This study deals with the feasibility and practicality to recover water and nutrients from fresh urine by means of evaporation/condensation. The evaporation process generated two distinct fractions: a condensate and a concentrate. The optimal percentage of evaporation (in volume) was found to be 80%, resulting in optimal condensate quality. Higher percentages of evaporation resulted in a deterioration of water quality, as urea decomposed into ammonia, followed by volatilization of the ammonia which ended up in the condensate. Following evaporation, struvite was recovered from the concentrate at an optimal Mg/N ratio of 1/1. The condensate was purified by filtration into two layers of soil and sand, followed by zeolites. Complete removal of N-NH3 and 91% removal of chemical oxygen demand (COD) could be achieved throughout this process. Finally, the condensate was disinfected by sodium hypochlorite, achieving over 6-log inactivation of MS2 bacteriophage at a dose of 1,200 mg min/L. In conclusion, this study shows that there exist some potential benefits to the production of high-quality water and fertilizer from urine. The value of struvite recovered from the concentrate was found to be equivalent to that of the water from the condensate, showing that both streams deserve equal attention.
For over a decade, there has been an ongoing shift of paradigm in the approach of wastewater treatment and technology (Guest et al. 2009). Instead of dissipating the resources, it has been increasingly acknowledged that there is a need to turn to maximum use of resources present in wastewater, which should really be referred to instead as ‘used water’ (Verstraete et al. 2009). Therefore, sewage should be considered as a renewable resource, from which value can be extracted in the form of water, energy and nutrients (mostly nitrogen, N and phosphorus, P). Based on an approximate cost of 0.25 €/m3, water represents about two-thirds of the total value of wastewater (Verstraete et al. 2009), making it cost-effective to implement water reuse in various parts of the world where there is shortage of water, as is the case in many areas and especially in most of the developing world. Although many experiences in the world have shown the promise of water reuse, it is less technically and economically feasible to recover N and P, which are too diluted in the wastewater matrix. Indeed, urine makes less than 1% of the total volume of domestic wastewater, but contributes 81% of the nitrogen and 50% of the phosphorus (Guest et al. 2009).
An approach to overcome this problem is sewage pre-concentration by chemically enhanced primary treatment, using iron and aluminium-based coagulants (Diamantis et al. 2013). Another approach consists of the crystallization process, which allows reducing waste production as well as recovering various materials (Aldaco et al. 2008). Although both techniques seem promising, they have yet to be applied in full-scale sewage treatment plants. A simpler and most cost-effective approach may lie in source separation and decentralized treatment of urine and faeces (Kulkarni 2010). One of the limitations commonly associated with source separation of human waste is the need to rethink our sewage infrastructure; however, nearly 2.6 billion people around the world still do not have access to sanitation and a large fraction also lacks access to clean water, especially in rural areas in the developing world (WHO & UNICEF 2010). This means that there is huge potential to implement such practice in the developing world, which, as shown above, would in turn provide valuable and recyclable resources such as energy, urea, salts, minerals and water.
Human urine is a complex aqueous solution that mostly contains water (over 95%), sodium chloride (NaCl) and urea (CO(NH2)2) as dominant compounds, but also potassium (K+), and phosphate in lower concentrations (WHO & UNICEF 2010). This makes urine an important source of nitrogen and phosphorus and, in fact, most of the nutrients that people eat and drink are passed on in their urine. The recovery of phosphorus from urine is especially of crucial importance because of the limited availability of phosphate rocks found in nature (Rodehutscord 2008). It has been estimated, indeed, that our reserves of phosphorus will last for less than a hundred years, after which there will be no other choice but to recover it from water bodies, due to the very slow regeneration of phosphate rocks over geological times. Lately it has been suggested that struvite, also known as magnesium ammonium phosphate and consisting of magnesium, ammonium and phosphorus in equal molar concentrations (MgNH4PO4.6H2O), could be obtained from sewage or urine and used in agriculture as a slow-release fertilizer (Guest et al. 2009; Verstraete et al. 2009).
