Industrial Crops and Products Single cell oil production ... - Unicamp [PDF]

Industrial Crops and Products 89 (2016) 478–485

Contents lists available at ScienceDirect

Industrial Crops and Products journal homepage: www.elsevier.com/locate/indcrop

Single cell oil production integrated to a sugarcane-mill: Conceptual design, process specifications and economic analysis using molasses as raw material J.P.F. Vieira a,b , J.L. Ienczak b , P.S. Costa a,b , C.E.V. Rossell b , T.T. Franco a , J.G.C. Pradella b,∗ a

School of Chemical Engineering University of Campinas, UNICAMP, CEP 13083-852 Campinas, São Paulo, Brazil Brazilian Laboratory of Science and Technology of Bioethanol (CTBE), Brazilian Center for Research in Energy and Materials (CNPEM) Rua Giuseppe Máximo Scolfaro, 10.000–Polo II de Alta Tecnologia, Caixa Postal 6192, 13083-970 Campinas, SP, Brazil b

a r t i c l e

i n f o

Article history: Received 26 January 2016 Received in revised form 21 May 2016 Accepted 25 May 2016 Available online 10 June 2016 Keywords: Single cell oil Sugarcane mill integration Molasses Process design Economic analysis

a b s t r a c t This study aimed to assess the single-cell oil production integrated to a sugarcane-mill and molasses as main raw material. Rhodotorula glutinis lipid production with low-cost raw sugarcane molasses was accessed. The envisage process relies in an optimized sugarcane molasses culture media in a fed-batch strategy to achieve a high-cell density high-lipid cell content. A Process Flow Design (PFD) integrated to a typical Brazilian sugar mill was proposed according to fermentation experiments. Performed economic analysis provided data for discussion on strategic equipment that can facilitate the viability to the process. Typical results of established protocol were total biomass concentration 62.25 g l−1 , lipid productivity 0.42 g l−1 h−1 , sugar conversion yield to lipids 0.21 g of lipid g−1 of total reducing sugar. Microbial lipid production and defatted biomass plant production was proposed at a nominal capacity of respectively 16,720 ton and 21,600 ton per year, running 8300 h per year. The projected selling prices of these products were US$ 1,300.00/ton of lipid and US$ 500.00/ton of defatted yeast. It was demonstrated that the industrial plant was potentially attractive when capital investment cost was decreased with the use of low cost epoxy-lined carbon steel stirred bioreactor. In this alternative, the internal rate of return (IRR) will be 24.61% leading the NPV (at 7% interest rate) of US$ 67,797,000/year (before taxes). © 2016 Elsevier B.V. All rights reserved.

1. Introduction The fossil fuels shortage and lack of sustainability raised question on its usage in a long-term basis. They are responsible to increase CO2 level in the atmosphere which is directly associated with global warming. It is a consensus that global energy towards to energy from renewable sources and consequently, there is renewed interest in the production and use of fuels from plants or organic waste (Macrelli et al., 2012). Production of a cheap renewable fuels to replace fossil fuels is subject in the political agendas of many countries, aimed at the development of a reliable energy source to ensure fuel security, promote rural development and to address climate change by reducing greenhouse gases emission (Macrelli et al., 2012). Among proposed renewable fuel, biodiesel (methyl and/or ethyl esters of long chain fatty acids of vegetable oils) presents as the fuel

∗ Corresponding author. E-mail address: [email protected] (J.G.C. Pradella). http://dx.doi.org/10.1016/j.indcrop.2016.05.046 0926-6690/© 2016 Elsevier B.V. All rights reserved.

with the highest potential to replace diesel oil (Leung et al., 2010). Usually the industry adopted for biodiesel production the transesterification reactions of vegetable oils (soybean oils, rapeseed oils, palm oils, and its waste cooking oils) with methanol or ethanol. Vegetal oil high price and low vegetable oil productivity has being hindered their use in large scale as its production costs are associated with raw material mainly with raw-material price up to 80% of its value (Haas et al., 2006). In recent years, much attention has been done to the exploration of microbial oils, which might become one potential oil source for biodiesel production (Li et al., 2007). Oils from microorganisms, known as single cell oils (SCO) emerge as a potential substitute for vegetable oils and presents 80–90% of the triglycerides content (Ratledge, 1991). On the other hand, SCO has become an important source of polyunsaturated fatty acids (PUFA). PUFA are present in cellular membrane of neurons, sensory cells, sub-cellular organelles and play important functional role as precursors of prostaglandins, lipoxins and leukotrienes. Moreover, specific PUFAs are recommended for the prevention and treatment of cardiovascular and inflammatory diseases, brain disorders and obesity. Rele-

J.P.F. Vieira et al. / Industrial Crops and Products 89 (2016) 478–485

479

Table 1 Kinetic parameters for some high-cell-density Rhodotorula glutinis and Rhodosporodium toruloides (Rhodotorula glutinis anamorph state) cultivation in bioreactor aiming to lipid production. Yeasts

␮max (h−1 )

Substrate

Xtotal (g l−1 )

Lipid (%; w w−1 )

Pr (g l−1 h−1 )

R. glutinis CCT 2182 R. glutinis NRRL Y 1091 R. toruloides Y4 R. toruloides Y4 R. glutinis CGMCC 2.703

0.235

Sugarcane molasses

Reference

62.2

54

0.42

Improved fed strategy; this study

0.260

Glucose

185.0

40

0.88

Pan et al. (1986)

0.130 – –

Glucose Glucose Corncob hydrolysate

106.5 127.3 70.8

67 62 47

0.54 0.57 0.17

Li et al. (2007) Li et al. (2007) Liu et al. (2015)

