Bioresource Technology 175 (2015) 517–528

Contents lists available at ScienceDirect

Bioresource Technology journal homepage: www.elsevier.com/locate/biortech

Technoeconomic analysis of large scale production of pre-emergent Pseudomonas fluorescens microbial bioherbicide in Canada Edmund Mupondwa ⇑, Xue Li, Susan Boyetchko, Russell Hynes, Jon Geissler Bioproducts and Bioprocesses, Science and Technology Branch, Agriculture and Agri-Food Canada, Government of Canada, Saskatoon Research Centre, 107 Science Place, Saskatoon, SK S7N0X2, Canada

h i g h l i g h t s  Commercial production of Pseudomonas fluorescens BRG100 bioherbicide was analysed. Ò

 Capital investment scaling and profitability are analysed using SuperPro Designer .  Total capital investment for a BRG100 fermentation plant is $17.55 million.  NPV shows that the fermentation plant is profitable over wide operating scale.  Small plants require need NPV breakeven prices but are less capital cost sensitive.

a r t i c l e

i n f o

Article history: Received 10 August 2014 Received in revised form 24 October 2014 Accepted 27 October 2014 Available online 4 November 2014 Keywords: Technoeconomic Bioherbicide bacterial fermentation Bioprocess design Capital investment Net present value

a b s t r a c t The study presents an ex ante technoeconomic analysis of commercial production of Pseudomonas fluorescens BRG100 bioherbicide in Canada. An engineering economic model is designed in SuperPro DesignerÒ to investigate capital investment scaling and profitability. Total capital investment for a stand-alone BRG100 fermentation plant at baseline capacity (two 33,000 L fermenters; 3602 tonnes annum1) is $17.55 million. Total annual operating cost is $14.76 million. Raw materials account for 50% of operating cost. The fermentation plant is profitable over wide operating scale, evaluated over a range of BRG100 prices and costs of capital. Smaller plants require higher NPV breakeven prices. However, larger plants are more sensitive to changes in the cost of capital. Unit production costs decrease as plant capacity increases, indicating scale economies. A plant operating for less than one year approaches positive NPV for periods as low as 2 months. These findings can support bioherbicide R&D investment and commercialization strategies. Crown Copyright Ó 2014 Published by Elsevier Ltd. All rights reserved.

1. Introduction Over the past two decades, there have been significant efforts aimed at the development and commercialization of microbial bioherbicides (bacteria, fungi, and virus) to control both pre- and postemergent grass and broad-leaf weeds (Bailey, 2010; Bailey and Falk, 2011; Bailey et al., 2010; Ash, 2010; Charudattan and Dinoor, 2000; Hynes and Boyetchko, 2006; Glare et al., 2012). These endeavours have been driven by growing concerns over environmental and health impacts of chemical herbicides (Glare et al., 2012), and a search for viable alternatives for controlling economically important weeds that have developed resistance to chemical herbicides (Beckie et al., 1999, 2013; Heap, 2014). There has also been a slowdown in the discovery of new chemical leads ⇑ Corresponding author. Tel.: +1 306 385 9360; fax: +1 306 385 9482. E-mail address: [email protected] (E. Mupondwa). http://dx.doi.org/10.1016/j.biortech.2014.10.130 0960-8524/Crown Copyright Ó 2014 Published by Elsevier Ltd. All rights reserved.

since 2005, with increasing difficulty in converting a new lead into a new product launch from at least 140,000 chemicals that must be screened to find one new, commercially acceptable, synthetic pesticide; this has led to a very drastic decline in new product launches from 2002 to 2010 (Bailey and Falk, 2011; Glare et al., 2012). This slowdown has been aggravated further by the wide and rapid adoption of new traits in primary crops such as corn, cotton, soybean, and canola containing herbicide tolerance and insect resistance genes. In spite of considerable efforts to commercialize bioherbicides and indeed other biopesticides, the sector is still characterised by small to medium-scale firms, and it continues to play a relatively minor role in crop protection as a percentage of the more than $40 billion world pesticide market. This is notwithstanding positive growth rates reported for the biopesticide sector whose market size was estimated at $1.3 billion in 2011 and projected to increase to $3.2 billion by 2017, representing a compound annual