In our previous report (Lefebvre et al. 2013), we disclosed a novel, sustainable and low-cost decentralized sanitary system that uses biochar or animal dung to (i) dry faeces, such that it can be combusted and converted into ashes and (ii) extract water from urine by boiling it under reduced-pressure conditions. Building on that experience, the objective of this work is to provide an overview of the feasibility and practicality to recover water and nutrients from fresh urine by means of evaporation/condensation, followed by distinct treatment processes for both the concentrate (to produce struvite) and the condensate (to recover clean water).
MATERIALS AND METHODS
Urine evaporation was conducted in the laboratory with a rotary evaporator (IKA RV 10, Singapore). For each experimental batch, 500 mL of urine, freshly collected from a panel of ten individuals, were evaporated at 78 °C and −200 mbar, and water samples were collected at different time intervals, when the volume of the condensate (and therefore of urine evaporated) reached 10%, 50%, 60%, 80%, 90%, 95% and 100%, respectively. The evaporation process therefore generated two distinct fractions: a condensate and a concentrate.
Struvite formation from the concentrate
Struvite formation was induced by addition of magnesium oxide (MgO) at different Mg/N molar ratios ranging from 1/1 to 3/1 into 100 mL of the concentrate fraction collected after 80% of the original volume of urine was evaporated. The mixture was allowed to stir for 40 min and then the crystals of struvite were recovered by filtration through a 0.45-μm glassfibre filter, followed by drying and weighing.
Condensate filtration and disinfection
Filtration experiments were carried out on the condensate fraction by gravity in an open tank (total volume of approximately 20 L). Two layers of filtration materials were used in the form of (i) a top layer (20-cm height) of soil obtained from a local plant nursery (dark-brown colour; density around 1.1 g/cm3, sandy loam texture) and (ii) a bottom layer (13-cm height) of two types of sands, consisting of silica fine sand (10-cm height, light-brown colour, diameter below 1 mm, density around 1.7 g/cm3) over coarse sand (3-cm height, grey-brown colour, diameter 2–5 mm, density around 2.0 g/cm3). Following this filtration tank was another tank of similar dimensions, where additional gravity-driven post-filtration treatment was provided in the form of a layer (15 cm) of zeolites (clinoptilolite, light-green colour, diameter 1–2 mm, density around 1.1 g/cm3).
Following filtration, disinfection studies were carried out on the condensate using sodium hypochlorite. MS2 coliphage ATCC 15597-B1 was chosen as a model of virus in water. Incubation with Escherichia coli was conducted at 37 °C and 120 rpm for 20 hours. The MS2 suspension was then centrifuged at 3,600 g for 10 min and the supernatant was passed through a 0.22 μm sterile filter. Detection of MS2 coliphage relied on the standard double agar overlay technique, following the method of Butkus et al. (2004).
pH and conductivity were determined according to Standard Methods (APHA 2005). Chemical oxygen demand (COD) was determined by the closed reflux titrimetric method, ammonia (N-NH3) was analysed by spectrophotometry, and potassium (K+) and phosphates were determined by ion chromatography (Dionex, USA), after filtration using 0.45-μm glassfibre filters.
RESULTS AND DISCUSSION
Optimization of evaporation rate
Indeed, a large fraction of the ammonia was found in the condensate and this was maximal at the beginning of the evaporation process. Figure 1(b) clearly shows that the concentration of N-NH3 was maximal (4,900 mg/L) when the percentage of urine evaporated attained 10%. It was then reduced, showing that the quality of the condensate recovered from urine after this initial fraction improved. In other words, a large amount of ammonia was released initially upon heating and the subsequent fraction of water evaporated was of better quality. When 90% of the urine was evaporated, the concentration of N-NH3 was at its lowest at 150 mg/L. However, further evaporation of urine provoked another increase of ammonia concentration in the condensate and this could be explained by further degradation of urea at high temperature, a phenomenon previously described by Liu et al. (2008). The results on ammonia were confirmed with the COD, which reached a minimum of 120 mg/L at 80% of water recovery and then increased when more condensate was recovered (Figure 1(b)). Again, this can be explained by urea volatilization ending up in the condensate fraction.
In conclusion, these results show that the COD and nutrient content (N, P and K) of the concentrate increased as water was withdrawn from this fraction, making it richer in nutrients and a more suitable fertilizer than raw urine. An optimal water recovery percentage of 80% led to the optimal quality of the condensate recovered. It should be noted here that the time required to evaporate 80% of the urine was 1.5 h under our experimental conditions.