Table 2 Equipment list of microbial lipid production. Name

Type

V-101 HX-101 PM-101 V-102 HX-105 V-103 PM-102 V-104 HX-102 PM-103 V-105 HX-103 V-106 PM-104 R-101 R-102 R-103 PM-105 PM-106 CF-101 V-109 HX-104 PM-108 V-110 SDR-101 HX-106 M-101 PM-107 PM-109 HG-101 DC-101 SL-102 TFE-101 V-108 HX-107 V-107 SL-103 V-111 PM-110 HX-108 PC-101

Receiver Tank Heat Exchanger Centrifugal Pump Flash Drum Heat Exchanger Receiver Tank Centrifugal Pump Receiver Tank Heat Exchanger Centrifugal Pump Flash Drum Heat Exchanger Receiver Tank Centrifugal Pump Stirred Reactor Stirred Reactor Stirred Reactor Centrifugal Pump Centrifugal Pump Centritech Centrifuge Receiver Tank Heat Exchanger Centrifugal Pump Flash Drum Spray Dryer Heat Exchanger Centrifugal Fan Centrifugal Pump Centrifugal Pump Homogenizer Decanter Centrifuge Silo/Bin Thin Film Evaporator Receiver Tank Heat Exchanger Blending Tank Silo/Bin Receiver Tank Centrifugal Pump Heat Exchanger Pneumatic Conveyor

Units 1 1 1 1 1 1 1 1 1 1 1 1 1 1 3 3 11 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1

Size 10.00 70.00 0.25 0.20 50.00 5.00 0.22 10.00 500.00 2.05 1.00 50.00 10.00 1.87 2.00 45.00 500.00 0.01 0.22 55.00 5.00 155.00 0.76 1.00 130.41 15.00 55,292.91 0.70 2.05 40.00 40.00 75.00 20.00 3.80 37.44 10.00 2.00 5.00 0.08 1.00 150.00

vant PUFA for those applications are ␣-linolenic acid (C18:3n–3, ALA), ␥-linolenic acid (C18:3n–6, GLA); eicosapentaenoic acid (C20:5n–3; EPA), and docosahexaenoic acid (C22:6 n–3, DHA) that are becoming high demanded in food and pharmaceutical industries. Heterotrophic fungi (Mortierella, Pythium, Mucor), and bacteria (Shewanella, Moritella) produce EPA-rich oil in relatively large amounts from simple carbon. Marine organisms, such as dinoflagellates and traustochytrid protists, such as Crypthecodinium cohnii and Traustochytrium accumulates more than 50% of their lipids as DHA using submerged cultivation in bioreactor being important sources of industrial PUFA producers (Bellou et al., 2016; Ratledge, 2013). SCO production of lipids is regarded as a partially growthassociated bioprocess. In general, it is explored in the literature as

m3 m2 kW m3 m2 m3 kW m3 m2 kW m3 m2 m3 kW m3 m3 m3 kW kW m3/h m3 m2 kW m3 m3 m2 m3/h kW kW m3/h m3/h m3 m2 m3 m2 m3 m3 m3 kW m2 m

Material

Purchase Cost ($/Unit)

CS CS SS316 CS CS CS SS316 CS CS SS316 CS CS CS SS316 SS316 SS316 SS316 SS316 SS316 SS316 CS CS SS316 CS SS316 CS CS SS316 SS316 SS316 SS316 Concrete SS316 CS CS SS316 CS CS SS316 CS CS

33,000 104,000 11,000 3000 85,000 20,000 10,000 33,000 339,000 27,000 8000 85,000 33,000 26,000 618,000 1,054,000 2,818,000 10,000 10,000 344,000 20,000 168,000 17,000 8000 312,000 41,000 19,000 17,000 27,000 161,000 293,000 75,000 499,000 20,000 72,000 265,000 75,000 20,000 10,000 32,000 97,000

a two phases bioprocess: a growth phase followed by a lipid production (accumulation) phase (Ratledge and Cohen, 2008). During growth phase microorganism is cultivated in a well balance culture media aiming to the cell proliferation at high specific growth rate to obtain a high-cell concentration. The accumulation phase is achieved afterwards due to a limited nutrient (nitrogen, phosphorus and/or sulfates) concentration and a non-limited carbon source policy (Papanikolaou et al., 2009). The high carbon-nutrient ratio triggers intracellular lipid accumulation from acetyl-CoA that is directed to the de novo lipid biosynthesis (Ratledge, 2014). Literature also reported the limitation of iron and oxygen as affecting the lipid accumulation (Granger et al., 1993; Hassan et al., 1993). The carbon source and investment costs in bioreactor are of great impact in SCO production cost. Therefore, high-productivity and

