1. Introduction
Maximum agricultural production is almost always the goal of any producer and is typically achieved with large inputs of inorganic fertilizers, such a nitrogen (N) and phosphorus (P). Many N fertilizers require large energy inputs to create through the Haber–Bosch process [
1], whereas most P fertilizers are produced from a finite supply of rock phosphate (RP), such as apatite and phosphorite, that must be first mined from the ground and processed [
2]. A more readily available source of potential fertilizer nutrients than the atmosphere or the ground, as in the present case with N and P, respectively, could be beneficial for agricultural production and other sectors of human society and the environment.
Wastewater is generated constantly by municipalities and many industries. Tiseo [
3] estimated that 67 billion m
3 year
−1 of wastewater are generated by Europe and North America combined. Many wastewaters, such as municipal, agricultural, and industrial, among others, contain large concentrations of potentially useful nutrients, such as N and P [
4,
5,
6]. Unless removed from the wastewater stream, excess nutrients may pose a burden to processing facilities, such as wastewater treatment plants (WWTPs), and/or the environment, if nutrient loads to receiving waters are large. UNESCO [
7] estimated that only 20% of the globally produced wastewaters are treated before being discharged back into receiving waters. Consequences of excess nutrients in wastewaters can include pipe clogging in the infrastructure, the creation of process inefficiencies, and greater costs associated with WWTPs [
8,
9] and cultural eutrophication, followed by the cascade of potential negative effects that may lead to serious environmental degradation [
10,
11,
12].
Struvite (MgNH4PO4·6H2O), a white, crystalline mineral, is one of several minerals that may precipitate in WWTP pipes under certain aqueous chemical conditions that can restrict or completely stop pipe flow. Though unintentional struvite formation is undesirable, intentional struvite formation may be beneficial as a means to recycle excess nutrients from waste streams, such as those that enter WWTPs. Intentional struvite formation has been accomplished through chemical precipitation, where Mg is dosed at a strategic location in a WWTP to purposefully cause struvite to form and remove or recycle P and N from the wastewater, reducing the potential to precipitate and clog pipes. Once collected, the precipitate can be further processed into pellets, packaged, and sold as a blended fertilizer-P and -N material, which was accomplished recently by Ostara Nutrient Recovery Technologies, Inc. (Vancouver, BC, Canada) in association with a municipal WWTP near Atlanta, GA.
More recently, electrochemical precipitation techniques have been developed and tested to create struvite from synthetic wastewater, where an electrical current is imposed on a Mg electrode with a stainless-steel counter-electrode [
13,
14,
15]. Magnesium is supplied to the P- and N-containing solution as the Mg electrode corrodes. Aside from eliminating the need for external chemical inputs to promote the reaction to form struvite, as with the chemical precipitation method, the electrochemical method generates hydrogen that can be subsequently captured and used as an alternative energy source [
16].
Both chemically precipitated (CPST) and electrochemically precipitated struvite (ECST) have been evaluated in a variety of settings for their potential use as an alternative, blended fertilizer-P and -N source compared to other commonly used, commercially available fertilizers, including triple superphosphate (TSP), monoammonium phosphate (MAP), diammonium phosphate (DAP), and RP. Studies have reported similar plant growth from struvite [
17,
18], whereas others have reported reduced agronomic effectiveness of struvite [
19,
20] compared to typical commercially available fertilizer-P sources. The behavior of CPST and ECST in various soil textures over time has been evaluated without plants in a series of laboratory incubations under moist [
21] and flooded [
22,
23] soil conditions. Ylagan et al. [
24] evaluated corn (
Zea mays) and soybean (
Glycine max) response to CPST and ECST in a greenhouse study. Omidire et al. [
25] conducted a 2 year field study on a P-deficient silt–loam soil in eastern Arkansas to evaluate rice (
Oryza sativa) response to CPST and ECST compared to TSP, MAP, DAP, and RP, where urea was used to balance the N among the various fertilizer materials. Results of these studies conducted in Arkansas indicate struvite’s behavior in soil and struvite’s performance with a variety of crops are at least comparable to those of other commonly used, commercially available fertilizer-P/-N materials.
