Reactor design for wastewater biorefineries
WASTEWATER BIOREFINERIES: RECOVERING VALUE WHILE PRODUCING CLEANER WATER
Reactor design for wastewater biorefineries
Presented at the 10th edition of the Renewable Resources and Biorefineries conference, hosted in Valladolid, Spain, from 4 - 6 June 2014. Website: rrbconference.com
Today I want to tell you about bioreactor design for wastewater biorefineries. Immediately, two questions should pop up in your brain:
- Why on earth would we want to use wastewater, something that is typically very dilute, and highly variable, not to mention often dangerous to our health or just smelly, to produce commodity products, aka to make a profit? And;
- What makes reactors the critical point in this discussion?
First, let me tell you what we're dealing with, to set things into context. We are considering wastewaters, ranging from complex municipal wastewaters, to a variety of industrial wastewater sources that may be more defined.
We can see from the data on this slide that wastewaters can be grouped according to three factors: volume, concentration, and complexity. For the most part these waters have huge flows, in the order of mega liters every day, and can be quite dilute, with the most common components in the order of milligrams per liter. What makes them hard to deal with is the level of complexity - these waters mostly tend to be highly variable, changing concentration and perhaps also composition the whole time, and they tend to be 'receptacles', meaning that the compounds that make their way into the water is not controlled, so you can get all sorts of heavy metals or toxic chemicals in these waters. Just as these are poisonous to us, they can possibly wreck havoc with microbial populations in your bioprocess as well. Wastewater biorefineries involve the recovery of valuable products, including water and nutrients, from wastewater as an integrated system rather than a unit process, and potentially provide a link between the users of water and those responsible for its management where resources are recovered in closed loop cycles.
So why bother?
Commodity bioproducts from renewable resources are often not economically competitive, and the challenge is threefold:
- The costs of the raw material, which could account for up to 80% of the total cost. A lot of research, as we can see at this conference, is about using wastes as 'free' raw material, but because of the suboptimal nature of the waste, often this method may make the total cost more expensive.
- Energy and sterilisation costs. Work by Harding (2009) and Richardson (2011) show that this contributes not only costs but also carry significant environmental impact, something that as biorefineries we try to avoid (even if only from a PR perspective sometimes!!)
- Downstream processing (DSP): "Product recovery is often difficult and expensive; for some recombinant-DNA-derived products, purification accounts for 80-90% of the total processing cost." (Doran, 1995) Purifying bioproducts is really hard.
Reactor design is key to improving all three of these factors, generally, and with the dilute, complex nature of wastewater, even more so.
Conventionally, reactor optimisation aims to reduce the reactor volume to reduce the energy invested per unit product, and aims to achieve a higher biomass concentration, which results in less DSP cost per unit product. Using wastewater as raw material is exactly the opposite of this! Using wastewater gives a conveniently low cost and highly available raw material, but why could it make sense for an economically viable bioprocess?
The lower substrate concentrations in wastewater require lower oxygen supply than in conventional bioprocesses, which is associated with cost savings. With less oxygen supply required, the stirring can be more gentle, which may reduce shear stress in shear-sensitive organisms.
Wastewater may also potentially have a matching nutrient requirement, in terms of Carbon, Nitrogen and Phosphate - the reason we get algae blooms in rivers downstream from sewage works, for example.
Lastly, of course, beneficiating wastewater could contribute to increased resource efficiency, and reduced environmental burden, leading to greater sustainability, but this is not what I want to focus on today.
Before I continue I have to stress though, that wastewater biorefineries only make sense IF:
- we consider the economic competitiveness of product against competing products, (and NOT (just) environmental or social gain)
- we allow the separation of the steps required for the cleaning of the water (polishing) from the productivity, as this allows greater flexibility. This may be separate units on the same plant, operated by the same company, or moving the (now cleaner) water to a dedicated treatment plant.
So if wastewater is indeed a promising raw material, what is needed from the reactor design?
For bioproducts from wastewater to work, reactors need to produce product in the face of large volumes and a complex medium. The resultant broth should not affect the environment adversely, and it should make downstream processing (DSP) easier.