480

J.P.F. Vieira et al. / Industrial Crops and Products 89 (2016) 478–485

high-lipid concentration and the use of low-cost carbon sources, such as agricultural residues are pivotal to achieve lipid bioprocess production at a competitive price (Ageitos et al., 2011). In this context several attempts has being carried out to develop a high performance bioprocess using a diversity of carbon sources. The two-phases (growth phase followed by accumulation phase) fedbatch process carried out for several investigators in the past years is an attractive procedure (Li et al., 2007; Zhu et al., 2008; Zhao et al., 2012; Munch et al., 2015). It is flexible to control the growth and lipid accumulation phases by modifying the feed flowrate of carbon source and limiting nutrient throughout the fermentation process and is largely applied in bioprocessing to achieve high-cell density high-productivity production of biomolecules. In theory, during the late stage of the growth phase, nitrogen (and/or phosphorus) sources exhaustion (or limitation) impairs the cellular growth, which leads to a reduction in the cell mass production rate and the channeling of the flux of carbon toward lipid biosynthesis. Recently well-succeeded attempts to produce lipids using solid-state fermentation directly from cellulosic indicated that this alternative route might be an interesting option to decrease lipid production cost (Cheirsilp and Kitcha, 2015). Sugarcane (Saccharum sp) molasses is a raw material rich in salts and carbohydrates (sucrose, glucose and fructose). Molasses is a residue from sugar (sucrose) production from sugarcane and is produced at a rate of about 40–60 kg of molasses/ton of processed sugarcane. (Meade and Chen, 1977). We recently demonstrated the potential of lipid production from sugarcane molasses using the oleaginous yeasts Rhodotorula glutinis, Rhodosporidium toruloides, Rhodotorula minuta and Lipomyces starkey (Vieira et al., 2014). Taking into account the amount of sugarcane devoted to sugar production, that was about 280 mton harvested sugarcane in the 2015/2016 season (CONAB, 2015), molasses production in Brazil was estimated to be 13.000.000 t/year, with carbohydrate content of about 85% m/m. This carbohydrate therefore is a potential carbon source to be used for SCO production. Moreover, the integration of SCO production in a sugarcane-mill would help to decrease lipid cost because of ease of availability of raw material, utilities (steam and electricity), effluent treatment and disposal and logistic distribution. This scheme was proposed in the past in order to produce bacterial poly-hydroxyalkanote biodegradable polymer from sugarcane carbohydrates (Rossell et al., 2001; Bueno Netto et al., 2000). In this context, this study aimed to develop a process design to produce lipid using sugarcane molasses at high-cell-density and high-lipid-cell content with a selected oleaginous yeast. The process was integrated in silico on a typical sugarcane mill Brazilian industry, and accessed its economic viability based on kinetic results for different scenarios.

2. Material and methods 2.1. Microorganisms and culture media R. glutinis CCT 2182 was obtained from the Andre Tozello Foundation Culture Collection (Campinas, SP, Brazil) and it was previously selected as detailed elsewhere (Vieira et al., 2014). The first seed culture medium was YPD, composed of yeast extract (3.0 g l−1 ), peptone (5.0 g l−1 ), and glucose (10 g l−1 ). The second seed culture employed a modified mineral medium containing (g l−1 ): glucose (20.0), (NH4 )2 SO4 (1.0), KH2 PO4 (1.0), Na2 HPO4 ·12 H2 O (1.0), Mg2 SO4 ·7H2 O (2.0), NaCl (1.0), CaCl2 (0.02), FeCl3 ·6H2 O (0.01) and yeast extract (2.0). The cultures were incubated in a shaker (Excella E24, New Brunswick Scientific, Edison, USA) for 24 h at 250 rpm, 28 ◦ C, and pH 5.8, and were then employed as inoculum in the bioreactor cultures. The main substrate used in the bioreactor culture media was Brazilian molasses (obtained from Usina da Pedra,

Table 3 Direct Fixed Capital Cost of microbial lipid production (considering 316L stainless steel bioreactor). Parameters

Value (US$)

% Investment

Factors

Equipment Purchase Cost (EPC) Installation Process Piping Instrumentation Insulation Electrical Buildings Yard Improvement Auxiliary Facilities Engineering Construction Contractor’s Fee Contingency Direct Fixed Capital Cost

39,442,000 12,205,000 2,367,000 3,944,000 394,000 789,000 2,367,000 394,000 2,367,000 6,427,000 6,427,000 3,856,000 7,712,000 88,691,000

44.47% 13.76% 2.67% 4.45% 0.44% 0.89% 2.67% 0.44% 2.67% 7.25% 7.25% 4.35% 8.70% 100.00%

1.00 EPC 0.31 EPC 0.06 EPC 0.10 EPC 0.01 EPC 0.02 EPC 0.06 EPC 0.01 EPC 0.06 EPC 0.16 EPC 0.16 EPC 0.10 EPC 0.20 EPC 2.25 EPC

Table 4 Operating cost of microbial lipid and deffated biomass production on industrial plant. Cost Item Raw Materials Molasses Amm. Sulfate Water Air Labor-Dependent Facility-Dependent Utilities Std Power Steam Cooling Water Chilled Water Total

Amount 123,557 9,586,721 330,174,407 3,212,261,717

194,548,440 898,276 26,521,961 5,070,426

Unit ton kg kg kg

kW-h ton ton ton

$

%

13,314,000 12,355,694 958,672 0 0 2,122,000 5,744,000 7,759,000 5,836,453 89,828 1,326,098 507,043 28,940,000

46.01

7.33 19.85 26.81

100.00

São Paulo, Brazil), composed of sucrose (356.87 ± 9.08 g l−1 ), glucose (152.35 ± 0.55 g l−1 ), fructose (121.30 ± 1.21 g l−1 ), and NH4 + (0.5 ± 0.05 g l−1 ), with a C/N ratio of 40. 2.2. Fed strategy In order to obtain high-cell-density with high lipid content, a carbon non-limited fed batch feeding strategy was used and described as following. A suspension of R. glutinis CCT 2182 previously prepared as in item 2.1 was inoculated at 10% of working volume in a bench bioreactor (Bioflo 115, New Brunswick Scientific, Edison, USA) with 2 l working volume. The bioreactor has an initial total reducing sugar (TRS) concentration from molasses 18.0 g l−1 , (NH4)2 SO4 2.3 g l−1 supplemented with 0.80999 g KH2 PO4 ; 0.73636 g Mg2 SO4 and 0.73636 g yeast extract per 10 g of total reducing sugar. Soon after inoculation, an exponential fed of sugarcane molasses solution (SCMS) at 300 g TRS l−1 was set up in order to keep TRS concentration in fermentation broth in a non-limited concentration (in the range of 10–20 g l−1 ), according to Eq. (1) F=