518

E. Mupondwa et al. / Bioresource Technology 175 (2015) 517–528

growth rate of over 15% over that period (Bailey and Mupondwa, 2006; CAB-International, 2010; Glare et al., 2012). Research and development of bioherbicides, including their commercialization, has not advanced as rapidly as that for insect and phytopathogen pests. For instance, Bacillus thuringiensis (Bt) based active ingredients marketed for control of crop and forestry lepidopteran insect pests account for over 90% of microbial biopesticides (Bailey and Mupondwa, 2006; CAB-International, 2010; Glare et al., 2012). The commercialization of bioherbicides has been constrained by factors including small market size confined to niche applications, uneconomic mass-production, and lack of commercial backers (Bailey and Falk, 2011). For instance, BioMal from the fungus Colletotrichum gloeosporiodies f. sp. malvae was Canada‘s first bioherbicide; it was discovered and developed during 10 years of research at Agriculture and Agri-Food Canada (AAFC) in collaboration with commercialization partner Philom Bios (based in Saskatoon) in the mid-1980s to 1990s. However, after a 2-year registration, the company halted the product due to decreased market, unprofitable commercialization cost, and high production cost (Bailey et al., 2010). In addressing these constraints, Canada has continued concerted research and development (R&D) endeavours to develop and commercialize microbial bioherbicides. Recent public sector R&D by the Science and Technology Branch of AAFC has focused on the discovery and development of active ingredients based on the bacterium Pseudomonas fluorescens for use in mainstream crops to control economically important annual grass weeds such as green foxtail and wild oats, two very important weeds in North American agricultural production regions. In particular, BRG100, a strain of the bacterium P. fluorescens, has been shown to possess 85–90% efficacy in weed control as a single treatment pre-emergent bioherbicide, by inhibiting weed seed germination and suppressing root growth. This new technology is described by AAFC inventors Boyetchko et al. (2005) in US Patent 6,881,567 issued to AAFC with a corresponding Canadian Patent 2,377,054 issued in 2006. The technology basically provides isolated microbial bacteria from a P. fluorescens strain that can be formulated as a granular bioherbicide and applied to soil before, during or after planting to a wide range of economically important weed species including green foxtail (Setaria viridis [L.] Beauv.), foxtail barley (Hordeum jubatum), crabgrass (Digitaria sanguinalis), annual ryegrass (Lolium rigidum), barnyard grass (Echinochloa crusgalli), yellow foxtail (Setaria glauca), Italian rye grass (Lolium multiflorum), Goose grass (Eleusine indica), and wild oat (Avena fatua) (Boyetchko et al., 2005). These weeds, in particular wild oats and green foxtail, are among the most bothersome weeds for farmers producing major cereal crops such as wheat, barley, rye, oat, triticale and other cereal crops, due to full season competition with crops, resulting in significant reduction in crop yield and major economic losses for farmers (Beckie et al., 1999). In terms of bioprocessing technology, BRG100 is mass produced via submerged liquid fermentation (as opposed to solid state fermentation) (Boyetchko et al., 2005). According to Stowell (1991) and Jackson et al. (1996), submerged liquid fermentation is considered to be the most economical and efficient commercial mass-production method for most microbial biopesticidal agents, as demonstrated by Churchill (1982) and Stowell (1991) for two commercial mycoherbicides, CollegoÒ (now re-registered as Low DownÒ) and DevineÒ, which are manufactured in the United States using submerged liquid fermentation. On the other hand, solid substrate fermentation (the growth of microorganisms on a solid matrix with low free water content) is less frequently used due to what is regarded as higher labour costs, technical challenges in preserving sterility, inability to control culture conditions, and problems in retrieving microbial spores from the feedstock (Churchill, 1982). Hence, in terms of commercial application,

effective commercialization of biopesticides has been largely influenced by economics of low-cost large-scale fermentation, production, and formulation of highly efficacious and stable microbial populations (Ash, 2010). In terms of formulation, BRG100 is formulated as a Pesta (Boyetchko et al., 2005). The term ‘Pesta’ refers to a pasta-like granular product made from a cereal grain flour in a process that encapsulates biocontrol agents as described for instance in Connick et al. (1991, 1996). In terms of product development, granular formulation ‘‘Pesta’’ was the first granular inoculum developed to deliver BRG100 as a grass weed bioherbicide, and which has been demonstrated to prolong the shelf-life of a dried encapsulated bioherbicide (Connick et al., 1991, 1996; Daigle et al., 2002). The role of formulation and efficacy has been well articulated (Boyetchko et al., 2002; Boyette et al., 1996; Daigle et al., 2002; Hynes and Boyetchko, 2011). Although BRG100 has potential for large-scale application in mainstream cropping systems, past challenges associated with the development of bioherbicides point to a shortage of investment models for guiding R&D within the biopesticide innovation chain. There is a dearth of studies in the public domain that have provided a complete technoeconomic evaluation of microbial bioherbicide fermentation technology especially at the crucial pre-commercialization phase, complete with capital investment analysis. A few exceptions include two recent technoeconomic analyses of bacterial biopesticide fermentation for the mass production of Bt bioinsecticide (Rowe and Margaritis, 2004; Brar et al., 2007). Rowe and Margaritis (2004) modelled a stand-alone fermentation plant in Ontario for mass production of Bt bioinsecticide whose total capital cost was estimated at $18 million and assumed to operate 24 h day1 330 days annum1 with a production capacity of 3  107 billion international units (BIU) annum1 based on low density fed-batch fermentation in a 281,000-L fermenter utilizing 322.6 kW of energy. This is a large capacity production process representing 8–15% of world annual Bt production estimated at 13,000 tonnes annum1 (Rowe and Margaritis, 2004). In a related study, Brar et al. (2007) conducted technoeconomic analysis of a similar 3  107 BIU annum1 stand-alone fermentation plant in southern Ontario based on earlier analysis by Rowe and Margaritis (2004). Their estimated total capital investment ranged from $18 to $21 million for various process scenarios. In both of these studies, the estimated capital cost was 1.5 times installed equipment cost. By comparison, total capital investment for a monoclonal antibody biopharmaceutical plant with 60,000 L bioreactors is around 12 times the total equipment cost (Flickinger, 2013). The objective of this study is to conduct a detailed ex ante technoeconomic analysis of the manufacturing process for the large-scale production of BRG100. The study provides a range of quantitative metrics for guiding research direction in the early phase of bioherbicide technology development. This includes a projection of costs associated with BRG100 mass production and other scale-up costs to enable future investment decisions, as well as parameters for determining some significant go/no-go decisions within the innovation chain.

