Vineyard Pruning Extracts as Natural Antioxidants for Biodiesel Stability: Experimental Tests and Preliminary Life Cycle Assessment

The control of the oxidative stability of biodiesel and blends of biodiesel with diesel is one of the major concerns of the biofuel industry. The oxidative degradation of biodiesel can be accelerated by several factors, and this is most critical in the so-called second generation biodiesel, which is produced from low-cost raw materials with lower environmental impacts. The addition of antioxidants is imperative to ensure the oxidative stability of biodiesel, and these are considered products of high commercial value. The antioxidants currently available on the market are from synthetic origin, so the existence/availability of alternative antioxidants of natural origin (less dependent on fossil sources) at a competitive price presents itself as a strong business opportunity. This work describes and characterizes a sustainable alternative to synthetic antioxidants used in the biodiesel market developed from extracts of vineyard pruning waste (VPW), which are naturally rich in phenolic compounds with antioxidant properties. A hydrothermal extraction process was applied as a more efficient and sustainable technology than the conventional one with the potential of the extracts as antioxidant additives in biodiesel evaluated in Rancitech equipment. The VPW extract showed comparable antioxidant activity as the commercial antioxidant butylated hydroxytoluene (BHT) typically used in biodiesel. The stability of the biodiesel is dependent from the amount of the extract added. Further, for the first time, the assessment of the environmental impacts of using natural extracts to control the oxidative stability of biodiesel in the production process is also discussed as a key factor of the process environmental sustainability.


List of Figures:
: a) VPW extract dissolved in benzyl alcohol (BA); b) VPW (1500 ppm) dissolved in BA and added to the biodiesel sample   :   Table S1: Life Cycle Inventory assumptions made for the design of the sub-processes of Vineyard Pruning Waste (VPW) extract production. Table S2: Inventory data of the VPW based antioxidant production process. All values are referred to the functional unit (1L biodiesel). Table S3: Tier 1 Emission Factors for Road transport -Heavy Duty Vehicles [1]. Table S4: Tier 1 emission factors for non-road machinery [2]. Table S5: Inventory data of the landfill of hydrochar. All values are referred to the functional unit (1L biodiesel). Table S6: Life Cycle Inventory assumptions made for synthetic BHT production process. Table S7: Inventory data of the BHT production process. All values are referred to the functional unit (1L biodiesel). Table S8: Chemical composition (Fatty acid methyl ester composition) of the sample biodiesel (antioxidant free) by gas-chromatography (GC) following the EN14103 Table S9: Some properties of the sample biodiesel (antioxidant free) Table S10: Flash point determination according Procedure C applicable to fatty acid methyl esters (FAME) as specified in EN 14214 [11] or ASTM D6751. Table S11: Processes contribution to the area of protection Human Health at EndPoint(H) level (weighting) of VPW extract dissolved in benzylic alcohol (cut-off 2%). Values expressed as single score (Pt) and percentage (%). Table S12: Inventory data of the Sensitivity Analysis-1. All values are referred to the functional unit (1L biodiesel). Table S13: Inventory data of the Sensitivity Analysis-2. All values are referred to the functional unit (1L biodiesel).   The analysis of performance and accuracy between the IP (induction period) results obtained in the new RANCITECH apparatus was done by comparison of the results obtained for the same sample batch of biodiesel. The obtained data is summarized in Figure S3 which presents a comparison between the results (average of replicas) the two blocks of the Rancimat (Rancimat 1 & Racimat 2) and the RANCITECH results (average of replicas) the four blocks (A1.1, A1.2, B2.1, B2.2). Comparing the results obtained in both studies it was possible to conclude that the RANCITECH can produce reliable values for the OS with a higher reproducibility and lower uncertainty. It was found a higher reproducibility and repeatability of the RANCITECH apparatus which was attributed to the improved air flow control and the better temperature control and homogeneity of the thermostable block.

List of Tables
Based on results obtained in this study, and in the uncertainty propagation analysis (considering the the partial uncertainties of air flow, time interval, temperature oscillation, and sampling) the following uncertainty equation (equation [S1]) was derived for the 1 Table S1: Life Cycle Inventory assumptions made for the design of the sub-processes of Vineyard Pruning Waste (VPW) extract production.