x Xo Vo ex t (So − S) YX/S

(1)

where, F is the volumetric flow rate of SCMS varied with time t, x is the specific growth rate, Yx/s is the cell mass yield factor from consumed TRS, Vo is the initial volume of fermentation broth in bioreactor, Xo is the initial cell concentration and So is the TRS concentration in the feed stream. The adopted parameters values for Eq. (1) were: So = 300 g l−1 x = 0.23 h−1 , Yx/s = 0.53 g biomass g−1 TRS; Vo = 1.5 l; and Xo = 1.00 g l−1 . In this phase the pH 4.8 was controlled by a solution of ammonium hydroxide (6.25% by volume) in order to provide a

J.P.F. Vieira et al. / Industrial Crops and Products 89 (2016) 478–485 Table 5 Economic evaluation considering 316L stainless steel bioreactor. Parameters

Value

Unit

Annual operating time Total Capital Investment Direct Fixed Capital Working Capital Startup Cost Capital Investment Charged to This Project Operating Cost Main Revenue Other Revenues Total Revenues Cost Basis Annual Rate Unit Production Cost Unit Production Revenue Gross Margin Return On Investment Payback Time IRR (Before Taxes) NPV (at 7.0% Interest)

8.300 95,137,000 88,691,000 2,012,000 4,435,000 95,137,000 28,940,000 22,572,000 10,804,339 33,376,000 16,719.84 1,730.86 1,996.20 13.29 7.62 13.13 8.05 8,323,000

h $ $ $ $ $ $/yr $/yr $/yr $/yr ton MP*/yr $/ton MP* $/ton MP* % % years % $

*MP: Main product.

non-limited nitrogen regime. As soon as the yeast cell concentration reached about 20.0 g l−1 , the pH control solution was changed from ammonium hydroxide solution to NaOH solution 3 M, in order to impose a nitrogen-limited regime, triggering therefore lipid accumulation. Temperature, pH, aeration, and agitation rate were automatically controlled at 28 ◦ C and 4.8. respectively. The aeration and agitation rate were automatically controlled in the interval of 0.5–2 vvm (1.25–5.0 l of air min−1 ), and 250–1200 rpm to maintain dissolved oxygen concentration in fermentation broth (D.O.) at 20% of air saturation. The experiments were carried in triplicate and results are presented as mean values and standard deviation. 2.3. Chemical analyses Sampling was withdraw periodically, centrifuged at 10.000g, 10 ◦ C, and chemical analyses performed in supernatant and yeast cell sediment. The concentrations of glucose, fructose, and sucrose were measured in the supernatant using high performance liquid chromatography (HPLC) Ultimate 3000, Dionex. The total concentration of reducing sugars (TRS) was calculated as the sum of the concentrations of the individual sugars. Ammonia nitrogen concentration was determined in the supernatant accordingly (Srienc et al., 1984); the total lipid concentration was determined gravimetrically and calorimetrically after lysis of the yeast (Fales, 1971; Bligh and Dyer, 1959). Disruption of the yeast cells and lipid recovery was performed as described (Vieira et al., 2014). Methyl esters of fatty acids were obtained by direct transesterification (Lewis et al., 2000) and determined by gas chromatography. Detailing of applied methodologies were described elsewhere (Vieira et al., 2014).

481

of 35 years; (iv) NPV interest rate of 7%; 10% and 12%; (v) 100% of outlay of investment in the 1st year; (vi) depreciation straight line over 30 year; (vii) salvage value of direct fixed capital of 5%; (viii) maintenance of 1.7% of direct fixed capital per year. (ix) insurance of 0.5 % of direct fixed capital per year; (x) taxes of 1.0 % of direct fixed capital per year; (xi) factory expense of 1.0 % of direct fixed capital per year; (xii) working capital of 30.0 % of raw material; utilities and labor; (xiii) startup cost of 5.0% of direct fixed capital; (xiv) selling taxes of 0% of gross profit. SuperPro Designer v8 (Intelligen, Inc., USA) was used for estimation of fixed capital and operating costs. Required power for agitation in fermenter Pf (kW) was estimated by Eq. (1) (Koutinas et al., 2014). 2.5. Lipid stoichiometry equation The adopted stoichiometry of lipid and yeast production from glucose (Eqs. (2) and (3)) was described elsewhere (Koutinas et al., 2014): C6 H12 O6 + 4.32O2 + 0.54C5.35 H9.85 O2.45 N1.5 → 1.12C4 H6.5 O1.9 N0.7 + 4.41CO2 + 5.02H2 O

C6 H12 O6 + 1.26O2 → 0.06C57 H104 O6 + 2.61CO2 + 2.92H2 O

(2)

(3)

Eq. (4) shows a modification of Eq. (2) for yeast biomass production from the ammonium ion present in the ammonium sulfate (NH3 + ) and glucose (stoichiometric balance). C6 H12 O6 + 2.87O2 + 0.26(NH4 )2 SO4 → 0.76C4 H6.5 O1.9 N0.7 + 2.97CO2 + 4.60H2 O + 0.26SO4 −2

(4)