The reduction in volume would certainly facilitate storage and transportation. One of the potential drawbacks to storing and transporting the concentrated urine could again be the potential losses of ammonia by evaporation, following Equations (1)–(3). Yet, this is not a concern that is specific to our process, as fresh urine is also known to experience hydrolysis of urea within a day of storage. This phenomenon, known as ureolysis, typically leads to an increase in pH and to the production of gaseous ammonia and bicarbonate ions, by Equations (1)–(3) (Hellström et al. 1999). Therefore, it would be useful in further studies to compare the losses of ammonia during storage and transportation of concentrated urine with that of fresh urine, possibly by using our decentralized sanitary system in the field (Lefebvre et al. 2013). In practice, Hellström et al. (1999) reported that such losses are generally small in the field.
Struvite production from the concentrate stream
It is apparent from Figure 2 that the optimal molar ratio of Mg/N equalled 1:1. In these conditions, the concentration in the concentrate was reduced from 1,018 to 61 mg/L, i.e., a 94% reduction. At the same time, a moderate reduction of N-NH3 was also observed from 550 to 410 mg/L, i.e., a 25% reduction. Increasing the dose of MgO did not significantly alter these results and it is evident that phosphate was the limiting factor. This finding is in accordance with Equation (3) (Mg, N and P should be in equimolar quantities) and also with the literature, where P-recovery as struvite in the range of 90–97% has indeed been reported from digested sludge under a similar Mg/N equimolar ratio (Arakane et al. 2006; Shimamura et al. 2006). In the optimal conditions, 4.7 ± 0.7 g-struvite/L-concentrate could be recovered in the typical form of a white crystalline substance. Based on an MgO cost of 150 €/tonne and a struvite value of 250 €/tonne, the potential benefit derived from struvite could be estimated at 0.9 €/m3 of concentrate or 0.2 €/m3 of raw urine. Of course, this estimate does not take into account capital and operational costs and, in comparison, others have estimated the total costs of producing struvite in the range of 220–2,750 €/tonne (Doyle & Parsons 2002; Neethling & Benisch 2004).
Water recovery from the condensate stream
Even though the condensate obtained after 80% of the urine was evaporated displayed the optimal quality, additional treatment is still required if water reuse or recycling is to be achieved. A polishing treatment in the form of filtration was therefore applied. Table 1 shows the water quality of the raw condensate and the quality of the final product after sequential filtration by soil, sand and finally zeolite (clinoptilolite). Complete removal of N-NH3 and 91% removal of COD could be achieved through this process, which is in accordance with the literature where up to 97% of the ammonium in stored urine could be transferred onto clinoptilolite through ion exchange (Beler-Baykal et al. 2011). It should be noted here that, once the filtration materials (soil, sand and zeolite) have adsorbed the nitrogen from urine, they could potentially be used as a soil conditioner. The application of the condensate on the soil would indeed be expected to raise the soil pH and increase its value as a fertilizer. However, the impact of such fertilization would require further investigation, with respect to crop production and pollution.
|Sequential filtration treatment|
|Sequential filtration treatment|
bdl = below detection limit.
In this study, we have shown that there exist some potential benefits to the production of high-quality water and fertilizer from urine. The monetary benefit that can potentially be made from crystallizing struvite from the concentrate was found to be equivalent to that made from recovering water from the condensate, showing that both streams deserve equal attention. However, the experiments made in the laboratory should be confirmed at full-scale using our prototype urine-diversion dehydration toilet. In particular, it is necessary to verify that the energy needed to evaporate the urine can either be offset by the value of the products (struvite and water) or can be provided by the burning of the faeces. Moreover, urine characteristics are known to change significantly throughout storage as the predominant form of nitrogen shifts from urea to ammonium and, in the field, it may not be practical to deploy a continuous system where urine is not allowed to settle. Finally, there is, of course, a need to identify the various other barriers to adoption, which could be political and social.
The authors would like to acknowledge the funding from the Bill & Melinda Gates Foundation (OPP1069475) for this project.