The interesting thing about DSP is that, even as this often contributes a lot to the total processing costs, the processing units are already very efficient and well developed. Reactor design in general needs to be better designed in order for DSP to be more effective.
We can't sterilise the stream as the energy costs would just be too great, so we need a way to focus our efforts on the biomass rather than the bulk fluid. In order to achieve that, we need to decouple the hydraulic and solid residence times. In short, we need a biofilm.
Because of the typically large and continuous flows, we can't store the liquid, so the process has to be continuous or semi-continuous.
The last factor is not directly related to the reactor but more to the market and management of these systems. We can't use just a little bit of the stream. We have to use the whole stream to make it attractive for the people involved, be they industrial partners or the municipality tasked with treating the water, to consider this approach. Therefore wastewater biorefineries is best suited to commodity products like biofuels, biopolymers or biobased building blocks rather than niche products like pharmaceuticals or pigments, and this selection influences the reactor design.
Looking at this graph from Nicolella et al (2000), with the substrate concentration in kg per cubic meter (which is the same as grams per liter) on a log scale on the horizontal axis and the flow rate in cubic meters per day on a log scale on the vertical axis, we can see a few things.
Conventional bioprocessing most often occur at more than 10g/L of substrate, where the biomass is quite happy being single cells. This is fine, but not in my current interest.
We don't want to operate in this area where it says 'problematic separation', because that means expensive DSP. This area where flocs are likely to occur are also difficult to process and require huge settling ponds.
Two areas on this graph look promising though; the static biofilms operating at slightly higher flows, indicated by this blue circle, and the particle biofilms at slightly higher substrate concentration, indicated by the orange circle.
From here we can consider the complexity of the stream. Remember that we can't add anything to the stream that would affect the environment once we discharge the effluent. No nasty chemicals, and I would also say no genetically modified organisms. So we don't have a lot of scope to modify the microbial community by force. The most robust and resilient biocatalysts would occur in a mixed community, able to withstand shock loads and hostile environments. They need to survive, but remember also, they need to produce products for us. This brings me to an important point: Wastewater biorefineries is not suited to all sorts of products. The product needs to meet commodity market needs, but ALSO need to serve an ecological function to the bioorganisms - it needs to give them a competitive advantage.
In terms of reactor design, the reactor needs to help provide this friendly environment, and to an extent that will depend on individual design requirements.
Next we need to enable effective downstream processing. We need to be able to get the product out easily. This means we don't want to sieve through all the millions of litres, but we also don't want to dissemble the reactor every time we need to get the product out - remember, we are working with continuous flows! Now, most DSP works with phase separation, gases from liquids like biogas, solids from liquids through precipitation, etc. Thus, we need this product produced in a different phase for easy DSP, and we need a general reactor design that makes it possible to get to the product.
I considered well-established, proven reactors generally used in wastewater treatment in this work, and this product recovery was the criteria that excluded most of these general reactor designs available. Only three designs remained potentially feasible. Of course, new research designs may eventually add to these.
The first is the rotating biological contactor, or RBC, and the second is the trickling filter, both very old and traditional reactor designs used in wastewater treatment. They both operate best at higher flows and slightly lower substrate concentrations.
The third one is a much newer technology called aerobic granular sludge, or AGS, which prefers slightly higher substrate concentrations, and is of a modular design to accommodate the slightly lower flow requirement.
I will now briefly consider each of these in turn to illustrate the basic idea, but bear in mind that the final reactor designs rely on accurate sizing, the eventual stream and product selected, and so on.
The trickle tower consists of media that allows biofilm growth on or through them, and the water flowing down across this biofilm. Trickle towers used to be constructed with stone, which limited their height and thus residence time, and the sheer weight of the stones crushed the lower materials leading to eventual failure and clogging. These days plastic and other light-weight media resolved this problem, as seen in the image top right, but one has to be careful about which media is selected to effectively enable product recovery.