The theoretical mass conversion yield of glucose and nitrogen from ammonium sulfate to yeast biomass are equal to 0.40, and 9.67, respectively. The theoretical mass conversion yield of glucose and nitrogen from yeast extract to yeast biomass and lipids are equal to 0.589, 9.35 and 0.295, respectively. According to the results obtained in this study, the conversions of glucose to yeast biomass and glucose to nitrogen are equal to 0.42, and 8.71, respectively. The conversion values were intermediate between cell growth with ammonium sulphate and cell growth with yeast extract due to supplementation of the medium with a mix of these nitrogen sources. However, the stoichiometric equation by Koutinas et al. (2014) did not represent the lipid fraction of the yeast biomass, which is approximately a lipid mass fraction of 7% (m/m). Then one can rearrange these equations for the cell mass production according to Eqs. (5) and (6), using respectively ammonium sulfate and yeast extract. C6 H12 O6 + 2.87O2 + 0.26(NH4 )2 SO4 → 0.76C3.57 H5.72 O1.86 N0.7

2.4. Economic analysis hypothesis Microbial lipid production and defatted biomass plant production was proposed at a nominal capacity of respectively 16,720 ton and 21,600 ton per year, running 8300 h per year. The designed industrial plant is supposed to sell the produced lipid and the defatted yeast as animal food supplement. The projected selling prices of these products were US$ 1,300.00/ton of lipid and US$ 500.00/ton of defatted yeast. Labor, molasses, and nitrogen were respectively US$ 5.29/h of labor, US$ 100.00/ton of molasses raw material and US$ 0.100/ton of nitrogen source. Purchase prices of water, saturated steam and electricity were set at US$ 0.050/ton water, US$ 0.100/ton steam and US$ 0.030/kW h. The project design economic evaluation used the following hypothesis: (i) construction period of 12 months; (ii) startup period of 6 months; (iii) project lifetime

+2.97CO2 + 4.60H2 O + 0.01C57 H104 O6 + 0.26SO4 −2

(5)

C6 H12 O6 + 4.23O2 + 0.54C5.35 H9.85 O2.45 N1.5 → 1.12C3.71 H1.70 O1.86 N0.72 + 4.41CO2 + 5.02H2 O +0.01C57 H104 O6

(6)

2.6. Power consumption The required power for agitation in fermenter Pf (kW) was estimated by Eq. (7) (Koutinas et al., 2014). Power consumption of other

482

J.P.F. Vieira et al. / Industrial Crops and Products 89 (2016) 478–485

Fig. 1. Time evolution for Rhodotorula glutinis CCT 2182 bioreactor cultivation in the non-carbon limited fed strategy and supplemented culture media: ( concentration; () TRS concentration; ( ) nitrogen concentration ( and standard deviation of triplicate experiments.

) total yeast biomass concentration; (

devices was estimated with the aid of Superpro® Designer according to the flow rate, density of the fluid estimated pressure pipeline and pressure on the equipment (see Table 5 ). Pf =

1 2.8Vf ˛

(7)

where:␣ = electrical motor efficiency (0.9)= Fraction working volume/size of fermenterVf = Fermenters working volume (m3 ) 3. Results and discussion 3.1. Bioreactor strategy for lipid production The present study proposed a SCO production approach that emulated bacterial polyhydroxyalkanoates cytoplasm accumulation applied in past studies (Diniz et al., 2004; Pradella et al., 2010). The protocol comprises the fed-batch high-cell-density growth phase followed by a fed-batch nitrogen-limited lipid accumulation phase (Fig. 1a). R. glutinis grown attained total yeast biomass concentration (X), maximum specific growth rate (x ) lipid productivity (Pr ) and lipid yield from TRS respectively of 62.25 g l−1 , 0.235 h−1 ; 0.425 g l−1 h−1 and 0.171 g of lipid g−1 of TRS. The cell mass kinetic displayed an exponential profile until approximately 24 h, indicating a balanced growth for proposed culture medium and conditions. Due to nitrogen limitation after 24 h, defatted biomass concentration (Xr) remained constant at approximately 30 g l−1 . Lipid accumulated in yeast cell climbed to 54% of dry cell weight at the end of experiment. A comparison of the kinetic data obtained in this study with the results obtained by other researchers that used R. glutinis (or its anamorph form Rhodosporodium toruloides) aiming the production of lipids from a diversity of carbohydrates was carried out (Table 1). Reported lipid production from glucose attained higher values of cell concentration and lipid productivity than obtained results and similar amount of cell-lipid content (Table 1). Although we tried maintenance of D.O. above 20% of air saturation during the experiments, this was not achieved between 40 h to 60 h fermentation time. In this period, the D.O. went down to less than 2% while aeration and agitation rate climb at its maximum levels, respectively, 5.0 l of air min−1 and 1200 rpm (data not shown). We want to emphasize