2. Methods 2.1. Ex ante analysis and definition of BRG100 Pseudomonas fluorescens BRG100 technology is still under development. Hence, this analysis is conducted as an ex ante study within the context of the innovation value chain model (Kline and Rosenberg, 1986) which simply refers to sequence of phases from discovery, technology scale-up, fermentation and formulation, mass-production, and commercialization. This study focuses

E. Mupondwa et al. / Bioresource Technology 175 (2015) 517–528

on the design of the fermentation plant and the development of technoeconomic parameters for deriving measures of economic viability. Since the study is ex ante, the data largely comprises experimental data and related empirical sources, engineering cost estimates from vendors, and costs of similar technologies. Throughout this paper, the term BRG100 will be used to distinguish two important aspects of this technology: (a) P. fluorescens BRG100 bacterial cells isolated from the rhizosphere of green foxtail; (b) the final BRG100 Pesta granular product (i.e., the bioherbicide) formulated to release and deliver the P. fluorescens BRG100 bacteria to inhibit or suppress weeds. This distinction between P. fluorescens BRG100 bacterial cells per se and the final BRG100 Pesta granular product recognizes that P. fluorescens BRG100 bacterial cells in their unformulated state are not in themselves ready for biological weed control; they must be properly formulated to ensure stability, efficient handling, storage, transportation, flowability through on-farm application equipment, protection of the bioherbicide from adverse environmental conditions, and promote bioherbicide efficacy and delivery at the target weeds. This includes the determination of application rates and integration into current farming practices by farmers using standard equipment. Formulation includes: (a) addition of inert substances (which do not possess bioherbicidal activity) to serve as carriers for the bacterial cells (the active ingredient) and improve application efficacy, safety, handling, and storage; (b) addition of an adjuvant to promote and sustain efficacy of the bacterial cells via protection from UV radiation and adverse environmental conditions. Hence, in the context of this study, the mixture of the active ingredient (bacterial cells) and inert ingredients (Pesta and other substances) is what constitutes the BRG100 bioherbicide formulation. Where brevity is required and context is clear, the final formulated BRG100 product will simply be referred to as BRG100. Overall, formulation is a key component of ongoing research at AAFC aimed at moving this technology further upstream along the innovation chain, recognizing that formulation inadequacies are a key factor in the number of bioherbicides withdrawn from the market, including microorganisms described in earlier biopesticidal screening studies that were frequently applied in formulations that are not commercially feasible or did not include formulations (Hynes and Boyetchko (2011) 2.2. Base case process description The BRG100 production process is designed in Intelligen Superpro DesignerÒ. A flowsheet for the major unit operations is presented in Fig. 1. It contains an upstream and downstream section. Overall, the process itself starts with laboratory culture cultivation, medium mixing, fermentation, Pesta mixing, extrusion, grinding, fluid bed drying, and end product packaging. The conceptual design in Intelligen SuperPro DesignerÒ uses experimental data from previous laboratory and field studies including strain isolation, storage, cultures, ‘‘Pesta’’ formulation, and field trial (Boyetchko et al., 2005; Daigle et al., 2002). The specific process starts with growing the BRG100 bacterial cells isolated from the rhizosphere of green foxtail in a nutrient medium in the lab first (sub-process area P-1, P-18, and P-22 in Fig. 1), subsequent to which the broth suspension was transferred to a larger fermenter (sub-processes P-3 and P-6). M9 medium, molasses, and zinc sulphate were prepared according to the composition weight ratio shown in Table 1, along with material costs for the base case. BRG100 bacterial production yield is defined in terms of colony forming units (CFU), which is a key quantitative output parameter because it provides the basis for specifying the capacity of the processing plant. CFU is bacteria with the ability to live and reproduce to form another colony of the same bacteria under specified conditions. The BRG100 bacterial seed culture (106–109 cfu mL1) was

519

inoculated in the medium and grown at 20 °C at 150 rpm for 48 h and scaled up at 1% inoculation ratio. Costs for medium M9 were assumed to be 10% cheaper relative to costs for pilot scale materials that vendors provided. A range of production options is possible, from small scale to large scale. In this study, the inoculation ratio was 1%, roughly from 330 mL to 33 L and to 3300 L in the base design case. There were two stirred fermenters (P-3 and P-6 in Fig. 1) to incubate BRG100 and broth medium at 20 °C for 48 h at 150 rpm to keep the process continuous in a large scale operation (this ex ante study does not investigate a batch process which, due to space limitations, is a topic of future research). After fermentation, the BRG100 bacterial suspension (approximately 3  109 cfu mL1) was mixed with flour (80/20 oat flour/maltose) (P-17) at a ratio of 1:3.3, after which the mixture was fed into a twin-screw extruder and granule cutter (P-12, P-8, P-10). The granules were dried in a fluid-bed dryer (P21) at 40 °C to a water activity level of 0.2–0.33. The resulting BRG100 bioherbicide granule product (P-20) contained approximately 108 cfu g1 bacteria whose shelf life can be up to 16 months. The final moisture content of the BRG100 bioherbicide granule was approximately 6%. In the base case design, the process flow diagram shown in Fig. 1 assumes that the BRG100 fermentation plant operates 24 h, 330 days per year, for a total of 7920 h a year. The number of operating days and plant size was subsequently modified in order to analyse fermentation plant economics beyond the base case, taking into consideration potential market size. 2.3. Cost estimation As indicated above, in the absence of data on current operating plants, the estimation of total cost has to rely on available empirical data and process simulation tool Intelligen SuperPro DesignerÒ, a valuable tool that enables the user to concurrently design and evaluate manufacturing and project economics. 2.3.1. Total capital cost Total capital investment refers to fixed costs associated with the project. It is the sum of the following costs over all upstream and downstream sections of the process: direct fixed capital, working capital, start-up and validation cost, up-front R&D cost, and upfront royalties. Direct Fixed Capital (DFC) costs include costs associated with a plant’s capital investment. It is the sum of direct costs (equipment, processing piping, instrumentation, buildings, facilities); indirect costs (engineering and construction); and miscellaneous costs (contractor’s fee and contingencies). Fixed capital cost can be estimated using various methods such as order-ofmagnitude methods, preliminary estimate methods (long factor method), definitive estimate methods, and detailed estimate method (Green and Perry, 2008). DFC in this study is estimated using the equipment purchase cost (PC) built-in module in Intelligen SuperPro DesignerÒ and as well as information provided by vendors for specific equipment such as water distiller. It is assumed unlisted equipment purchase cost is 0.10 PC. Parameters and assumptions used to estimate DFC are summarized in Table 2. 2.3.2. Operating cost Operating cost is the cost of producing the product. It includes costs related to the demand for a number of resources: materials, consumables, labour-dependent cost, utilities, waste treatment, facility-dependent costs, laboratory cost, transportation, miscellaneous operating cost, advertising and selling cost, running royalties, and failed product disposal cost. Parameters and assumptions used to estimate operating cost are also summarized in

