Photogranules have been formed in a range of laboratories under a variety of environmental conditions over the last years. Even though many aspects in the life cycle of photogranules remain poorly understood and crucial differences exists between studies, there are several commonalities that have been documented in the literature to favour photogranulation. In the following section, we synthesize these factors.

4.1 Introducing or producing an initial phototrophic community

Frequently in the laboratory, but possibly also at the industrial scale, a new generation of photogranules needs to be de-novo generated on-site from a starting material. At the laboratory scale, the starting material is in many cases activated sludge or aerobic granular sludge, both of which do not typically contain a substantial phototrophic community.

4.1.1 Achieving a critical mass of suitable phototrophs

In this review, we have so far only considered the production of hydrodynamic photogranules, i.e., photogranules that are produced in the hydrodynamic environment of a photobioreactor. It is also possible to produce photogranules in a hydrostatic environment (Milferstedt et al. 2017a). The production in the absence of hydrodynamic shear and washout is unique among the ways of generating granular sludge. An unconsolidated matrix, e.g., activated sludge, often contained in a small, closed environment of several millilitres can form hydrostatic photogranules, but this does not always succeed (Joosten et al. 2023).

Common to both photogranulation approaches appears to be the requirement of a critical mass of phototrophic microorganisms to initiate photogranulation. In the literature, two strategies are typically followed to obtain this critical mass (Fig. 3).

Fig. 3

Strategy for production of a critical mass of phototrophic microorganisms required for photogranule formation. The critical mass of phototrophs can be reached by augmenting the starting matrix with phototrophs or by enriching phototrophs under the appropriate conditions

The first strategy consists of forming photogranules directly from activated sludge or aerobic granules (Zhang et al. 2018; Zhong et al. 2024). To do so, it is necessary to enrich the small number of phototrophs present in the starting matrix to a sufficient quantity. It was demonstrated that incubation of the same activated sludge in an SBR and a photo-SBR can respectively yield both aerobic granules and photogranules (Sales et al. 2022).

Alternatively, the critical mass of phototrophs can be obtained by adding directly to the starting matrix cyanobacterial isolates or previously formed photogranules from hydrodynamic or hydrostatic incubations (Zhang et al. 2022; Sales et al. 2022; Wang et al. 2022).

Hydrostatic photogranules used for inoculation are typically about 1 cm large and are often reported to disintegrate rapidly under stirring when they are used to inoculate a reactor under a hydrodynamic regime (Milferstedt et al. 2017b; Abouhend et al. 2018). The rapid decay suggests that the hydrostatic photogranules themselves do not produce offspring photogranules, but that the decayed photogranules may supply the critical mass of microorganisms for photogranulation. Augmentation with phototrophic cultures may also be followed by a subsequent enrichment phase.

When following the enrichment strategy for hydrostatic photogranules, i.e., using a starting matrix without the addition of a phototrophic inoculum, aliquots of the seemingly identical matrix may develop into photosynthetic aggregates with different morphologies such as microbial mats or photogranules. Between replicates, the morphotype may also vary, leading to a distribution of morphotypes. It was shown that in hydrostatic cultivations, this distribution can be changed by adding specific cyanobacteria (Joosten et al. 2023). The addition of sufficient amounts of specific cyanobacteria will steer morphology formation towards the photogranule morphotype. The addition of phototrophs has the tendency to homogenize the results of hydrostatic incubations.

Adding specific photogranule-forming cyanobacterial strains to an activated sludge matrix could favour the formation of photogranules. However, excessive addition of a non-photogranulating cyanobacteria strain in the inoculum may prevent photogranulation entirely (Joosten et al. 2023). The resulting morphotype may be a property inherent to the cyanobacterial strain that was added.

Researchers were able to narrow in on some specific organisms capable of promoting photogranulation both under hydrostatic and hydrodynamic conditions (Zhu et al. 2023; Joosten et al. 2023; Kong et al. 2023a). The ecosystem engineers responsible for the formation of photogranules frequently appear to be filamentous cyanobacteria (but not always) belonging to the order Oscillatoriales. However, not all Oscillatoriales are able to form photogranules (Joosten et al. 2023).

Moreover, it was found that these ecosystem engineers may be absent from the mature photogranules (Joosten et al. 2023). Its transitory presence and its absence at a later stage during the hydrostatic development reveals the transient presence and activity of the ecosystem engineer as an important step in the life cycle of a photogranule. Should additional research support this finding, mature hydrostatic photogranules may turn out not to be the optimal choice for inoculation, given their potentially lower concentration of the desired cyanobacteria.