) total lipids

) defatted biomass concentration. Points are mean values

that in the proposed protocol the broth dissolved oxygen concentration was automatically controlled by variation of aeration and agitation rate. Cultivation in previous studies (Liu et al., 2015; Zhao et al., 2012) used oxygen-air enrichment in order to supply the high oxygen demand due to high-cell density cultivation. These strategies however would pose an undesirable cost increment in the SCO production and generally are only justified for high-value products. Lignocellulosic hydrolysate composed a mixture of glucose and xylose, submitted to detoxification and supplemented with salt solution was also recently explored as alternative carbon source in order to decrease the impact of raw material cost. However, very low Pr was obtained (about 0.2 g l−1 h) (Liu et al., 2015; Anschau et al., 2014) (Table 1). This would probably hamper its commercial viability due to the high investment costs in bioreactor. The results presented by this study led to lower results in terms of cell density and cell-lipid content but have the merit of the usage of sugarcane molasses, a carbon low-cost agricultural residue commercialized at about US$ 100/ton in this country. 3.2. Proposed process flow diagram Exploitation of microorganisms as a source of bioenergy production could meet technical and economic conditions as biofuel resource. Some advantages has being raised in literature as cost competitiveness with petroleum fuels, air quality improvement (e.g. CO2 sequestration), and minimal water usage requirements (Tabatabaei et al., 2011). Some studies of new technologies integrated in sugar mills has been done recently (Dias et al., 2011; Furlan et al., 2013). Residual sugars such as molasses obtained from the sucrose commercial sugar refining is a potential carbon source material candidate as it is composed mainly by sucrose, glucose and fructose, salts and vitamins (Hugot, 1969; Meade and Chen, 1977). Its usage was recently proposed as raw material for lipid production utilizing ligneous yeasts (Vieira et al., 2014). A Process Flow Diagram (PFD) for microbial lipid production from molasses was therefore proposed based on the developed protocols and includes the main streams and equipment involved in this process (Fig. 2). The stream components and its specification are shown (Annex A, Table A3). In short, culture medium composed of sugarcane molasses supplemented (MO001/002 + NS001/002)

J.P.F. Vieira et al. / Industrial Crops and Products 89 (2016) 478–485

483

Fig. 2. Process flow diagram (PFD) for microbial lipid production from sugarcane molasses.

are pumping to heat exchangers (HX-101 and HX-102) to attain adequate disinfection level and received in tanks V-103 and V-106. Culture media ME-101 from V-103 is supplied to bioreactors R101 and R-102 to produce yeast inoculum. Culture media ME-102 from V-106 is supplied to R-103 bioreactors for lipid process production. Yeast oleaginous suspension (YO-001) from bioreactor is

centrifuged in CF-101 and the produced yeast cream is storage in receiver tank V-109. Yeast concentration and drying is carried out in flash drum (V-110) followed by spray drying process (SDR-101). The dried yeast Y-003 is suspended in hexane HX-001 in the blending tank V-107 and disrupted in the high-pressure cell homogenizer HG-101. Defatted biomass (RYD-001) is separated from hexane oil

484

J.P.F. Vieira et al. / Industrial Crops and Products 89 (2016) 478–485

solution (OS-001) in decanter DC-101, cleaned from solvent (not shown) to render 2.60 ton h−1 of product. Microbial lipids (SCO104) are recovered from the hexane oil solution (OS-001) in the thin film evaporator TFE-101 to render of 2.01 ton h−1 of product.

3.3. Economic analysis The PFD summarized in Fig. 2 possess a main inlet stream of molasses that is bioprocess carbon source acquired at US$ 100.00/ton and two main outlets microbial lipid and defatted biomass with a selling price fixed at respectively US$ 1,300.00/ton of lipid and US$ 500.00/ton of defatted yeast. The microbial lipid value was set at US$ 1330/ton as an intermediary value of the vegetable oil practiced in US during 2003–2013 periods (Koutinas et al., 2014). The defatted biomass value was set at US$ 500/ton to emulate soybean meal quotation in the same period. The list of equipment specifications and unitary cost of acquisition of PFD (Fig. 2) is presented in Table 2. Bioreactor train is by far the most expensive item of investment cost accounting followed by separation equipment being therefore lipid productivity and lipid concentration two of the most important key process variable to be optimized. Based on mass balance and the cost of acquisition of equipment for the process described in Fig. 2 the Direct Fixed Capital Cost is presented in Table 3. It is worth to mention that in the present alternative it was choose to use full-sterilized stainless steel 316L bioreactors conventionally used the bio industry, and as expected the DFCC climb to a value very high of US$ 88.691.000. Operating cost for the proposed flow diagram was therefore calculated taking into account the annual production of microbial oil and defatted biomass. Table 4 shows the operating costs of microbial lipid production of the proposed industrial plant. Based on Tables 3 and 4 a profitability economic analysis of the proposed industrial plant for the 316L stainless steel bioreactor option was carried out and summarized in Table 5. The analysis indicated that the lipid plus defatted yeast production based in the present data was a marginally economic feasible operation. The IRR attained 7.62% per year and the NPV value (at 7% IRR per year) was about US$ 8,323,000 (before taxes) for molasses price at US$ 100.00/ton and the selling prices of microbial lipid and defatted biomass respectively at US$ 1,300.00/ton of lipid and US$ 500.00/ton of defatted yeast. Ratledge and Cohen (2008) reported an expected price of microbial oil to be US$ 3000/ton in order to make the process economic attractive, not considering the cost of the carbon source. Therefore, the selling price would expect to be higher than this value if it would consider the carbon source cost. This is in agreement with Koutinas et al. (2014) that reported a minimum lipid oil price of about US$ 4000/ton at a carbon source cost of US$ 100/ton, (the same value used in the present study). Even so, this value would not compete with the vegetable oil price that varied between US$ 1000 to US$ 2000/ton during 2003–2013 periods (Koutinas et al., 2014). In present analysis the Stainless steel 316L stirred tank bioreactor (R-102 and R-103, Table 2) is by far the most expensive item of capital investment in the production plant representing about 90% of equipment purchase cost hence impacting total investment cost (Table 3) and making a very tight economic operation. Therefore, it is of the most importance to have alternative bioreactor equipment in order to alleviate the investment cost for SCO production. Epoxy-lined carbon steel bioreactor is a commonly used equipment in ethanol production in Brazil (JP Vieira, personal communication). Estimative of capital, operating costs, as well the main parameters of the process economic evaluation for this alternative configuration was therefore carried out (Table 6). The economic analyses were made on the same proposed PFD (Fig. 2).

