Substrate’s composition effect on methane production
The obtained TS values shown in Table 1 indicate that the seaweed residues lost moisture content during the extraction process, due to osmotic gradients of solvents used in the extraction itself and its related chemical reactions. In fact, it is well known that algal biomass exhibits very high levels of moisture content. Water content in seaweeds ranges between 78 and 90% (Burton et al. 2009; Marinho-Soriano et al. 2006) depending on species and period of harvesting, while in this study it was found to be between 65 and 80%, thus resulting in higher TS% w/w compared to fresh biomass. The VS fraction of the residues remained very high (above 70%) for all species despite the extraction processes, with values in line with literature for un-extracted seaweed (Horn 2009), suggesting high biogas potential. Total COD (tCOD) concentrations found in the residues are in line with those obtained by (Nkemka and Murto 2010) from seaweed leachate. However, they are up to twofold below values reported by (Gurung et al. 2012), where raw seaweeds were used. The explanation of this behaviour is the composition of the sample, which may consistently vary across seasons and among seaweed species. Also, sample preparation procedure has a significant effect, i.e. organic solids concentration in the batch reactors can be selected arbitrarily for testing or be based on previous experience. In fact, sample settings of this study are closer to trials set up by (Nkemka et al. 2010), in terms of solids concentration and biomass harvesting period (September). This result is encouraging, as tCOD values from raw seaweed are in range with those from the residues under investigation, confirming that a valuable amount of organic matter is present.
Methane production is known to be positively correlated to the contents of carbon and hydrogen, while being negatively related to the oxygen content. The high percentages of hydrogen and sulphur (Table 2) already indicate that formation of corrosive nitrogen and sulphur containing emissions is likely to occur. In Table 2, the content of carbon and hydrogen suggests methane yields close to those obtainable from starch (415 mL gVS−1) (Angelidaki and Ellegaard 2004). High content of nitrogen and sulphur may also lead to inhibition of the methanogenic phase, besides forming toxic gases such as ammonia and hydrogen sulfide.
C/N ratios ranging from 20 to 30 are considered optimal for AD, as if this ratio is very low nitrogen will be released and accumulated in the form of ammonium ion (NH4
+) (Chandra et al. 2011). As it can be observed in Table 2, residues of FS, FV, AN, LD are perfectly in range of the ideal conditions, while digesters containing LH and UR residues should be complemented with another waste to balance the C/N ratio. Nitrogen rich and carbon rich substrates should be, respectively added to the seaweed digester containing LH and UR.
Theoretical and effective methane yields
Results of the fermentation experiment are reported in Table 3. Tedesco et al. (Tedesco et al. 2013, 2014a, b; Vanegas and Bartlett 2013a, b) reported biogas and methane yields from L. digitata and hyperborea, Ulva and Fucus species in Ireland. The mentioned studies will serve as a comparison to the biogas yields obtained in this investigation, as they were conducted on fresh un-extracted seaweed biomass.
Biogas yields obtained from Fucus spp. residues (about 100 mL CH4 gVS−1) were found slightly above the highest values obtained by (Tedesco et al. 2013) conducted for 21 days in which the inoculum-to-substrate ratio (ISR) was 1:1 w/w and the biomass-to-water mass ratio was 1:20 w/w. The amount of TS in the reactors was thus doubled in this study’s fermentation, while the volume of inoculum was instead one third of that used in this experiment. Beside the different ISR settings between (Tedesco et al. 2013) and this study, seasonal variation in biochemical composition has also an influence, as methane yields are meant to be higher in warm seasons and lower in the cold ones. This is due to higher carbohydrates stored by the plant during spring and summer. Biomass composition being equal, the larger the initial amount of inoculum in the reactors and the faster the digestion occurs. In general, considering the different ISRs and biochemical seasonal variation, these studies would suggest that Fucus spp. residues should maintain their biogas yield around 200–230 mL gVS−1 for digestion at mesophilic temperatures throughout the year.
Biogas yields from L. digitata and hyperborea are in line with average biogas and methane yields obtained by Tedesco et al. (Tedesco et al. 2014a, b). Lower yields were obtained by (Vanegas and Bartlett 2013a, b) for LD, FS and Ulva spp. than those achieved in this study. However, this could be attributed to the use of bovine slurry as inoculum, which creates an increasingly acid environment for the bacteria due to accumulation of volatile fatty acids (VFAs) in the bioreactors. In fact, Vanegas et al. had to adjust the pH within the reactors by adding a mixture of NaOH and 10% KHCO3 to the reactors to re-establish the ideal pH range and boost the buffer capacity of the system. It should be pointed out that previous investigations carried out by the authors and the literature (Chynoweth et al. 1981; Hanssen et al. 1987) found the use of sludge as inoculum to actively contribute to the digestion stability and contrast VFAs accumulation followed pH decrease, as this has happened frequently when digesting seaweeds. It can be concluded that sludge should be considered as first inoculum/co-substrate option when planning a seaweed digesting facility.
