performance and efficiency, resulting in marked financial costs. Pathogenic biofilms have also proven detrimental to human health. Biofilm infections, such as pneumonia in cystic fibrosis patients, chronic wounds, chronic otitis media, and implant- and catheter-associated infections, affect millions of people in the developed world each year, and many deaths occur as a consequence (Bjarnsholt 2013). Foodborne diseases – often caused by biofilm-forming pathogens – are a public health concern throughout the world (Srey et al. 2013). The development of multispecies biofilms on teeth (i.e., dental plaque), and their associated bacterial pathogenesis, can lead to gum disease and tooth decay (Kolenbrander et al. 2010). Biofilms may also be undesirable in the open water environment. For example, algal mat formation on water bodies is a component of the eutrophication process. Finally, biofilms that develop on the interior walls of pipes that comprise a potable water distribution system can lead to additional chlorine demand, coliform growth, pipe corrosion, poor water taste, and foul odor (Hallam et al. 2001).

On the other hand, biofilms, can be controlled and used beneficially for the treatment of water (defined herein as potable water, municipal and industrial wastewater, fresh/brackish/salt water bodies, groundwater) as well as in water stream-based biological resource recovery systems. The investigation of biofilms in the water environment will be classified for the purpose of this review into three major categories: (1) biofilm ecology, (2) biofilm reactor technology and design, and (3) biofilm modeling. Biofilm ecology is defined here as the study of components and processes that take place in the biofilm. Biofilm reactor technology and design encapsulates the development, design, operation, and optimization of bioreactors that target controlled biofilm utilization. Biofilm modeling is the development and application of various computational approaches to simulate, predict, or synthesize the processes occurring in biofilms and biofilm reactors.

The term biofilm refers to the microbes and associated deposits on a surface embedded in the matrix of EPS. The broader term, biofilm system, includes other components affecting the biofilm, and usually consists of, at least, the substratum (on which the biofilm forms) and the bulk phase (which flows over the biofilm). This paper reviews key research and practical events related to these areas of biofilm study, focusing on research and practice-related trends in biofilm-related biology, biofilm reactors, and models of particular relevance to challenges and opportunities regarding biofilms and biofilm systems.

The biology of biofilms includes a diverse array of topics. The current focus of biofilm biology is dedicated to applying state-of-the-art approaches to evaluate biofilm ecology in relation to structure and function, including the identification of factors that drive biofilm formation and dispersal.

The symphonic application of biological methods is essential to understand microbial films biology. The currently and often used combined application of quantitative polymerase chain reaction (qPCR), fluorescent in situ hybridization (FISH), advanced 2-D microscopy, and micro-scale chemical sensors has allowed biofilm researchers to create a better vision of biofilm make-up – including both the cellular matter and their excretions – than ever before. This insight has proven valuable to advancing the understanding of biofilm structure and function. qPCR has been used to further our understanding of biofilm structure and function, and the roles that biofilms play in a bioreactor (Kim et al. 2011). Applying qPCR combined with micro-dissection has allowed one to quantify the stratification of functional guilds in biofilms (Terada et al. 2010). FISH is a technique that is based on hybridizing a fluorescently labelled DNA probe to (typically for bacterial investigations) complementary sequences present in the bacterium's 16S rRNA. Phylogenetically distinct groups of bacteria can be simultaneously visualized by proper choice of DNA probes. When properly applied to biofilms – and in combination with the right microscopic method and detection method (often multi-channel confocal laser scanning microscopy (CLSM)) the technique allows one to identify the spatial organization and relative location of different bacterial groups (Okabe et al. 1999; Vlaeminck et al. 2010). CLSM, transmission electron microscopy, and soft X-ray scanning transmission X-ray microscopy have been used to map the distribution of macromolecular sub-components (e.g., polysaccharides, proteins, lipids, and nucleic acids) of biofilm cells and their associated EPS matrix (Lawrence et al. 2003). More recently, optical coherence tomography has been applied to visualize the mesoscale structure of biofilms (Wagner et al. 2010), and confocal Raman spectroscopy has provided a tool for studying the chemical heterogeneities of biofilms in situ (Sandt et al. 2007).