Table 6 Economic evaluation considering epoxy lined carbon steel bioreactor. Parameters

Value

Unit

Annual operating time Total Capital Investment Direct Fixed Capital Working Capital Startup Cost Capital Investment Charged to This Project Operating Cost Raw Materials Labor-Dependent Facility-Dependent Utilities Main Revenue Other Revenues Total Revenues Cost Basis Annual Rate Unit Production Cost Unit Production Revenue Gross Margin Return On Investment Payback Time IRR (Before Taxes) NPV (at 7.0% Interest)

8.300 47,022,000 42,867,000 2,012,000 2,143,000 47,022,000 25,972,000 13,314,000 2,122,000 2,776,000 7,759,000 22,572,000 10,804,339 33,376,000 16,719.84 1,553.36 1,996.20 22.18 18.63 5.37 24.61 67,797,000

h $ $ $ $ $ $/yr $ $ $ $ $/yr $/yr $/yr ton MP/yr $/ton MP $/ton MP % % years % $

*MP: Main product.

The use of this equipment would bring the investment cost from about US$ 88,691,000 (Table 5) down to US$ 42,867,000 (Table 6). Consequently, in this alternative the internal rate of return (IRR) would climb up to 24.61% leading the NPV (at 7% interest rate) of US$ 67,797,000/year (before taxes). Therefore, the proposed modification brings the process potentially attractive. The increase of lipid productivity (mass of lipid/volume of bioreactor × hour) and the use of low-cost carbon source are obvious choices to decrease SCO cost production. Coproducts such as defatted biomass selling at US$ 500/ton considered in the present study would help to benefit industrial plant cash flow. Another alternative preclude in literature is the use of microbes (naturally or molecular biologically constructed) able to store high-values fatty acids, such as the ␻-3 fatty acids, to help increase industrial plant cash flow (Ageitos et al., 2011). However, we think that the use of low-cost bioreactor as proposed in the present study appears to be a simple alternative with interesting potential deserving attention to be explored in the future. 4. Conclusions The growth and lipid accumulation phases of R. glutinis on sugarcane molasses as raw material. was studied in bioreactor and best attained typically total biomass concentration 62.25 g l−1 , lipid productivity 0.42 g l−1 h−1 , sugar conversion yield to lipids 0.21 g of lipid g−1 of total reducing sugar. An industrial bioprocess conceptual design including mass balance, list of equipment, cost and specifications was established. Economic performed analysis indicated the use of low cost bioreactor built in carbon steel coated with epoxy resin to reduce CAPEX for a substantial increase in NPV to make project be economically attractive. Acknowledgements The present project was partially financed by FAPESP. The authors want to thanks to the Brazilian Bioethanol Science and Technology Laboratory (CTBE) of National Center of Energy and Material (CNPEM) for the use of its installation and the technical assistance.

J.P.F. Vieira et al. / Industrial Crops and Products 89 (2016) 478–485

Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.indcrop.2016.05. 046. References Ageitos, J.M., Vallejo, J.A., Veiga-Crespo, P., Villa, T.G., 2011. Oily yeasts as oleaginous cell factories. Appl. Microbiol. Biotechnol. 90, 1219–1227. Anschau, A., Xavier, M.C., Hernalsteens, S., Franco, T.T., 2014. Effect of feeding strategies on lipid production by Lipomyces starkeyi. Bioresour. Technol. 157, 214–222. Bellou, S., Triantaphyllidou, I.-E., Aggeli, D., Elazzazy, A.M., Baeshen, M.N., Aggelis, G., 2016. Microbial oils as food additivies: recente approach for improvig microbial oil production and its polyunsaturated fatty acid content. Curr. Opin. Biotechnol. 37, 24–35. Bligh, E.G., Dyer, W.J., 1959. A rapid method of total lipid extraction and purification. Can. J. Biochem. Physiol. 37, 911–917. Bueno Netto, C., Craveiro, A.M., Pradella, J.G.C., Oliveira, M.S., Maiorano, A.E., Pinto, A.G., Matsubara, R.M., 2000. Processo para produzir de polihidroxialcanoatos a partir de ac¸úcares extraídos de cana-de-ac¸úcar. BR Patent PI 9103116-8. CONAB, 2015. Acompanhamento da Safra Brasileira 2015 Vol. 2 – Safra 2015/16, available at www.conab.gov.br (accessed 20.05.16.). Cheirsilp, B., Kitcha, S., 2015. Solid state fermentation by cellulolytic oleaginous fungi for direct conversion of lignocellulosic biomass into lipids: fed-batch and repeated-batch fermentations. Ind. Crops Prod. 66, 73–80. Dias, M.O., Cunha, M.P., Jesus, C.D., Rocha, G.J., Pradella, J.G.C., Rossell, C.E., Filho, R.M., Bonomi, A., 2011. Second generation ethanol in Brazil: can it compete with electricity production? Bioresour. Technol. 102, 8964–8971. Diniz, S.C., Taciro, M.K., Gomez, J.G.C., Pradella, J.G.C., 2004. High-cell-density cultivation of Pseudomonas putida IPT-046 and medium-chain-length polyhydroxyalkanoate production from sugar cane carbohydrates. Appl. Biochem. Biotechnol. 119, 51–69. Fales, F.W., 1971. Evaluation of a spectrophotometric method for determination of total fecal lipid. Clin. Chem. 17, 1103–1108. Furlan, F.F., Filho, R.T., Pinto, F.H.P.B., Costa, C.B.B., Cruz, A.J.G., Giordano, R.L.C., Giordano, R.C., 2013. Bioelectricity versus bioethanol from sugarcane bagasse: is it worth being flexible? Biotechnol. Biofuels 6, 1–12. Granger, L.M., Perlot, P., Goma, G., Pareilleux, A., 1993. Efficiency of fatty acid synthesis by oleaginous yeasts prediction of yield and fatty acid cell content from consumed C/N ratio by a simple method. Biotechnol. Bioenergy 42, 1151–1156. Haas, M.J., McAloon, A.J., Yee, W.C., Foglia, T.A., 2006. A process model to estimate biodiesel production costs. Bioresour. Technol. 97, 671–678. Hassan, M., Blanc, P.J., Granger, L.M., Pareilleux, A., Goma, G., 1993. Lipid production by an unsaturated fatty acid auxotroph of the oleaginous yeast Apiotrichum curvatum grown in single stage continuous culture. Appl. Microbiol. Biotechnol. 40, 483–488. Hugot M.,1969. Manual da Engenharia Ac¸ucareira, Ed. Mestre Jou: São Paulo - SP, Brasil. Koutinas, A.A., Chatzifragkou, A., Kopsahelis, N., Papanikolaou, S., Kookos, I.K., 2014. Design and techno-economic evaluation of microbial oil production as a renewable resource for biodiesel and oleochemical production. Fuel 116, 566–577.