520

E. Mupondwa et al. / Bioresource Technology 175 (2015) 517–528

Fig. 1. Flow diagram of bioherbicides production using SuperPro DesignerÒ.

Table 1 Composition of BRG100 medium and material costs for the base case. Component Medium preparation Sodium phosphate dibasic Na2HPO4 Potassium phosphate KH2PO4 Ammonium chloride NH4Cl Sodium chloride NaCl Sucrose Magnesium sulphate MgSO4.7H2O Thiamine-HCl Calcium chloride dihydrate 0.1 M CaCl22H2O Zinc sulphate Molasses Oat flour Maltose Water Paper bag

Composition in medium (g L1)

Unit cost ($ kg1)

Annual amount (kg)

6 3 1 2 2 1.23 0.005 0.015 0.06 20

57 63 40 8 27 37 563 68 37 2 2 0.6 0.002 0.1/entity

6,936 3,468 1,156 2,312 2,312 1,422 6 17 69 23,119 2,904,026 726,026 1,098,372 360,199 entities

Table 2. Labour rate is basic wage rate plus expenditures on benefits, supervision, supplies, and administration charge.

difference between discounted annual cash flows from the sale of BGR100 and production costs as in Eq. (1):

2.4. Financial and profitability analysis of BRG100 fermentation plant

NPV ¼ I0 þ

Profitability analysis and projection of revenues for the BRG100 fermentation plant is based on maximization of net present value (NPV) and internal rate of return (IRR) for given initial capital investment, taking into consideration operating costs and annual cash flow over the 20 year life of the bioherbicide fermentation plant, commencing in 2014 and terminating in 2034. General finance principles for capital investment analysis are used (Ross et al., 2007). NPV is generally used as a metric for determining the acceptability of an investment whose benefits can be quantitatively measured in monetary terms. It is the sum of all costs and benefits over the life of the investment, discounted at the opportunity cost of capital. Here, the opportunity cost of capital is the expected rate of return that investors in the BRG100 fermentation plant would forgo by selecting the BRG100 investment versus alternative investments. In this study, NPV is computed as the

where I0 is the initial investment in the BRG100 fermentation plant; CFAt is annual cash flow from assets and is given by CFAt = (TRt  TCt  DEPt) (1  T) + DEPt; TRt is total revenue before tax; TCt is total cost before tax; DEPt is depreciation over the life of the fermentation plant; T is the corporate marginal tax rate; SVN is salvage value; d is the discount rate or cost of capital; and t = 1, 2, . . ., N denotes year with N terminal time. From an investor’s perspective, this computation simply converts all the components of CFA to their present values, accounting for the timing of income, expenditures, and tax payments during the life of the fermentation plant. Other key evaluation parameters include: (a) 8-month construction period; (b) 3-month start-up period; (c) 10-year depreciation period; (d) operating capacity for each year (80% in 2014; 90% in 2015; 100% thereafter). Return on investment (ROI) is computed to provide an additional performance parameter for evaluating the

N X CFAt SV N t þ ð1 þ dÞN t¼1 ð1 þ dÞ

ð1Þ

521

E. Mupondwa et al. / Bioresource Technology 175 (2015) 517–528 Table 2 Capital cost and operating cost assumptions and parameters for BRG100 fermentation plant. Capital cost

Operating cost

Items

Estimation assumption

Items

Estimation assumption

1. Direct cost (DC) Total equipment purchase cost (PC) Installation A. Piping B. Instrumentation C. Insulation D. Electrical facilities E. Buildings F. Yard improvement G. Auxiliary facilities 2. Indirect cost (IC) H. Engineering I. Construction 3. Other cost (OC) Contractor’s fee Contingency 4. Directed fixed capital (DFC) 5. Working capital 6. Start-up and validation cost 7. Up front R&D 8. Up front royalties

DC = PC + installation + A + B + C + D + E + F + G Listed equip. purchase cost + unlisted equip. purchase cost Installation cost of listed equip. + unlisted equip. 0.35  PC 0.40  PC 0.03  PC 0.10  PC 0.45  PC 0.15  PC 0.40  PC IC = H + I 0.25  DC 0.35  DC

Materials Consumables Maintenance Depreciation Insurance Local taxes Factory expenses Labour Labour basic rate Benefits factor Operating supplies factor Supervision factor Administration factor Lumped rate Direct time utilization Laboratory costs Utilities

From mass balance Shake flask 6% of DFC Depreciated 10% of DFC 1% of DFC 2% of DFC 5% of DFC Sum (labour demand  Wage rate) $17 h1 0.17 0.1 0.2 0.5 $20 h1 70% 15% of labour cost Electricity unit cost $0.09 kWh1 Chilled water unit cost $0.40 MT1 Cooling water unit cost $0.10 MT1 Steam unit cost $4.20 MT1 Fixed $200,000 year1 Fixed $10,000 year1 0