For a newly developed, alternative fertilizer material to find a market niche, the material must be economically viable as well, not just agronomically effective. Although struvite has been recognized as a viable fertilizer product since the late 1950s, initial commercial production was limited due to large manufacturing costs [
26]. Issues related to transportation, storage, composition, and purity also hindered commercial-scale development. However, in recent years, technological advancements, coupled with concern for resource and environmental conservation, have led to the implementation of large-scale commercial struvite production across several countries, such as Germany, the Netherlands, Japan, Canada, and the United States [
27].
The most widely used struvite product in the United States is Crystal Green
®, produced using Ostara Pearl technology [
28]. Like most technological innovations, Crystal Green
® has not entered the market as the lowest-cost option [
28,
29], with prices being roughly double those of other fertilizer-P materials in 2019 and 2020. Field trials conducted by Ostara have shown greater yields for crops grown using Crystal Green
®, resulting in returns on investment upwards of 4:1, but researchers note that fluctuations in price for crops and alternative fertilizers can impact year-to-year profitability [
29,
30].
Omidire et al. [
31] used preliminary results after one year of field trials in rice, corn, and soybean to evaluate CPST and ECST relative to TSP and other commonly used, commercially available fertilizers. Market prices from CPST currently exceed those of other commonly used fertilizer-P sources [
32,
33]; therefore, greater yields and/or lower total fertilization costs will likely need to occur to allow the struvite materials to economically compete with current production practices using other commonly used fertilizer-P sources. However, a more formal economic evaluation of struvite material use in row-crop agriculture has not been conducted to date.
In addition to assessment of economic viability, the environmental implications of struvite use are also an important consideration. Life cycle assessment (LCA) provides a widely accepted framework to assess struvite’s potential environmental impacts under various scenarios as a potential replacement for other commonly used, commercially available fertilizers. Life cycle assessment has recently been used to evaluate implications of red rice on food security [
34] and agricultural water management [
35] associated with rice production in the Lower Mississippi Delta Region, encompassing Arkansas, Louisiana, and Mississippi. Considering struvite has been characterized as having slow-release characteristics under certain soil conditions [
17,
19,
36,
37,
38], it is possible that the in-season physiological timing of plant demand for P and N will be better matched with P and N release from struvite than the dissolution timing of other fertilizer sources. Consequently, N volatilization and/or denitrification losses may be minimized relative to those from other P and/or N fertilizer sources, such that the use of struvite may result in lower greenhouse gas (GHG) emissions.
The potential societal and environmental benefits of using recycled nutrients from wastewaters as fertilizer materials in large-scale row-crop agricultural production warrant investigation. Furthermore, considering Arkansas, as the largest rice-producing state, has consistently accounted for nearly 50% of the total annual rice production in the United States in recent decades [
39], evaluating the economic and environmental impacts of struvite use in rice production in Arkansas is more than justified. Thus, the objective of this study was to evaluate the economic and global warming implications of using struvite as a fertilizer-P source for flood-irrigated rice relative to other commonly used commercially available fertilizer-P sources in Arkansas. Despite similar rice yields among fertilizer-P treatments within a growing season, it was hypothesized that struvite-P treatments will result in lower net returns compared to other fertilizer-P sources given the expected greater market prices for ECST and CPST, on account of CPST’s relative newness and ECST’s non-existence in the market yet, compared to other fertilizer-P sources. It was also hypothesized that TSP, the most commonly used fertilizer-P source in the region, would economically out-perform all other fertilizer-P sources evaluated. Furthermore, it was hypothesized that the struvite-P treatments and application of recovered-P sources in flood-irrigated rice production would provide environmental benefits to the rice production system by reducing the global warming potential (GWP) compared to the application of conventional fertilizers.