Therefore, the importance of phototrophic microorganisms for the formation of photogranules has been demonstrated, not only as mere oxygen providers but also as critical ecosystem engineers. The full mechanism behind the life cycle of photogranules remains however unclear.

4.1.2 The complexity of providing light

Light is obviously a key parameter for photogranulation. Although providing light to grow phototrophic populations may seem trivial at first, there is a myriad of ways to do so and of parameters to take into account, resulting in great complexity. We consider here three aspects of illumination: light intensity, light spectrum and alternating dark/light phases.

Light intensity is usually measured using illuminance and provided in Photosynthetic Photon Flux Density (PPFD), which corresponds to the amount of Photosynthetically Active Radiation (PAR) reaching a surface. PAR is defined as the fraction of the light spectrum usable by phototrophic organisms for photosynthesis and is usually defined as the range between 400 and 700 nm. Note that the definition of PAR may differ according to the organisms in question, e.g., the light requirement for purple bacteria is in the infrared part of the spectrum. Their PAR is thus in the wavelength range beyond 900 nm. PPFD is typically measured using so-called PAR meters and is expressed in µmol photons∙m−2∙s−1.

Measuring light intensity using irradiance (e.g., W∙m−2) rather than illuminance offers a more direct relation to energy, but risks hindering comparison between studies. In fact, a conversion between irradiance and illuminance units requires that the spectrum of the light that is used is provided. This is necessary to exclude photons that are outside of the usable wavelength range, and to correct for the different amounts of energy by wavelength. We discourage units based on human perception of light such as lux, as they do not give information on the light usable by phototrophs.

It should be mentioned that, while most studies specify the light intensity used for their experimental setup, this value is only meaningful when location and details about the measurement are also provided. Considering that light intensity decreases proportionally to the square of the distance from its source, even distances that seem small in a complex reactor setup can induce significant measurement errors. The actual light intensity in the reactor will depend on many factors such as biomass concentration, medium turbidity, presence of a biofilm on the reactor wall and reactor wall thickness. For maximum accuracy, we recommend measuring light intensity on reactor walls, or directly within the reactor when possible and to at least specify where, e.g., in water or air, and how the measurement was performed. This is especially important considering that even rather low differences in incoming light intensity, such as 30 µmol∙m−2∙s−1, can drastically change the microbial composition in photogranules (Liu et al. 2023). Differences on this order of magnitude can be easily produced, for example by stray daylight through a window.

Photogranules were reported to be successfully formed using a wide range of light intensities, from 18 µmol m−2∙s−1 (Ansari et al. 2021) up to 500 µmol∙m−2∙s−1 (Trebuch et al. 2020). Most studies were conducted within a range of 100 to 250 µmol∙m−2∙s−1. A combination of light intensity of 450 µmol∙m−2∙s−1 and strong stirring was shown to hinder the formation of photogranules and favour growth of microalgae over cyanobacteria (Gikonyo et al. 2021a).

Lowering the light intensity from 150 to 100 µmol∙m−2∙s−1 reportedly favoured photogranulation and enhanced the growth of filamentous cyanobacteria over filamentous microalgae (Abouhend et al. 2018). Contrarily, under mechanically aerated SBR conditions, a light intensity of 140 µmol∙m−2∙s−1 favoured photogranulation compared to 80 and 110 µmol∙m−2∙s−1 (Liu et al. 2023). Overall, it seems that no consensus has been reached regarding optimal light intensity. In several cases, the way light intensity is measured is not detailed, which further complicates comparison of values between studies.

Light intensity influences the microbial composition and activity of the phototrophic layer of photogranules, e.g., low light intensities tend to favour development of cyanobacteria compared to microalgae. Higher biomass growth rate, bigger granule size, stronger oxygen production as well as significant differences in microbial compositions were reported with stronger light in photogranules formed in SBR reactors operated at light intensities between 45 and 225 µmol∙m−2∙s−1 (Meng et al. 2019b).

Photogranules in a hydrodynamic reactor environment regularly experience changes in the exposure to light. These changes occur at different time scales. Alternating phases of illumination and darkness on the order of several minutes to hours are often implemented in the cyclic operation of sequencing batch photobioreactors. This operation mode is very common as light is switched off to stop oxygen production and to enhance anoxic and anaerobic bioprocesses. At the same time, moving photogranules in a laboratory reactor environment experience variable illumination at a temporal scale of less than a second through self-shading and on the order of seconds by changing distances to the light source (Fig. 4). Phases of darkness help relieving phototrophic microorganisms of photoinduced stress and allow recovery and maintenance of their photosystems.