Microbial ecology is an essential component of biofilm studies because of the desire to control biofilm development, biochemical transformation processes, and dispersion. Davies et al. (1998) suggested that a cell-to-cell signal is involved in the development of Pseudomonas aeruginosa biofilms. These findings implied involvement of an intercellular signal molecule in the growth of P. aeruginosa biofilms, which suggests possible targets to control biofilm growth, for example, on catheters, in cystic fibrosis, and in other environments where problematic P. aeruginosa biofilms persist. Shrout et al. (2006) documented the impact of quorum sensing and swarming motility on P. aeruginosa biofilm formation as being nutritionally conditional. Nitric oxide (NO) is an important gaseous messenger molecule in a biological system that is produced by one cell, penetrates through membranes, and regulates the function of another cell (Zetterström 2009). This discovery presented an entirely new principle for signaling in biological systems. Various NO donors of clinical and industrial significance have been demonstrated viable, in a laboratory system, for dispersal in single- and multispecies biofilms (Barraud et al. 2009). Applications in natural formed biofilms have, however, not yet been reported.

Research on biofilm reactors has been the source of an interesting new metabolic pathway. The anaerobic ammonium oxidation (anammox) process was discovered in a pilot-scale denitrifying fluidized bed biofilm reactor. From this system, a highly enriched microbial community was obtained, dominated by a single deep-branching planctomycete, Candidatus Brocadia anammoxidans (Jetten et al. 2001). Since that time, the utilization of anammox microorganisms in biofilm reactors has proven popular, cost effective, and efficient.

The continued development of knowledge about phototrophic biofilms has elucidated their utility for nutrient removal from wastewater, heavy metal accumulation and water detoxification, oil degradation, agriculture, aquaculture, and sulfide removal from contaminated waste streams (Roeselers et al. 2008).

Microorganisms in biofilms live in a self-produced gelatinous matrix of EPS, consisting primarily of polysaccharides, proteins, nucleic acids and lipids. EPS provide biofilms with mechanical stability, mediates bacterial adhesion to surfaces, and serves as the three-dimensional polymer network that interconnects and transiently immobilizes bacterial cells inside a biofilm. EPS are also capable of entrapping, or bioflocculating, biodegradable and non-biodegradable particulates in the polymeric matrix (Boltz & LaMotta 2007). Conceptually, some basic functions, properties and constituents of the EPS matrix are known, but the kinetics of EPS production and consumption, their contribution to the conversion of materials entrapped within them, and their contribution to metabolic kinetics and biochemical transformation rates owing to microbial growth in a biofilm are poorly defined. Thus, biofilm models explicitly describing EPS (e.g., Alpkvist et al. 2006; Celler et al. 2014) are scarce and lack the measurement (i.e., quantification) of fundamental mechanical properties. Hence, unlocking this not yet well-defined aspect of biofilms remains a challenge to researchers (Flemming & Wingender 2010). EPS play an important role in biofilms, including the agglomeration of selenium (Gonzalez-Gil et al. 2016). EPS extraction methods are still not well validated (Pellicer-Nàcher et al. 2013a, 2013b); several types of EPS matrix polymers are not solubilized in standard extraction methods (Lin et al. 2013), whereas the methods to measure polysaccharides and proteins easily give biased results (Le & Stuckey 2016).

The uptake and biochemical transformation of micro-constituents (including pharmaceuticals) that can occur during wastewater treatment (Jelic et al. 2011) is still a significant challenge to biofilm researchers and treatment system designers. Kim et al. (2009) compared the removal efficiencies of micro-constituents classified as trace organic chemicals (including endocrine disrupting compounds (EDCs) and estrogenic activity). Results suggest that the system with a biofilm compartment out-performed the suspended growth control process. Thus, bioreactors having a biofilm compartment, such as integrated fixed-film activated sludge (IFAS) systems, may be beneficial for enhancing the removal of estrogens and at least some trace organics. These researchers found further evidence for removal by heterotrophic biodegradation, rather than by sorption or removal by nitrifiers. This is significant, given the apparent correlation of ammonia-nitrogen oxidation with the metabolism of specific EDCs, while the biochemical transformation of other EDC types fails to correlate with nitrification. Torresi et al. (2016) suggest that biofilm thickness influences the biodiversity of nitrifying biofilms grown in moving bed biofilm reactors (MBBRs), and that this parameter influences a biofilm's capacity for micro-pollutant removal. The biochemistry and microbiology of micro-pollutant transformation – in context of biofilms – is under active investigation, and the identification of the responsible organisms, the role of different functional guilds, the contribution of co- vs primary metabolisms, and the significance of biofilm redox conditions are all under examination.