3. Results and Discussion
3.1. Rice Yield Response
Rice yields differed among fertilizer-P treatments between years (
p < 0.05) [
25]. In 2019, rice yields did not differ among fertilizer-P treatments, ranging from 14.4 Mg ha
−1 in the control to 15.6 Mg ha
−1 in both the TSP and RP treatments (
Table 4) [
25]. However, rice yields in every fertilizer-P treatment were lower in 2020 than in their respective treatment in 2019 due to a combination of differences in growing-season weather conditions and different initial soil properties [
25]. For practical purposes, the field studies shifted locations by ~100 m from 2019 to 2020, which resulted in a shift in soil pH from 7.4 in 2019 to 7.8 in 2020 plus increased extractable soil Ca to result in likely greater Ca-P precipitation in the soil, lowering the plant availability of the added fertilizer P [
25]. Furthermore, wetter pre-flood soil conditions from greater rainfall in 2019 likely promoted greater fertilizer-P dissolution and plant availability compared to the drier pre-flood period in 2020 [
25]. In 2020, rice yields from TSP, MAP, DAP, RP, and control did not differ and were greater than from ECST (
Table 4) [
25]. Rice yields in 2020 from CPST and ECST were similar to those from MAP, DAP, RP, and the control (
Table 4) despite the fertilizer-P materials having differing expected solubilities [
25].
3.2. Economic Evaluation
Although different fertilizer-P treatments did not always produce statistically different in-field yields in 2019 and 2020 (
Table 4), net revenues can still vary greatly across treatments, as profits are based on actual costs and yields observed rather than statistically significant yield differences. Estimated total fertilizer costs, total revenues, and net revenues by fertilizer-P treatment are summarized in
Table 5. Total revenue associated with the TSP treatment exceeded total revenue to all other treatments in both years. Even with some fertilizer-N cost savings provided by both CPST and ECST, total fertilization costs of both struvite treatments ranked first and third highest, respectively, in 2019 and 2020. With the exception of MAP in 2019, estimated economic net returns from TSP outperformed net returns from all other fertilizer-P sources in both years. An analysis of two-year total and average annual net revenues by fertilizer-P source (
Table 6) shows that TSP had the largest total net revenues and average revenues across the two-year study period. Across all fertilizer-P treatments, returns from CPST numerically ranked seventh and sixth for 2019 and 2020, respectively, where returns from ECST numerically ranked fifth and seventh for 2019 and 2020, respectively (
Table 7). In both years, CPST and ECST produced lower net revenues than those of MAP, DAP, TSP, and RP, which are all commercially available fertilizer-P sources (
Table 7).
Although urea-N rates applied to the struvite fertilizer treatments to balance fertilizer-N applications across all fertilizer treatments were similar to rates applied to most other treatment plots, the current market price of struvite fertilizer remains more than double many conventional fertilizer products (e.g., on average, USD 0.86 kg
−1 vs. USD 0.42 kg
−1, respectively). Therefore, total costs of P and N fertilization for the struvite treatments were about 24 to 65% greater than those for TSP, depending on the specific fertilizer-P source and year. Furthermore, average measured yields from the struvite treatments were generally numerically lower, though not always significantly lower, than from other fertilizer-P treatments. Yields from CPST numerically ranked sixth both years, whereas yields from ECST numerically ranked fourth in 2019 and seventh, even behind the control treatment, in 2020 (
Table 4). The combination of the relatively greater P and N fertilization costs and the lower yields led to the relatively poor estimated economic performance of the struvite materials as fertilizer-P sources across both years.
Though CPST has a place in the fertilizer market already as a recycled nutrient source, ECST’s actual costs have yet to be fully determined and vetted, as the electrochemical technology is in its infancy and still being developed as a potentially viable technique to recycle nutrients from wastewaters. Furthermore, one must also consider the additional benefits of both CPST and ECST as fertilizer-P sources generated by removing excess P and N from waste streams that can decrease the P and N loads in WWTPs, which may lead to reduced operational costs, and can potentially decrease the P and N loads returned to the natural environment in receiving waters.
3.3. Global Warming Potential Evaluation
Estimated GWP of rice production using CPST was 0.58 and 0.70 kg CO
2 eq kg
−1 rice in 2019 and 2020, respectively, whereas estimated GWP using ECST was 0.57 and 0.80 kg CO
2 eq kg
−1 rice 2019 and 2020, respectively (
Figure 2). In 2019, estimated GWP numerically differed between struvite materials by <3%, whereas in 2020, estimated GWP was >15% numerically greater for ECST than CPST. Annual differences in estimated GWP were at least partially related to annual differences in rice yields and the crop nutrient composition from the ECST material between years. Rice yields were similar between CPST and ECST in each year, but yields were 39 and 23% lower in 2020 compared to 2019 from ECST and CPST, respectively (
Table 4). Furthermore, in 2020, due to lower P concentrations from the different synthetic wastewater solutions prepared and processed, the amount of ECST applied was 14% greater compared to in 2019. A greater amount of struvite was necessary to provide the consistent fertilizer-P rate across all treatment scenarios (i.e., 67.6 kg P
2O
5 ha
−1). Consequently, the lower rice yields in 2020 compared to 2019 clearly had a substantial effect on the estimated GWP from the flood-irrigated rice production system in eastern Arkansas.