0.05(DC + IC) 0.10(DC + IC) DFC = DC + IC + OC 30 days of labour, raw materials, utilities, waste treatment 5% of DFC $100,000 0

efficiency of investing in the BRG100 fermentation plant, calculated simply as net profit from BRG100 sales divided by the cost of the investment, expressed as a percentage. Investors can use NPV to indicate the value of a capital investment. An additional parameter (not calculated here for brevity) is the internal rate of return (IRR) which is the discount rate equating discounted benefits and costs. The decision rule is to consider a capital investment if IRR exceeds an investor’s minimum required rate of return or cost of capital, depending on the industry and level of acceptable risk. Equity investors in the bioherbicide plant may prefer IRR values that significantly exceed the cost of capital, as an indicator of how much value is added to the bioherbicide business. In terms of production economic theory, nominal profits of the bioherbicide fermentation plant can be conceptually specified as a profit function (Chambers, 1988):

pðp; r; zÞ  max fp  q  r  xðq; x : zÞ 2 Sg qx

ð2Þ

which is a function of BRG100 fermentation technology, where p is a vector of BRG100 output prices; r is a vector of input prices; q is a vector of BRG100 output; x is a vector of inputs used to produce a kilogram of BRG100 granules; z is a vector of fixed factors; and S is a closed, bounded, smooth, and strictly convex set of all feasible combinations of BRG100 production inputs and outputs, i.e., the production possibility set for the bioherbicide fermentation plant. Fermentation costs can affect the profitability of a bioherbicide fermentation plant. In this case, it is useful to evaluate economies of scale associated with the plant at various scales of production since increasing the scale of production in fermenters could lower production costs and enhance the competitiveness of BRG100 relative to conventional chemical pesticides. Economies of scale refer to the reduction in the bioherbicide fermentation plant’s average cost as BRG100 production increases. The behaviour of long-run average cost is typically used by production economists and engineers to describe economies of scale. In this context, a bioherbicide plant can exhibit three types of returns to scale: constant, decreasing, or increasing. These concepts are mentioned here only cursorily since they are important in the managerial economics of a fermentation plant, and indeed any processing plant requiring large capital investment outlays. Given the significance of capacity as a decision variable for bioherbicide plant managers in managing

R&D Advertising and selling Running royalty expense

production costs, this analysis uses a scaling factor to estimate capital cost for BRG100 fermentation plant capacities ranging from 360 tonnes to 36,020 tonnes annum1. The scaling involves a power function (capacity ratio with exponent) whereby a nonlinear cost relationship is specified to estimate the cost of a new fermentation plant from empirical costs for an equivalent existing plant of different sizes (small-, medium-, and large scale operation). This relationship is formalized in Eq. (3) as follows:

C ne ¼ C ex

 b Q ne Q ex

or ln C ne ¼ ln C ex þ b ln

  Q ne Q ex

ð3Þ

where Cne is the estimated capital cost of the liquid state fermentation production technology with capacity Qne; Cex is the empirical capital cost of an existing similar fermentation plant with capacity Qex. The exponent b is a scaling factor (real constant number) describing the economic and financial impact of changing the size of the BRG100 bioherbicide plant. The scaling factor b is set as 0.6 (sixth-tenths-factor-rule) for order-of-magnitude estimation of new equipment cost (Green and Perry, 2008). In this case, b = 0.6 value implies that doubling the size of the BRG100 bioherbicide plant increases capital cost by 60%. It is clear that b = 1.0 implies a pari passu (i.e., proportionate) increase in capital cost as fermentation capacity increases; b < 1 implies a less than proportional increase in capital cost as capacity increases; b > 1 implies a greater than proportional increase in capital cost. As stated earlier, these three conditions (b = 1, b < 1, and b > 1) are referred to in economics as constant returns to scale, increasing returns to scale, and decreasing returns to scale respectively (Chambers, 1988). These conditions also imply that there is an economic and financial advantage in increasing the scale of operation for a large capital investment, but only to a point, beyond which it becomes uneconomical to expand capacity. This is an empirical issue not explored further in this study. In general, b usually varies from 0.5 to 0.9, as a function of type of processing plant, with b = 0.6 typically used for chemical processing plants (Green and Perry, 2008). 2.4.1. BRG100 prices and calculation of revenue Evaluating a range of prices at which BRG100 can be sold enables profitability comparisons with conventional chemical herbicides. In terms of price data, there are no bioherbicides registered