Verdict: We included trickle towers in my work as neither of the other two reactors I talk about here could quite cope with my microbe's requirement for oxygen. Trickle towers allow growth at the air-liquid interface, but it remains to be seen if product recovery will be adequate.
The rotating biological contactor, or RBC, is a partially-submerged attached growth bioreactor, with flat, circular disks turning slowly on a horizontal shaft. It is mechanically simple, has a low energy requirement, a modular character and is easy to operate. Currently, however, this limits the process flexibility and range of wastewaters, but I believe that targeted research can improve this greatly.
Traditionally the corrugated disks that we saw in the trickle tower have been used, but this may cause problems of unbalanced growth which results in mechanical wear. Disk innovation is improving this aspect: this image in the bottom left shows an alternative, hybrid model that I think gives a more sophisticated biofilm growing area.
Product recovery can occur via removal of individual disks, or of a shear force, either by a physical spatula sort of thing or through strong periodic air sparging.
Verdict: I think the RBC is very well suited to the biorefinery concept with a lot of flexibility. It is my 'old faithful'.
The last reactor design discussed here is the aerobic granular sludge reactor developed at TU Delft and which, if the hype is to be believed, ushers in a revolution in wastewater treatment, and I think may do the same for the wastewater biorefinery concept.
The AGS consists of granules, defined as “aggregates of microbial origin, which do not coagulate under reduced hydrodynamic shear (also known as ‘sludge bulking’) and which settle significantly faster than activated sludge flocs: 15 seconds vs 20 minutes (de Kreuk et al 2007), as seen in the image at bottom right. This allows efficient biomass retention making compact reactors with integrated sludge separation feasible. It is a sequentially operated batch reactor (SBR), so fed discontinuously, and so requires modular construction.
Verdict: I am very excited about the AGS technology, but it seems to require specific skills in operation and is very new. I do think it comes closest at bringing bioprocess engineering into the realm of wastewater treatment. It is not well suited to very dilute waters, and requires low amounts of settling inerts, which tend to accumulate in the system. It is the 'supermodel' of my reactors, temperamental but sophisticated and sexy.
Two examples where TU Delft have used this reactor in a biorefinery type context are a partnership with Pacques, and a partnership with the chocolate maker Mars to produce biopolymers from wastewater.
In conclusion, I discussed some aspects that need to be considered for bioreactor design if one wants to use wastewater as raw material. I think these considerations may be useful for general reactor design as well. For bioproducts from wastewater to work, reactors need to produce product in the face of large volumes and a complex medium. The resultant broth should not affect the environment adversely, and it should make downstream processing easier.
More generally, reactor design needs to appreciate the wider system. As a unit process, they need to be optimised for overall system performance, and not designed for maximised productivity in isolation of other units downstream. It is also critical to remember that for this to work in the long term we need to consider the ECONOMICAL viability of the product against competing products, rather than the environmental or social gain, right from the start, right from the design stages.
We have just starting a project exploring the global state of wastewater biorefineries, what wastewaters are suitable, what products could be produced using this concept, what is currently being done. Any comments, input or suggestions would be very much appreciated.
In closing, I wish to acknowledge the South African National Research Foundation and the Water Research Commission for their generous funding, and multiple mentors, only some of whom are listed here.
and, Thank you very much for your kind attention.
- Harrison S & Verster B, 2014. Reactor Design For Wastewater Biorefineries: Recovering Value While Producing Cleaner Water , article in draft.
- WRC report due to be published soon
- Nicolella C, van Loosdrecht MCM, Heijnen SJ, 2000. Particle-based biofilm reactor technology. TibTech, 18, 312-320.
- Kleerebezem R & van Loosdrecht MCM, 2007. Mixed culture biotechnology for bioenergy production. Current Opinion in Biotechnology, 18(3), 207-212.
- Harding, K.G., 2009. A generic approach to environmental assessment of microbial bioprocesses through Life Cycle Assessment (LCA), PhD dissertation, University of Cape Town.
- Richardson, C., 2011. Investigating the role of reactor design for maximum environmental benefit of algal oil for biodiesel. M.Sc dissertation, Department of Chemical Engineering, University of Cape Town.