Compared to the other fertilizer-P sources, the estimated GWP for both ECST and CPST in 2019 was lower than from the control, TSP, and RP, but was greater than from DAP and MAP (
Figure 2). In 2020, estimated GWP for CPST was lower than RP, but was greater than the control, DAP, MAP, and TSP (
Figure 2). For ECST in 2020, estimated GWP was greater than the control and all other conventional fertilizers (
Figure 2), at least partially due to the combination of a different nutrient composition and lower rice yields in 2020 than in 2019.
In 2019, there was a small numeric increase in rice yield (58 kg ha
−1) for ECST compared to DAP, but estimated GWP was not substantially reduced. Most likely the environmental burdens related to the production of struvite were not sufficient to outrank the impact of producing the DAP, nor decrease ECST’s relative contribution to the estimated GWP footprint. Furthermore, for the DAP and MAP treatments, there were GHG credits (
Figure 3), which were due to the co-production of fertilizer-N in the fertilizer-P production process. The multioutput process in the production process of DAP and MAP also delivered the co-product of diammonium phosphate (i.e., 1 kg MAP and DAP each also co-produced 0.211 and 0.391 kg N, respectively), which substituted the use of raw materials involved, mainly nitric acid and energy [
51].
A minor reduction (0.01 kg CO
2 eq kg
−1 rice) in the impact occurred for ECST compared to RP in 2019, even though the rice yield for RP was 191 kg ha
−1 greater and the impact for RP was contributed mostly by the production of phosphate rock, which was 1.02 × 10
−2 kg CO
2 eq kg
−1 rice. Similarly, estimated GWP for ECST in 2019 was greater compared to MAP, which was due to a combination of reduced yield (i.e., lower by 205 kg ha
−1) and greater estimated GHG emissions during struvite production compared to the production of MAP (
Figure 2,
Table 2).
Figure 3 summarizes the relative contributions of numerous inputs and agronomic processes in the rice production system across the different fertilizer-P-source treatments. For CPST, of the total GHG emissions (i.e., GWP) estimated per 1 kg of rice produced for both years, approximately 66% was associated with field emissions, which was the combined effects of N
2O, CH
4, and CO
2 emissions. Nitrous oxide and CO
2 emissions accounting for the field emissions were related to the applied urea. For ECST, the contribution due to field emissions was approximately 64% across both years (
Figure 3). Similarly, across both years, the contribution due to N
2O emissions alone was 6%, whereas the contributions from CH
4 emissions was around 88% (
Figure 3). Similar ranges for the contributions of N
2O and CH
4 were estimated for the non-struvite P application scenarios.
Seed production contributed around 1% to the estimated total GHG emissions for all treatment scenarios, including the struvite treatments, whereas the combined contribution from agri-chemicals (i.e., fertilizers, pesticides, and micro-nutrients;
Table 2) was approximately 10% for CPST in both years and was 12 and 8% in 2019 and 2020, respectively, for ECST (
Figure 3). Fuel consumption contributed about 6% for CPST and ECST in both years. Irrigation contributed approximately 20% for both CPST and ECST in 2019, whereas in 2020, irrigation contributed approximately 21% for CPST and 20% for ECST (
Figure 4). The estimated credits towards the GHG emissions due to the substitution of synthetic fertilizer by the struvite material was approximately 4% for CPST across both years (
Figure 3). For ECST, the estimated GHG emissions reduction was approximately 2% in 2019, but there was no net GHG emissions reduction for ECST in 2020, as the struvite production process itself had greater GHG emissions than the calculated credit (i.e., about 2% of the total GHG was added due to the struvite production process). For ECST in 2020, the added environmental burden was mainly due to the lower P concentrations compared to the ECST batch used in 2019 (
Table 2), and approximately 2% was the added GHG emissions (
Figure 3).