Fig. 4

Time scales of light exposure that photogranules may experience in a reactor set-up. a Through the movement inside the photoreactor, photogranules may be shielded from the light source for short amounts of times (self-shading). b On their path through the photoreactor, a photogranule may be farther away from the light source for several seconds. c A photoreactor may experience times of darkness during the operation on the order of minutes to hours

Alternating light/dark phases, at the scale of minutes, has been reported to significantly influence photoefficiency, biofilm thickness and geometry of phototrophic biofilms (Gao et al. 2023). Therefore, it is possible that alternating light phases influence photogranule growth and operation. Dedicated studies at different time scales of alternations may provide the required insight.

Light can be considered one of the substrates of a photobioreactor. Analogies to bioreactors that are fed chemical substrates can therefore be drawn. In the case of aerobic granular sludge, granulation was greatly influenced whether the reactor is fed while being aerated or not. Indeed, feeding in the absence of oxygen promoted the growth of phosphate and glycogen accumulating organisms which greatly improved settleability, density as well as phosphate removal capacity of the biomass (Haaksman et al. 2020). It is unknown whether this holds significance to the formation of photogranules. However, it is possible that light availability during feeding (i.e., feeding under aerobic or anoxic conditions) is important for promoting the growth of microorganisms such as phosphate-accumulating organisms. This is likely to influence pollutant removal. At this time, it is however unclear whether it also affects photogranule formation.

While the large majority of available studies focuses on artificial lighting, hydrostatic as well as hydrodynamic OPGs can also be generated using sunlight (Park and Dolan 2015; Trebuch et al. 2023c). Formation of photogranules and operation of photogranular reactors using only sunlight remains an interesting option with a low environmental impact.

4.1.3 Initial batch operation ensures phototrophic enrichment?

The fastest conversion of activated sludge into photogranules, currently reported in the literature, occurred in eight days (Gikonyo et al. 2021a). This was achieved using activated sludge under hydrodynamic batch conditions. However, the reported SVI and particle size distribution suggest that only part of the biomass was turned into photogranules at the time. Photogranulation was also achieved in less than two weeks in SBR with a 3 days fed-batch phase after inoculation with hydrostatically formed photogranules (Ansari et al. 2019).

It is important to note that studies mentioning the time required to achieve photogranulation often do not specify the criteria used to determine when the sludge was considered photogranulated. Therefore, comparing the time required for the formation of photogranules across studies may be somewhat subjective. However, other studies that do not include a batch or relatively long fed-batch phase commonly reported significantly longer formation times, usually by one or a few weeks.

While the choice of a specific start-up strategy is typically not explained in a publication, a batch or fed-batch phase gives a competitive advantage to the phototrophic community by providing light as sole source of externally provided energy and prevents the loss of potentially critical microorganisms through washout.

Therefore, and although no dedicated research was conducted on the matter, there is an interest in investigating different start-up strategies to speed-up and enhance the reliability of photogranulation.

4.1.4 Initial sludge concentration linked to abundance of phototrophs?

The initial concentration of activated sludge, when used as a starting matrix to form photogranules, has been reported to influence photogranulation dynamics as well as the properties of the resulting photogranules.

At light intensities between 117 and 450 µmol∙m−2∙s−1, formation of photogranules failed with initial activated sludge concentration above 3900 mgMLSS∙L−1 but succeeded at about 2650 mgMLSS∙L−1 (Gikonyo et al. 2021a). Lower biomass yield, lower EPS production and worse settleability were reported at initial concentrations of 1100 mgTSS∙L−1 compared to 900 mgTSS∙L−1 when inoculated with hydrostatically formed photogranules at a light intensity of 150 µmol∙m−2∙s−1 (Ansari et al. 2019). Therefore, even small differences in the initial concentration could possibly have an impact on photogranulation.

These results imply the existence of an optimal sludge concentration for starting-up a photogranule reactor. This concentration probably depends on light supply and inoculum characteristics, such as concentration and diversity of the initial phototrophic community. Therefore, it is possible that a too low seeding concentration fails to form photogranules because it lacks critical phototrophs. The optimal sludge concentration may therefore be the compromise between a sufficient concentration of phototrophic microorganisms and limited light penetration due to high suspended solids concentration. The inoculum concentration is not often provided in the literature and seems to be generally neglected when starting a photogranular reactor.