485

Leung, D.Y.C., Wu, X., Leung, M.K.H., 2010. A review on biodiesel production using catalyzed transesterification. Appl. Energy 87, 1083–1095. Lewis, T., Nichols, P.D., McMeekin, T.A., 2000. Evaluation of extraction methods for recovery of fatty acids from lipid producing microheterotrophs. J. Microbiol. Meth. 43, 107–116. Li, Y., Zhao, Z., Bai, F., 2007. High density cultivation of oleaginous yeast Rhodosporidium toruloides Y4 in fed-batch culture. Enzyme Microb. Technol. 41, 312–317. Liu, Y., Wang, Y., Liu, H., Zhang, J., 2015. Enhanced lipid production with undetoxified corncob hydrolysate by Rhodotorula glutinis using a high cell density culture strategy. Bioresour. Technol. 180, 32–39. Macrelli, S., Mogensen, J., Zacchi, G., 2012. Techno-economic evaluation of 2nd generation bioethanol production from sugar cane bagasse and leaves integrated with the sugar-based ethanol process. Biotechnol. Biofuels 5, 1–16. Meade, G.P., Chen, J.C.P., 1977. Cane Sugar Handbook, 10 ed. Wiley, New York. Munch, G., Sestric, R., Sparling, R., Levin, D.B., Cicek, N., 2015. Lipid production in the under-characterized oleaginous yeasts, Rhodosporidium babjevae and Rhodosporidium diobovatum, from biodiesel-derived waste glycerol. Bioresour. Technol. 185, 49–55. Pan, J.G., Kwak, N.Y., Rhee, J.S., 1986. High density cell culture of Rhodotorula glutinis using oxygen-enriched air. Biotechnol. Lett. 8, 715–718. Papanikolaou, S., Chatzifragkou, A., Fakas, S., Galiotou-Panayotou, M., Komaitis, M., Nicaud, J.M., Aggelis, G., 2009. Biosynthesis of lipids and organic acids by Yarrowia lipolytica strains cultivated on glucose. Eur. J. Lipid Sci. Technol. 111, 1221–1232. Pradella, J.G.C., Taciro, M.K., Mateus, A.M.P., 2010. High-cell-density poly (3-hydroxybutyrate) production from sucrose using Burkholderia sacchari culture in airlift bioreactor. Bioresour. Technol. 21, 8355–8360. Ratledge, C., Cohen, Z., 2008. Microbial and algal oils: do they have a future for biodiesel or as commodity oils? Lipid Technol. 20, 155–160. Ratledge, C., 1991. Microorganisms for lipids. Acta Biotechnol. 11, 429–438. Ratledge, C., 2013. Microbial oil: an introductory overview of current status and future prospects. OCL 20, D602. Ratledge, C., 2014. The role of malic enzyme as the provider of NADPH in oleaginous microorganisms: a reappraisal and unsolved problems. Biotechnol. Lett. 36, 1556–1568. Rossell, C.E.V., Mantelatto, P.E., Nonato, R.N., 2001. Integrated production of biodegradable plastic, sugar and ethanol. Appl. Microbiol. Biotechnol. 57, 1–5. Srienc, F., Arnold, B., Bailey, J.E., 1984. Characterization of intracellular accumulation of poly-beta-hydroxybutyrate (PHB) in individual cells of Alcaligenes eutrophus H16 by flow cytometry. Biotechnol. Bioenergy 26, 982–987. Tabatabaei, M., Tohidfar, M., Jouzani, G.S., Safarnejad, M., Pazouki, M., 2011. Biodiesel production from genetically engineered microalgae: future of bioenergy in Iran. Renew. Sust. Energy Rev. 15, 1918–1927. Vieira, J.P., Ienczak, J.L., Rossell, C.E., Pradella, J.G.C., Franco, T.T., 2014. Microbial lipid production: screening with yeasts grown on Brazilian molasses. Biotechnol. Lett. 36, 2433–2442. Zhao, X., Peng, F., Du, W., Liu, C., Liu, D., 2012. Effects of some inhibitors on the growth and lipid accumulation of oleaginous yeast Rhodosporidium toruloides and preparation of biodiesel by enzymatic transesterification of the lipid. Bioprocess Biosyst. Eng., 1–12. Zhu, L.Y., Zong, M.H., Wu, H., 2008. Efficient lipid production with Trichosporon fermentans and its use for biodiesel preparation. Bioresour. Technol. 99, 7881–7885.