The contributions during the production and application of struvite estimated per 1 kg rice are shown in
Figure 4. In 2019 and 2020, the added burdens (per 1 kg rice) in the CPST production process were NaOH (0.0044 to 0.0054 kg CO
2 eq) and magnesium oxide (0.007 to 0.0085 kg CO
2 eq), whereas the credit GHG emissions due to the substitution of equivalent amounts of synthetic fertilizers amounted to −0.03 to −0.04 kg CO
2 eq. However, in the case of ECST, the added burdens in 2019 and 2020 were electricity (0.0007 to 0.0011 kg CO
2 eq) and magnesium (0.019 to 0.03 kg CO
2 eq), whereas the credited GHG emissions due to the substitution of fertilizers were smaller than the added burdens, with a credit of only −0.03 to −0.013 kg CO
2 eq per kg rice. The equivalent mass of struvite applied per kg rice was 0.01 and 0.022 kg for ECST, respectively, in 2019 and 2020, whereas for CPST it was 0.02 and 0.03 kg kg
−1 rice, respectively, in 2019 and 2020.
3.4. Implications
Because actual market prices for ECST do not exist yet, ECST was assumed to have similar prices to CPST, which is an established struvite product. As such, not only were the fertilization costs for ECST and CPST greater than for the other commonly used, commercially available fertilizer-P sources, but measured yields were also often numerically lower, markedly lower in 2020, than field-measured yields for the other fertilizer-P sources. Improvements in yield response, or struvite prices much lower than those applied in this analysis, are needed before struvite can realistically become an economically viable alternative fertilizer-P source for flood-irrigated rice production. However, the additional economic benefits provided from the use of recycled nutrients—such as P and N load reductions in WWTPs, and P and N load reductions to receiving waters and the environment increasing future resource use efficiency as mineral-P resources are further depleted, among others, which were not accounted for in this study, may, in time, close the economic gap between present commercially available fertilizer-P sources and alternatives, such as the two struvite materials.
From the LCIA, it was also shown that the nutrient concentrations of the wastewater can greatly affect GHG emissions, since total nutrient recovery was the driving parameter for the substitutable amounts of synthetic fertilizers and the related impact of producing them. For example, with the reduced amount of recovered nutrients in 2020 for ECST, more struvite was necessary to apply to provide the consistent fertilizer-P rate across all treatments (i.e., 67.6 kg P
2O
5 ha
−1;
Table 2) and, instead of offsetting the GHG emissions, the added impact due to the struvite was 0.019 kg CO
2 eq kg
−1 rice, given that rice yield did not increase. Furthermore, for the GHG emissions, the added burden due to raw materials used in the struvite production process was greater than the credit (i.e., 0.0312 compared to −0.013 kg CO
2 eq kg
−1 rice, respectively). In 2019, for ECST, the total impact induced from the struvite system to rice system was instead −0.01 kg CO
2 eq kg
−1 rice, in which the added burden was 0.0196 kg CO
2 eq kg
−1 rice and the credit offered by the struvite was −0.03 kg CO
2 eq kg
−1 rice. Similarly, the lower rice yields in 2020 compared to 2019 across all fertilizer-P sources also affected the estimated GHG emissions per kg rice produced, as lower nutrient removal occurred from harvesting, which left more nutrients in the field potentially contributing to negative environmental consequences. In contrast to ECST, for CPST, the induced impact due to the struvite production process to the rice system was −0.024 and −0.029 kg CO
2 eq kg
−1 rice in 2019 and 2020, respectively.
Apart from mitigating GHG emissions, which is primarily related to the production of synthetic fertilizer before their use in the field, the benefits of resource recovery from struvite are related to community-scale wastewater treatment facilities, such as with respect to energy recovery, water reuse, and, importantly, nutrient recycling. Considering these prospects, it is imperative that other environmental benefits, such as changes in eutrophication potential (i.e., nutrient enrichment of terrestrial and aquatic ecosystems), fossil fuel resource depletion potential, and acidification potential, be further evaluated. This study is limited to having only addressed these additional environmental impact categories, but other impacts need to be addressed as well. It is also relevant to evaluate struvite recovery from different wastewater streams, including manure management from cattle and swine farms, and to compare various other municipal wastewater treatment technologies with respect to nutrient recovery potential.