4.1.5 Impact of macronutrients on photogranule formation

Under hydrostatic conditions (see Sect. 4.1.1), photogranulation success was documented to be influenced by the initial concentration of dissolved inorganic nitrogen. Photogranulation had a higher success rate with ammonium concentrations above 12 mg-N·L−1. In this study, absence of ammonium in the initial mixture led to the failure of granulation. The failed cultivations were defined by low internal cohesiveness and disintegration of the aggregates after gentle stirring (Stauch-White et al. 2017).

Unsuccessful cultures exhibited significantly higher concentrations of chlorophyll-b, suggesting a greater prevalence of eukaryotic microalgae in these failed cultivations. It was suggested that the initial concentration of inorganic nitrogen may have favoured the growth of cyanobacteria over microalgae (Stauch-White et al. 2017). However, in another study, successful hydrostatic photogranulation was negatively correlated with initial ammonium concentration and strongly positively correlated with the initial nitrate concentration (Castro et al. 2024). Under hydrodynamic conditions, urea as nitrogen source was reported to lead to thicker biofilms compared to ammonium and nitrate. The nitrogen source also significantly influenced the composition of EPS and the microbial composition of photogranules (Li et al. 2024b).

Certain cyanobacteria can perform atmospheric nitrogen fixation. They can be found and cultivated in photogranules, and can provide the aggregates the ability to sustain growth in a nitrogen-deficient environment (Trebuch et al. 2023b). However, under hydrodynamic conditions, even in the presence of nitrogen-fixing cyanobacteria, nitrogen deficiency reportedly led to irregular and loose photogranule morphologies. Some granules developed intragranular void areas, but this did not influence their physical stability nor settleability (Trebuch et al. 2023b). A sudden 50% reduction of the ammonium concentration initially caused partial degranulation, followed by an increased growth of hydrodynamic photogranules (Sales et al. 2022). Photogranules can, however, withstand low carbon content in the influent and maintain their stability and settleability at C: N ratios as low as 1, despite reduced biomass growth and EPS production (Zhao et al. 2018).

The carbon type can also influence photogranulation. Starting from activated sludge, formation of hydrodynamic photogranules failed when feeding only inorganic carbon as carbon source. Cultivations fed with the same total carbon concentrations, but with 50% or 100% organic carbon successfully formed photogranules. Photogranules fed with only organic carbon showed increased biomass growth, increased EPS production and growth of cyanobacteria, better settleability and reduced effluent turbidity, as well as a denser structure (Li et al. 2024a).

Therefore, nitrogen source and concentration has shown to influence the properties of photogranules and their key microorganisms. Despite a potentially strong importance, the literature lacks sufficient data to fully understand the impact of nitrogen on photogranulation, particularly regarding the effects of ammonium, nitrite and nitrate concentrations on the formation and stability of photogranules.

4.2 Providing a favourable environment for photogranulation4.2.1 Sequencing batch reactors as a suitable environment for photogranulation

Sequencing batch reactors (SBR) are the most popular reactor operating condition to cultivate biological aggregates, including photogranules.

The SBR operating mode easily allows uncoupling of hydraulic retention time from sludge retention time while sorting particles based on their settleability. It is operationally easy in an SBR to exert a strong selection pressure towards granulation by applying short settling times. This selection pressure via settleability is believed to be critical for the formation of granules (see Sect. 4.2.2). It is also easy to operate the SBR under alternating feast and famine conditions, which is regarded as critical for the formation of aerobic granular sludge (Liu et al. 2016). Whether the application of feast and famine conditions is beneficial for the formation of photogranules has however not been reported to date.

The SBR operation is known to favour the growth of filamentous microorganisms (Liu and Liu 2006), possibly including filamentous cyanobacteria, some of which are believed to drive photogranulation (Milferstedt et al. 2017b; Kong et al. 2023a).

Formation of photogranules has been achieved in various reactor setups other than SBR (Table 3). Maintaining these granules over time may however require additional efforts, for instance to keep a sufficiently strong selection pressure on settleability. It is possible to maintain a strong selection pressure in a continuous stirred tank reactor (CSTR), but this requires using an external or internal settler (Gikonyo et al. 2023).