Biofilms can be controlled and harnessed to provide the basis for their utilization for water treatment via biofilm reactors. The presence of biofilms may also be undesirable in a biological water treatment system, however, and can cause operational difficulties that increase the expense of treatment.

Biofilm reactors represent the primary means to harness the usefulness of biofilms for the treatment of water(s). Biofilms in these reactors serve as a principal mechanism for the biological transformation of nutrients that are regarded as environmental pollutants (e.g., biodegradable organic matter, nitrogen, and phosphorus). Several types of biofilm reactors have been utilized for water treatment, but currently much focus is on MBBRs and IFAS processes, membrane-supported biofilm reactors (MBfRs), and granular processes.

MBBRs and IFAS processes are mature technologies that continue to evolve. State-of-the-art MBBRs and IFAS processes use submerged free-moving biofilm carriers and can be used for carbon oxidation, nitrification, denitrification, and deammonification (Rusten et al. 2006; McQuarrie & Boltz 2011; Odegaard et al. 2014). Recent research has offered expanded insight into the role of these biofilm carrier types on mass transfer, and the impact of hydrodynamics on related biochemical transformation processes (Herrling et al. 2014; Melcer & Schuler 2014). Globally, there are more than 1,200 full-scale, operating MBBRs having a capacity of 200 population equivalent (p.e.) or greater. It is estimated that approximately 25% of these units are IFAS. MBBRs having a capacity less than 200 p.e. are numbered more than 7,000, globally. More than 100 MBBRs exist for nitrification in aquaculture. It is estimated that there is an equal distribution of MBBRs amongst industrial and municipal wastewater treatment facilities designed to treat waste streams for p.e. greater than 200. The geographic distribution of these installations is estimated as:

• facilities greater than 200 p.e. – 40% in Europe, 30% in North America, 20% in continental Asia and the South Pacific (not including India), and 10% in Africa;

• facilities less than 200 p.e. (including onsite facilities) – 80% in Europe, 10% in North America, and 10% in continental Asia and the South Pacific (not including India).

An MBBR-based process at the Lillehammer wastewater treatment plant (WWTP), Lillehammer, Norway, for the treatment of municipal wastewater has been described by Rusten et al. (1995) and an example IFAS installation has been documented at the Fields Point Wastewater Treatment Facility, Rhode Island, USA. The MBBR is an effective platform for simultaneous partial nitritation and deammonification. The AnitaMOX™ process is a commercially available system that has MBBR/IFAS appurtenances and exploits the partial nitritation/anammox process (PN/A) (Veuillet et al. 2014). A full-scale AnitaMOX system exists at the Sjölunda WWTP, Malmö, Sweden (Christensson et al. 2013).