522

E. Mupondwa et al. / Bioresource Technology 175 (2015) 517–528

for use on agricultural crops in Canada currently, and most successful commercial applications of biopesticides hitherto involve bioinsecticides and biofungicides in high-value crops. In the US similarly, there are very few bioherbicides on the market, with a majority targeted at the organic market, and with little or no data on application costs. Hence, there are no readily available bioherbicide market prices that can be employed as a comparative benchmark for computing potential revenue generated by the BRG100 bioherbicide fermentation plant. A comparison with current application rates and costs for conventional chemical herbicides used in target weeds provides an approximate reference point. Using the Government of Saskatchewan Guide to Crop Protection (Saskatchewan Agriculture, 2014a), there is a wide range of application costs for grass weeds such as green foxtail, wild oat, and the like which include herbicides such as alachlor, acetochlor, clethodim, dimethenamid-P, S-Ethyl dipropylthiocarbamate (EPTC), rimsulphuron, and S-metolachlor. In terms of extensive usage, a herbicide such as EPTC was first registered in 1958 and is used as a pre-emergence and post-emergence thiocarbamate herbicide against grass and broadleaf weeds, with extensive use in major crops in Canada and the US. The range of application cost for these chemical herbicides is $13–$100 ha1, with a majority of costs clustered in the $25–$50 ha1 range per kilogram of active ingredient. In fact, a number of relevant herbicides listed in the Guide to Crop Protection have application rates in the 1 kg ha1 range (for granular) and 1 L ha1 for liquid or wettable formulations. The Government of Saskatchewan Crop Planning Guide (Saskatchewan Agriculture, 2014b) provides farmers with cost data for developing their individual farm crop production budgets. It recommends herbicide costs averaging $50 ha1 for the Dark Brown and Black Soil Zones for major grains and oilseeds seeded on stubble. These application rates are obviously mainly for chemical products and it is recognized that formulations for a biological organism required to achieve efficacy at these application rates generally need special optimization for efficacy. In a marketing sense, it is also desirable to compare these chemical pesticide application rates with bioherbicide products. Unfortunately, there are very few commercial bioherbicides that provide information on application rates. It is often a challenge to obtain this information because technical details pertaining to a majority of formulated commercial bioherbicides in not available in the public domain, either through scientific publication or published patents. In view of this deficiency, one could use examples of entomopathogenic bacteria such as B. thuringiensis (Bt) which is the most successful commercial biopesticide, notably the subspecies: Bt kurstaki against lepidopteran pests. There are also other Bt subspecies: Bt israelensis against mosquito larvae and Bt tenebrionis against larvae of the Colorado potato beetle, as well as Bacillus sphaericus which is applied to control mosquito larvae. Examples of biofungicides used in Canada include ones based on the species Bacillus subtilis. Large-scale industrial production of Bt biopesticides is undertaken using liquid state fermentation, although production can still be performed using solid or semi-solid state fermentation. The Guide to Crop Protection lists two biopesticides used in mainstream Canadian Agriculture: DipelÒ (a bioinsecticide by Valent BioSciences) and SerenadeÒ (a biofungicide by DowAgroSciences). DipelÒ is a formulation containing Bt var. Kurstaki strain ABTS351 fermentation solids, spores, and insecticidal toxins, and used for control of a range of lepidopteran pests. It is registered in Canada for sunflower, timothy, corn, and potato to control sunflower moth, Essex (European) skipper, European corn borer larvae, and cabbage loopers. The application rates are as follows: 0.568 kg–1.112 kg ha1 when used in corn and potato fields; 0.314 kg–0.625 kg ha1 for use in sunflower; and a much lower rate of 0.141 kg–0.274 kg ha1 when used in forage grasses. The

biofungicide SerenadeÒ is registered for use in bean, chickpea, lentil, pea (all types), and soybean to control white mould or stem rot (Sclerotinia sclerotiorum), Botrytis blight or pod rot (Botrytis cinerea), and Sclerotinia stem rot (S. sclerotiorum). There are two Serenade formulations: SerenadeÒ Max (a wettable powder formulated using 14.6% B. subtilis QST713 strain); and SerenadeÒ CPB formulated as an aqueous suspension using 1.34% B. subtilis QST 713 strain. SerenareÒ Max (wettable powder) has several application rates depending on the crop and pest targeted. Recommended application rates are as follows: 2.97 kg–5.93 kg ha1 (for use on bean, chickpea, lentil, pea, and soybean (for S. sclerotiorum) and potato (for White mould S. sclerotiorum and Potato early blight Alternaria solani); 0.247 kg–0.988 kg ha1 (for use on soybean and canola to control Brown spot (Septoria glycines), Frogeye (Cercospora sojina), and Sclerotinia stem rot (Sclerotinia sclerotiorum). SerenadeÒ CPB (aqueous suspension) has the following application rates: 3.954 L–15.074 L ha1 (for use on bean, chickpea, lentil, pea, and soybean (for S. sclerotiorum); 0.988 L–3.953 L ha1 (for use on soybean and canola to control Brown spot Brown spot (S. glycines), Frogeye (C. sojina), and Sclerotinia stem rot (S. sclerotiorum). Another commercial product for a comparative benchmark is NodulatorÒ (by Becker Underwood), a peat-based nitrogen-fixing Rhizobia formulation delivering viable cells of Rhizobium legumiosarium bv. Viciae, which is a Gram-negative soil bacteria that fixes nitrogen. However, the fermentation production systems used in the industry are similar. The pelletized granules of NodulatorÒ are applied in-furrow at the rate of 7 kg ha1; the granules are designed to remain intact for uniform, steady flowability using standard farm application equipment. The above comparisons provide a gist of types of application rates for commercial delivery of weed control products. The idea obviously is not to provide a direct comparison. Indeed, the use of bioinsecticide DipelÒ and biofungicide SerenadeÒ application rates is a stretch given that these are not bioherbicides. It is hoped, however, that in the absence of adequate commercial examples for comparable bioherbicides, the use of Bt examples provides at least some tangible albeit rudimentary means of evaluating assumptions by BRG100 researchers regarding targeted application rates for this innovation. As noted in Section 2.2 above, a BRG100 granular inoculum formulation called Pesta is the final step (P-20) (in Fig 1) resulting in the production of bioherbicide granules for delivering P. fluorescens BRG100 bacterial herbicide, formulated into a pesta granular formulation using a single screw extruder followed by fluidized bed drying. The Pesta provides an extended shelf life of up to 16 months (Boyetchko unpublished data, AAFC, Saskatoon Research Centre). The application of BRG100 granular formulation is the first example of a soil bacterium used as a pre-emergent bioherbicide, and hence sets a precedent. For fungi, there are several examples of foliar and soil-application to control both pre- and post-emergent weeds, as elaborated in AAFC’s Phoma macrostoma technology, a broad spectrum bioherbicide for turf and agriculture (Bailey and Falk, 2011). AAFC researchers are continuing with their endeavour to optimize formulation both in terms of delivery and placement of sufficient amounts of BRG100 product to inhibit or suppress germination of the weed, which is one of the key challenges in bioherbicide product development. Recently, Hynes and Boyetchko (2011) reported modifications to the formulation to promote disintegration and dispersion of the granular bioherbicide. Fig. 2 presents experimental results vis-à-vis efficacy of formulations with and without the BRG100 Pesta on green foxtail at AAFC Research Farm using application rates ranging from 0.50 to 2 kg ha1. The application of BRG100 Pesta involved experimental plots in-furrow with green foxtail. For commercial field application, BRG100 Pesta will be applied in-furrow or side-banded with crop seed. Specific experimental results are unpublished given the work in progress. The