Considering only the use of struvite-derived fertilizer-P sources (i.e., CPST or ECST) in the field for crop production, if the struvite materials do indeed dissolve and release P at a slower rate, to better match the timing of the plants’ nutrient needs with the availability of the fertilizer nutrients compared to other commonly used, commercially available fertilizer-P sources, which still requires additional research to confirm, the expected result would be greater nutrient-use efficiency, and hence less potential fertilizer loss, particularly for N from denitrification, consequently reducing GHG emissions. The potential benefit of reduced GHG emissions from the use of struvite-derived, wastewater-recycled fertilizer nutrients in the field would be expected to extend beyond flood-irrigated rice production to other upland crops, such as corn and soybean.
Presently, struvite use in row-crop agriculture is still relatively new, but is developing. Consequently, even preliminary economic and LCA evaluations, such as the current study, are as yet quite limited to non-existent. Furthermore, at the present time, the ECST production process is experimental, and thus the ultimate scaled-up and optimized operations of the ECST process may have lower environmental impacts than predicted in this work. Furthermore, during the ECST process, hydrogen is produced as a byproduct, but potential benefits were not investigated in this study. Hydrogen as a potential fuel source may provide additional environmental benefits to the whole ECST production system. The environmental impact and energy efficiency of hydrogen, however, will depend on how the hydrogen is produced [
72,
73]. Considering the average CO
2 emissions, or energy intensity, of typical fossil fuels is about 70 g CO
2 MJ
−1 [
74], the benefits of capturing hydrogen can be estimated from its potential to substitute equivalent energy/fuel in the market. Likewise, the environmental benefits of recycled nutrients, such as P and N load reductions in WWTPs, and P and N load reductions to receiving waters and the environment, among others, were not quantified in this study, which could further mitigate the estimated negative environmental impacts, such as freshwater eutrophication. A commercial-scale, operational WWTP, integrated with an up-scaled struvite production process, could further guide research and management directions to improve the environmental footprint of WWTP processes.
4. Conclusions
This study aimed to provide a preliminary evaluation of the economic and global warming implications of struvite as a fertilizer-P source for flood-irrigated rice production in Arkansas relative to other commonly used, commercially available fertilizer-P sources. Partial budget analyses supported the hypothesis that ECST and CPST would produce lower net revenues than TSP, MAP, DAP, and RP, in part because the ECST process is still in the experimental phase and robust cost information is not yet available. Economic analyses also revealed that economic returns for TSP exceeded returns to all other fertilizer-P sources over the two study years. Although not unexpected, results suggest that, without substantial yield improvements, fertilizer cost reductions, and/or government policies or subsidies that provide incentives for using struvite products, struvite fertilizers will likely not be adopted in place of the more common fertilizer-P sources because of struvite’s current cost disadvantage; thus, struvite fertilizers are not likely to be widely adopted in the near future for use in Arkansas rice production. Similar results were identified for estimated GWP, particularly for ECST and its production process and when P recovery from the wastewater is relatively low. The estimated GWP of struvite use in a rice production system was greatly influenced by the P and N recovery from the struvite and struvite’s potential to substitute for conventional synthetic fertilizers. The combined implications of lower P concentration and lower rice yields from ECST in 2020 compared to 2019 were clearly manifested in the lower overall environmental footprint for ECST compared across fertilizer-P treatments and between the two years. However, when nutrient concentrations and yields were larger, as occurred in 2019, the GHG emissions from both struvite materials were at least comparable, and at times lower, to those of other commercially available fertilizer-P sources commonly used in rice production. Continued research and refinement of the ECST process may lead to greater efficiencies, hence potentially lowering costs, such that ECST may become an economically competitive fertilizer-P source for use in agricultural crop production beyond just rice. As results indicate competitive yield and environmental outcomes, it is clear that struvite, whether CPST or ECST, has potential to become a sustainable fertilizer-P-source alternative for use in large-scale production agriculture once the costs of ECST production are better understood.