Finally, previous unpublished work from the authors identified a biogas yield of 336 mL gVS−1 from Ascophyllum nodosum (March), while Hanssen et al. (Hanssen et al. 1987) found a yield of 280 mL gVS−1 (September). These values are 30–42% higher than those from the AN residues from this study. Furthermore, research conducted in Scotland (Edyvean et al. 1988) reported biogas yields up to 376–360 mL gVS−1 from AN residues after alginate extraction only. The low biogas yield obtained from AN residues in this investigation can be attributed to some extent to the multiple extractions performed, which are not limited to alginic acid, and are, therefore conceptually coherent. The biochemical composition of AN has been found by the literature to be much less affected by seasonal variation compared to most brown seaweeds (Black 1948). Biogas and methane yields from AN could so be considered rather stable during the year and relied upon, other conditions being equal. However, this is currently being experimentally verified for AN residues following biorefinery extractions in a broader investigation.
Interestingly, the overall methane content (Table 3) in the biogas that ranged between 40 and 45% for all species. Such a small variation is certainly connected to the fact that the C/N ratio was set to the same value in all reactors. This highlights the importance and the role played by this parameter in the methanation process. Values of NH3 and H2S were very high and even higher values are envisaged at plant scale. This would either slow down or prevent methane formation, as excessive NH3 and H2S concentrations are toxic to methanogens. Furthermore, these organic acid gases are corrosive to pipes and engines, as well as dangerous to human health in the detected concentrations. Consequently, the biogas obtained should be cleared up before use in combined heat and power (CHP) units or their formation should be chemically prevented via the addition of iron-based chemicals to the digester (Streefland et al. 2010).
The SMP and actual methane yields from the analysed residues were compared with theoretical yields from existing literature in Ireland (SPM*) from un-extracted seaweed in Fig. 1, while Fig. 2 shows the cumulative biogas yields achieved across the digestion period. It can be noticed that the derived SPM yields are close to those theoretically achievable from most anaerobically digested lignocelluloses such as maize (420 mL gVS−1) and Miscanthus (488 mL gVS−1), as reported by (Lübken et al. 2010). Furthermore, most theoretical yields are in line with those achieved by a very recent study conducted in Ireland during the summer 2013 on the coasts of Cork (Allen et al. 2015). However, practical yields are considerably below the theoretical maximum as only a fraction of VS is normally destroyed and transformed into methane by fermenting microorganisms. The volatile solids reduction is generally associated to the recalcitrance of the substrate to biological degradation, as it clearly shows in Fig. 2. In this study, however, such a result also partially reflects the short digestion time used, i.e. 15 days.
Actual methane yields from AN and UR residues were the lowest obtained. AN is known to contain polyphenols which are difficult to degrade and can inhibit anaerobic digestion. Hence, AN residues would produce more methane when pre-treated with formaldehyde (Horn 2009; Vanegas and Bartlett 2013a), which would eliminate issues associated with polyphenols poisoning. Ulva spp. conversion to methane is believed to have suffered from excessive concentrations of dissolved NH3, which developed an inhibition. In fact, UR was the substrate with the highest nitrogen content (Table 2). This is also demonstrated by extremely low gaseous ammonia concentration in the resulting biogas, i.e. 20 ppm, which indicates that most of the nitrogen remained dissolved in the substrate, and had a toxic effect on methanogens. This suggests that other substrates different from sludge and much poorer in nitrogen should be tested for co-digestion with this particular residue, with a C/N above 25. Furthermore, such low methane yields can also be attributed to low tCOD value in the bioreactor.
Laminaria spp. residues exhibit the highest potential for methane production, with a yield of 187–195 mL CH4 gVS−1. However, these substrates realised only 43% of their potential, while an average of about 20% of the theoretical yield was produced by the other species. The actual methane yields shown in Fig. 1 were obtained with no pre-treatment and by fermenting the residues after coarse chopping and initial pH adjustments only. A variety of pre-treatments have been proven to be very effective at enhancing the methane conversion efficiency from algal biomass, but they would increase the operational cost of the anaerobic digestion plant, sometimes making the whole process not economically viable as most pre-treatments are energy intensive or impact the process sustainability due to the use of chemicals (Jard et al. 2013). However, as the structural integrity of the seaweed is preserved after the compounds extraction (residues provided were not reduced in particle size), an initial milling/shredding of the residues will be indispensable to the automation of the process and to maximise the surface contact with bacteria within the digester (Bernat et al. 2015).
AN, Fucus spp. and UR residues performed very poorly compared to their methane potential (BI 17–22%). Technically there is a large room for improvement of the BI, and consequently of the methane yields. Previous work (Tedesco et al. 2014a) has indicated that for example, mechanical comminution of the biomass can enhance the methane yields from Laminaria spp. up to 53% (290 mL CH4 gVS−1), and unpublished work indicated that ground AN can reach up to 170 mL CH4 gVS−1 around the end of August.
Ulva spp. are abundant among the so-called ‘drift seaweeds’. These species’ plant structure is very fragile and storms are able to detach them from their roots, often causing green tides in Ireland (Allen et al. 2013). Given the low methane yield of UR residues and the general uncertain availability of the stock whether farmed or wild harvested, it would not be recommendable to rely on this substrate to produce gaseous fuel. Fucus, Laminaria and Aschophyllum spp. are sub-tidal plants more tightly attached to the rocks of the sea floor. If in the future, seaweed will be cultivated in Ireland for food, bioactive compounds and energy extraction, the latter mentioned species (FS, FV, LD, LH and AN) hold the highest potential for all these applications.