523

E. Mupondwa et al. / Bioresource Technology 175 (2015) 517–528

Fig. 2. Efficacy of formulations with and without the BRG100 Pesta on green foxtail at AAFC Research Farm (AAFC unpublished data).

researchers are also exploring potential additional benefits of BRG100 in integrated weed management (IWM) based on observations that reducing the competition of the weed ‘‘in-crop’’ may enhance bioherbicidal potential. Based on the above elaboration and application rates being optimized by AAFC researchers, this study uses an application rate of 1 kg ha1 for BRG100 to investigate profitability. The above comparative application rates are also consistent with compatibility tests recently conducted by Casella et al. (2010) in their study evaluating the efficacy and technological feasibility of three bioherbicide candidates (Drechslera gigantea, Exserohilum rostratum, and Exserohilum longirostratum) against an Italian population of green foxtail. In order to investigate the relationship between profitability and product price, a price range of $5–$50 kg1 is employed. To further account for potential wide stochastic movement in prices caused by seasonal variations in the demand for bioherbicides, a BRG100 selling price of $100 kg1 is also included (taking into consideration the range of Crop Protection Guide application costs of $13–$100 ha1 for chemical herbicides, with a majority of costs clustered in the $25–$50 ha1 range per kilogram of active ingredient as noted earlier). The price of $100 kg1 is clearly an extreme and hence the computed NPV corresponding to such higher price ranges will be inflated to this degree. This point is discussed further in the results section. 2.4.2. Interest expenses, inflation, and investment risk In order to account for inflation, all costs and prices were adjusted annually using a mean inflation rate in 2013 prices. All dollar values are in Canadian dollars, although approximate parity with the US dollar can be assumed for purposes of comparison. The baseline cost of borrowed capital is based on Bank of Canada prime rate and includes the Government of Canada bond rate, a market risk premium, and a liquidity risk premium which adds up to the unlevered cost of capital. The market risk premium is the difference between the expected return on an investment and risk-free rate (Ross et al., 2007). In Canada, the Canadian federal government treasury bill is a risk-free rate. Liquidity risk premium is the component of the nominal interest rate representing compensation for investors for any lack of liquidity, based on the assumption that investors always prefer liquid assets (Ross et al., 2007). Three discount rates (d = 7%, 10%, and 15%) are used in this study to represent the cost of capital and compute present discounted value of costs and revenues as well as assess the sensitivity of the BRG100 fermentation plant to various opportunity costs of fermentation plant investment capital. Typically, a private firm is less willing to accept and spread risks compared to a government project; hence the discount rate for a private investor will typically be higher relative to the social discount rate often

applied in the evaluation of publicly funded research and development projects.

3. Results and discussion 3.1. Capital cost Table 3 provides details of the distribution of capital costs along with major equipment costs. Total equipment cost for a BRG100 bioherbicide fermentation plant at baseline capacity (two fermenters each having a working volume of 33,000 L) is $2.58 million. Approximately 50% of the equipment cost is associated with two fermenters. Total DFC for the plant at this capacity (3602 tonnes annum1) is $15.31 million and total capital investment including working capital, R&D, and cost of start-up and validation is approximately $17.55 million, which is 7 times the equipment purchase cost.

Table 3 Capital costs for BRG100 fermentation plant in the base case. Items Direct Fixed Capital Cost (DFC) Equipment Purchase Cost Receiver Tank Water Distiller Fermenter Extruder Grinder Belt conveyor Screw conveyor Centrifugal pump Fluid bed dryer Blending tank Pesta mixer tank Unlisted equipment Installation Process piping Instrumentation Insulation Electrical Buildings Yard improvement Auxiliary facilities Engineering Construction Contractor’s fee Contingency Working capital Start-up and validation cost Up-front R&D Up-front royalties Total capital investment

Unit price ($)

50,000 12,000 673,000 207,000 74,000 53,000 9,000 9,000 33,000 167,000 230,000

Costs ($) 15,313,000 2,582,000 100,000 12,000 1,346,000 207,000 74,000 53,000 9,000 63,000 33,000 167,000 230,000 258,000 886,000 904,000 1,033,000 77,000 258,000 1,162,000 387,000 1,033,000 2,081,000 2,913,000 666,000 1,332,000 965,000 766,000 500,000 10,000 17,554,000