Granular biomass development and utilization in a sequencing batch reactor (SBR) has proven an effective and highly promising environmental biotechnology for the treatment of contaminated water streams. Aerobic granules can be formed and maintained in SBRs (de Kreuk et al. 2007). The potential for stable aerobic granule formation was reported by Beun et al. (1999). Currently, more than 25 WWTPs are operating or under construction on four continents, including Europe (five in the Netherlands), South America, Africa and Australia, that will utilize aerobic granular biomass processes. All of these WWTPs are designed for biological nutrient removal from municipal wastewaters. The largest capacity constructed to date has a capacity 517,000 p.e., with an average daily flow of 55,000 m³/day, in Rio de Janeiro, Brazil. A commercially available aerobic granular sludge system that has been used for successful biological nutrient removal from screened/degritted wastewater or primary effluent is named NEREDA™. A full-scale NEREDA process at Garmerwolde WWTP, The Netherlands, has been described in the literature (Pronk et al. 2015). The NEREDA process maintains a constant liquid/biomass volume. The filling, settling, and decanting steps occur simultaneously during approximately 25–33% of the operational period. The remainder of operation is reserved for aeration (i.e., reaction period). Approximately 10–15 minutes is required to achieve reactor quiescence before a next cycle can start with influent feeding from the bottom. These typical operational parameters, along with appropriate influent wastewater characteristics, result in effluents having TN <5 g/m³ and TP <1 g/m³. These simple bioreactors are, essentially, an empty tank with fine-bubble aeration and an influent wastewater distribution system along the tank bottom. The treated effluent flows over an effluent weir situated along the top of the tank. The bioreactor has no mixers, but does have an effluent discharge via overflow weirs and a waste sludge collection system (which is situated near the top of the settling sludge bed to promote wasting of more slowly settling sludge). Another approach to benefit from granular biomass is to use a cyclone or screens for the selective retention of granular biomass. Granules have also been used for PN/A systems of high ammonia-nitrogen concentration waste streams from digested sludge dewatering and anaerobically treated wastewater in processes such as ANAMMOX™ (van der Star et al. 2007) and DEMON™ (Wett 2007).

Another biofilm reactor type that exhibits great potential is the MBfR. The potential diverse range of applications for this process is a formidable strength. Gas-delivery to the liquid phase in these systems happens by means of a membrane (tubular, hollow-fiber, or flat) on which the biofilm directly grows. Electron donor and electron acceptor are subject to counter-diffusion through the biofilm from the bulk of the liquid and from the membrane lumen. Two systems have been promoted: (1) the hydrogen-based MBfR (Rittmann 2006), which delivers hydrogen as electron donor to a biofilm, and (2) the oxygen/air-based MBfR (Syron & Casey 2008), which delivers oxygen as electron acceptor to the biofilm. The latter is also known as the membrane aerated biofilm reactor (Martin & Nerenberg 2012). Hydrogen-based MBfRs have been demonstrated viable for the biochemical transformation of nitrate, nitrite, perchlorate, bromate, selenate/selenite, arsenate, and chromate to name only some. As the MBfR allows for a higher control of electron donor/acceptor delivery, biofilms with defined or strong redox stratification can be developed for a simultaneous oxic/anoxic process such as nitritation/anammox. Commercially available MBfRs exist. The membrane aerated biofilm reactor may be procured in North America as the ZeeLung™ process (Côte et al. 2015), and in Ireland as the OxyMem™ process. These processes are well suited for combined carbon oxidation and nitrification, nitrification, denitrification, partial nitritation and deammonification. A single unit demonstrating the ZeeLung™ system exists, treating approximately 2,300 p.e. for tertiary nitrification at the O'Brien Water Reclamation Plant, Chicago, Illinois, USA. At least nine full-scale OxyMem processes exist, collectively treating a ranging of flows and meeting a diverse array of treatment objectives throughout Japan, Sweden, Spain, United Kingdom, Ireland, and Brazil.

The continuous enhancement to and implementation of new water quality regulations, and the discovery of new processes, have made mature biofilm reactor types relevant to current trends and challenges that face this community. For example, the US Environmental Protection Agency has enacted a primary drinking water standard that requires selenium concentrations to be less than 0.05 mg/L. This regulation has impacted the agriculture, mining, and power (coal and oil) industries, to name a few. The use of expensive reagents and the production of hazardous residues make the use of physicochemical treatment impractical. As a result, the biological transformation of selenate and selenite to elemental selenium is preferred. Biofilm reactors capable of operating under anaerobic conditions are required; hence particulate biofilm reactors, as described by Nicolella et al. (2000), are of renewed interest. Similarly, processes such as SANI (Wang et al. 2009) and DEAMOX (Kalyuzhnyi et al. 2006) have made use of biofilm reactors such as deep-bed filters and upflow anaerobic sludge blankets.