Table 3 Phytoplankton in vivo main absorption and fluorescence bands, extracted from Seppälä (2003)

While using an SBR is a convenient option, other reactor types such as CSTR can provide many advantages, especially in full-scale applications. Indeed, CSTR are easier to operate and control than SBR. Filling and draining steps require strong pumping systems and unless multiple SBR are operated at the same time, the volume that can be treated at a time is limited (Kent et al. 2018). Retrofitting current infrastructure, most often CSTR systems, would also be significantly easier if photogranulation was mastered in CSTR. Other reactor types, such as high-rate algal ponds may be a better fit if sunlight is intended as the main light source. Thus, even if SBR is currently by far the most popular reactor type and operation for photogranulation, there is an interest in investigating photogranulation in other reactor types.

4.2.2 Washout of suspended biomass as selection pressure

Any condition that favours specific microbial populations qualifies as a selection pressure. In the literature, forcing non-settling biomass out of the system and retaining only settling aggregates is described as one of the most crucial selection pressures towards granulation in the aerobic granular sludge technology (Liu et al. 2005; Kent et al. 2018). Even though hydrostatic photogranules may form in an environment without washout, settling may be a strong selection pressure in the case of the formation of hydrodynamic photogranules. Competing planktonic biomass is removed because it settles much slower than photogranules, which favours the selection and growth of the microbial communities in fast-settling aggregates. Forming photogranules and maintaining them over time requires that slow-settling biomass is removed from the system. This can be achieved using gravity settling (Fig. 5a). Once photogranular biomass is established, it may be desirable to select a specific range of photogranule size to optimize a specific biological activity and biomass yield (Abouhend et al. 2020). While this kind of photogranule selection is currently not documented in the literature, the use of sieves or similar equipment may also be useful to retain a target photogranule size within the photoreactor (Fig. 5b).

Fig. 5

Biomass selection strategies to retain photogranular biomass in a system. a Gravity settling retains photogranules based on their density and morphology. It is the most commonly used strategy to form and maintain photogranules. The time allowed for settling determines which particle classes are retained in the system. b Retention of intermediate particle size classes may be achieved by sieving. This strategy may be applied to retain photogranules with a desirable biological activity

In an SBR setup, selection pressure by settling can be applied by using short settling times at the end of the cycles when stirring is stopped and before partially emptying the reactor. After settling, the treated water is withdrawn, respecting the volume exchange ratio, i.e., the percentage of bulk liquid that is removed at the end of the SBR cycle.

In practice, enforcing the selection pressure through a short settling time requires that the necessary time to remove the treated water from the reactor is much shorter than the time for settling, so that no additional partial settling occurs during the withdrawal time. This may be difficult to achieve for larger reactors, for very short settling time (less than a few minutes) and for high volume exchange ratios, as typical lab-scale pumps may not be fast enough.

Volume exchange ratio and cycle duration determine the hydraulic retention time of the system. As these factors are not independent, modifying any of these parameters will result in the application of a different selection pressure. Volume exchange ratio and hydraulic retention time are also directly linked with other important parameters such as the organic loading rate. For this reason, to strengthen or weaken selection pressure, it may be wise to rely on settling time alone.

Shorter hydraulic retention times were reported to lead to an increase in selection pressure, which accelerated photogranulation, improved biomass settleability and productivity in SBR under aerated conditions (Trebuch et al. 2020). As discussed above, however, it is possible that part of these observations was due to the higher organic loading rate induced by shorter hydraulic retention time (Table 4).

Table 4 Examples of reported photogranulation success under widely different experimental conditions

Keeping biomass too long in the system can lead to instability. In fact, solids retention time (SRT) influences photogranulation and properties such as the size distribution, settleability, EPS production and the microbial community composition of the photogranules (Buitrón and Coronado-Apodaca 2022; Zhi et al. 2023b). These authors recommend keeping SRT between 10 and 12 days. The absence of biomass wastage (i.e., no regulation of the solids retention time) caused a negative shift in photogranule size distribution, suggesting partial degranulation (Ansari et al. 2019). An uncontrolled solids retention time is known to be responsible for the development of filamentous outgrowth and poor settleability in aerobic granular sludge (Liu and Liu 2006).

Biomass selection therefore largely influences the formation and the properties of photogranules. The extent of this influence, especially in the case of settling-based selection pressure, remains unknown, but warrants further, dedicated research.

4.2.3 The role of shear in photogranulation

Hydrodynamic shear stress through stirring or aeration is commonly described as one of the most crucial parameters to generate aerobic and anaerobic granular sludge (Liu and Tay