524

E. Mupondwa et al. / Bioresource Technology 175 (2015) 517–528

3.2. Operating cost Estimated operating cost is presented in Table 4. A plant producing 3602 tonnes of BRG100 annum1 has total annual operating cost of $14.76 million, for a unit production cost of $4.10 kg1. Raw materials are the most important item, accounting for nearly 50% of the operating cost. Facility-dependent costs account for 23% of total operating cost. Labour cost and laboratory cost come in third and fourth, respectively. Costs were also evaluated at fermentation plant scales of 360–36,020 tonnes annum1. Fig. 3 depicts unit production and investment costs versus BRG100 production scale, and shows that unit production costs decrease as plant capacity increases; however the rate of decrease slows down significantly after plant capacity reaches 7204 tonnes year1 (22 tonnes day1), a reflection of economies of scale. The scale of production has a significant impact on unit costs of operating, with substantial reductions in cost associated with higher scales of operation. 3.3. Profitability and uncertainty analysis Profitability and uncertainty analysis was conducted within Intelligen SuperPro DesignerÒ in terms of the sensitivity of NPV to changes in base case parameters; namely: changing production scale (from 360 to 36,020 tonnes annum1, using 3602 tonnes annum1 as baseline); changing BRG100 selling price ($5–$100 kg1 based on price points for alternatives described in Section 2.4.1); variation in the cost of capital (7–15%). Fig. 4 shows that at the base case capacity of 3602 tonnes annum1 and given a baseline 7% cost of capital, NPV ranges between $21.6 million (for BRG100 selling price of $5 kg1) to $3.1 billion (at a BRG100 selling price of $100 kg1 – which is the extreme case). Fig. 5 presents the impact on NPV of increasing the cost of capital from 7% to 15% while facing a much lower BRG100 price of $10 kg1. This is assessed for small to large plant sizes. Although large BRG100 plants can generate comparatively higher NPV at lower prices, they have steeper NPV curves, suggesting that they are more sensitive to changes in the opportunity cost of capital. The results provide insight into prices required to sustain profitability across various plant sizes when facing increased capital costs and reduced prices. One of the issues in the competitiveness of a BRG100 fermentation plant at various capacities relates to its response to very low prices and high cost of capital. Fig. 6 depicts this impact for a fermentation plant facing 15% cost of capital, and a much more stringent BRG100 retail price band of $3–20 kg1 over a reduced production capacity range of 360–7204 tonnes annum1. This result demonstrates the importance of price mark-up for certain scales of operation. At lower BRG100 prices and a 15% cost of capital, smaller BRG100 fermentation plants require considerably higher NPV break-even prices. For instance, at a BRG100 price of $3 kg1, all BRG100 plants with capacities of 6000 tonnes annum1 or less

generate negative NPV while prices of $4, $5 and $10 kg1 result in negative NPV for BRG100 fermentation plant capacities of less than 5000, 3000, and 1000 tonnes annum1 respectively. As shown earlier in Fig. 3 above (Section 3.2), unit production and investment costs are disproportionately higher for this range of capacity compared to fermentation capacity ranges of 7000 tonnes year1 (22 tonnes day1) and higher, a reflection of economies of scale. The smallest BRG100 fermentation plant capacity modelled here is 360 tonnes annum1, which is roughly 1.1 tonnes day1, and whose unit operating cost for one kg of bioherbicide is more than $16, resulting in negative NPV if the market determined price is $10 kg1. If the selling price is set at $10, the plant could reach positive NPV after producing 1000 tonnes of BRG100 each year, which corresponds to a production rate of 3 tonnes day1. Overall, the results again suggest profitability over a noticeable range of operating scale, price variance, and capital cost variance. However, a drastic reduction in price coupled with a high cost of capital and constrained output could adversely affect the viability of the operation. A pertinent question relates to market size; whether the scales of BRG100 production analysed here seem reasonable in relation to potential market size for BRG100. The scope of weeds covered in this invention clearly demonstrates the breadth of potential commercial application of BRG100 in conventional agriculture currently dominated by chemical herbicides. As noted in the introductory section, this potential is further enhanced by the fact that most of these chemical herbicides can no longer be used effectively in weed control because a majority of weeds, including some of the most prevalent and troublesome weeds like green foxtail and wild oat, have developed resistance to chemical herbicides (Beckie et al., 2013; Heap, 2014). In spite of this potential, bioherbicides are often difficult to commercialize for various reasons including low uptake by large industries (Ash, 2010; Charudattan and Dinoor, 2000; Glare et al., 2012; Bailey and Falk, 2011). In this case, the BRG100 fermentation plant may only operate periodically during the year if commercialization constraints limit BRG100 to niche market applications. Table 5 summarises results corresponding to the BRG100 fermentation plant operating less than one year relative to the base case. It can be seen that the plant could approach positive NPV for operating periods as low as 2 months. However, it would be constrained when facing low prices and high capital costs (as in Fig 6 above). Operating the plant for more than 6 months would reduce the investment payback period by less than 2 years. In spite of the positive NPV obtained over a wide range of capacity for this technology, the market for herbicides limits the amount that can be feasibly and realistically absorbed by the market. Hence, it is important to view these computed NPV values in terms of the share of the herbicide market that BRG100 can capture in mainstream on-farm crop applications by farmers. For instance,

Table 4 Operating cost for BRG100 fermentation plant in the base case. Cost items

Annual cost ($ annum1)

Unit cost ($ kg1 of BRG100)

Raw materials Labour-dependent Facility-dependent Laboratory/QC/QA Consumables Utilities Electricity Steam Cooling water Miscellaneous Advertising/selling Total

7,128,461 3,183,418 3,453,093 477,513 1,099 303,746 168,748 41,484 93,514 200,000 10,000 14,757,330

1.98 0.88 0.96 0.13