Finally, biofilms have recently been thoroughly investigated for their capacity to biologically generate electricity, the so-called microbial fuel cell (MFC). Liu et al. (2004) demonstrated that MFCs can produce electricity while biologically converting complex compounds present in municipal wastewater. There are several different means for constructing an MFC. Logan et al. (2006) presented means for constructing MFCs, compared devices on an equivalent basis, and reviewed an array of related scientific principles, ranging from environmental engineering to microbiology and electrochemistry. The creation of an MFC that can yield sufficient electrical output for economically viable production and utilization eludes researchers, and remains a challenge for biofilm scientists and engineers. MFCs, when operated in the electrolysis mode as microbial electrolysis cells, can produce useful chemical products such as hydrogen and ethanol (Zhou et al. 2013).

The deleterious role of biofilms on membranes is also an area of concern to process designers and biofilm researchers. Membrane biofouling is a costly operational concern, for example as a feed-spacer problem in spiral-wound membranes (Vrouwenvelder et al. 2009). The role that quorum sensing plays in dispersing biofilms has led biofilm researchers to seek membrane biofouling control measures via quorum sensing (Yeon et al. 2009). Alternatively, Vrouwenvelder et al. (2011) presented a scenario for controlling spiral-wound membrane biofouling by reducing flow pace, modified feed-spacer design, and an advanced cleaning strategy. Another approach to dealing with undesired biofilms that grow on membranes is to tolerate their existence, and focus on increasing hydraulic conductivity of the growing biofilms rather than trying to prevent their formation; ultimately one may benefit from the biological activity in a biofilm to improve permeate quality (Chomiak et al. 2015).

Biofilm models are essential both to the study and development of fundamental biofilm research and the development and implementation of biofilm reactors (Morgenroth et al. 2000). A consensus description and comparison of biofilm models was presented by Wanner et al. (2006). This effort led to the widespread development and application of one-dimensional biofilm models as an engineering tool (Boltz et al. 2010). Nevertheless, multi-dimensional models (e.g., Picioreanu et al. 2004) have enhanced virtually every form of biofilm research and system development. A clear dichotomy has existed between the use of biofilm models as a research resource and the more recent use as an engineering tool. Bioreactor hydrodynamics has a substantial influence on the degree of uncertainty that is affiliated with the use of mechanistic biofilm models to describe a biofilm reactor (Boltz & Daigger 2010). Therefore, accounting for the importance of bulk-liquid hydrodynamics and system idiosyncrasies (e.g., biofilm carrier type and transport) via simulation has become an integral consideration for biofilm and biofilm reactor modelers (Kagawa et al. 2015; Boltz et al. (submitted)). Biofilm models have become an increasingly important tool for biofilm researchers and biofilm reactor designers who are interested in the most relevant topics in environmental biotechnology, including greenhouse gas emissions (Van Hulle et al. 2012; Sabba et al. 2017), phototrophic biofilms (Wolf et al. 2007), biofouling in membrane separation systems (Radu et al. 2015) and MFCs (Picioreanu et al. 2007).

Fundamental principles describing biofilms exist as a result of focused research, practical application, and modeling. The use of reactors for the treatment of municipal and industrial wastewaters is a common beneficial use of biofilms. Applied research exists that provides a basis for the mechanistic understanding of biofilm systems. The empirical information derived from such applied research has been used to develop design criteria for biofilm reactors and remains the basis for the design of many biofilm reactor types despite the emergence of mathematical models as reliable tools for research and practice. There is a gap between our current understanding of biofilm fundamentals and reactor-scale empirical information, represented by the dichotomy in the literature between our knowledge about and use at the micro-scale (biofilm) and macro-scale (reactor). Lewandowski & Boltz (2011) highlighted this division by describing state-of-the-art basic research and practice oriented beneficial use of biofilm systems for the sanitation of water.

The contributions of the following colleagues to both our knowledge of biofilms and biofilm reactors and the contents of this review are gratefully acknowledge: Jeff Peeters, General Electric Water Division; Eoin Syron, OxyMem; Kim Sørensen, WABAG; Robert Nerenberg, University of Notre Dame, Hallvard Odegaard, NTU. This manuscript is based on a contribution by the authors in the IWA report Global Trends & Challenges in Water Science, Research and Managementand is published with IWA's permission.