1.VPW Harvest
The harvest is performed by mechanical box-pruning of 50 ha of "Quinta dos Carvalhais" vineyard (Mangualde, Portugal), assuming the use of a dedicated machinery with average fuel consumption 17.0 L diesel/ha [3]. The harvested vineyard pruning are then collected and transported to the vineyards borders by a support tractor, whose fuel consumption is estimated in 3L diesel/ha [4]. For this calculations, a vineyard pruning density of 720 kg/m 3 with a 50 % w/w moisture content (MC) is considered [5]. Table S2 reports the emission factor assumed for non-road agricultural machineries used for the calculation of emission to air [2]. A consumption of 834.1 kg/y of agricultural diesel is expected. The harvest of vineyard pruning avoided the production of the equivalent amount of wood waste, which avoid consequently the impacts associated to its combustion in a biomass furnace (Table S2) (Waste wood, untreated, {ROW}| heat production, untreated waste wood, at furnace 1000-5000 kW | Consequence, U).

VPW Transport to "Quinta dos Carvalhais"
The collected VPW are then transported from the vineyard border to "Quinta dos Carvalhais" by a truck with 7 m 3 capacity. Assuming an average distance of 10 km and an estimated fuel consumption of 25 L/100km [6][6], 87.2 kg/y fuel are required. Table S3 reports the emission factor used for the calculation of the emission to air for road transportheavy duty vehicles [7][7].

VPW Grinding.
VPW are ground by a wood branches pulverizer, 1000 kg/h, 39.5 kW (53 Hp) capacity able to grind VPW up to 4 mm. The estimate time for pulverizing is 64.5 h/y. As first approach, it is assumed that this particle size allows the same yields of antioxidant activity of 229 mgGAE/g dw considered for the present work [8].

Dried VPW Transport to SWE Plant (Porto, Portugal).
The ground VPW are transported to SWE plant by Euro VI trucks of 32 m 3 capacity.
Assuming an average distance of 150 km, and an estimated fuel consumption of 30 L/100km [6], 674.8 kg/y fuel are required. The emission factor assumed for road transportheavy duty vehicles used for the calculation of emission to air are reported in Table S3 [7]. S10 Table S1: cont.

VPW Feed (Screw-feed).
For the loading of the ground VPW a standard pitch, single flight a 4 m long, 1000 kg/h capacity, 112 rpm, 0.12 kW screw conveyor was hypothesized. The design of the equipment was performed according to [9], taking sawdust as reference material with an overload and drive efficiency factors of 3 and 0.88, respectively. The water required for the SWE process was 586 m 3 /y (ground VPW:water ratio = 10), and it is assumed to be loaded using the grid pressure, so no additional energy requirements are considered for water feed;

6.Subcritical Water Extraction
The impact associated to SWE process were calculated based on the enthalpy change required to rise the ground VPW and the water from NTP conditions (20 , 1.01325 bar) ℃ to 280 and 80 bar (SWE operational conditions). The calculation was performed using ℃ average specific heat for water ( , kJ/kg. ) calculated between 20 and 280 ℃ ℃ ℃ according to Eq.S2, which gives 4.73 kJ/kg. C.
In lack of specific experimental data available, the cp for grind VPW was considered constant and equal to 0.9 kJ/kg. [10]. ℃ The energy required for SWE process was calculated as enthalpy change according to Eq.S3: where m is the total mass (kg) of ground VPW and water, cp/ are the specific/average specific heat (kJ/kg. °C), respectively, T 2 is the SWE operational temperature (280 ) and ℃ T 1 is the initial room temperature (20 ). The calculated heat amount required for SWE ℃ according to [11], is 235.7 MWh/y.

Flash and Separation of hydrochar from liquid VPW extract
To allow heat recover and energy saving, 174 MWh of heat in form of water vapor are recovered hypothesizing the sudden pressure release from SWE reactor from 80 bar to almost atmospheric pressure (1.1 bar). This flash allows to vaporize the hot liquid water present in the SWE reactor, recovering a hot vapor stream and reducing the remaining water vapor that must be evaporated at the end by 46 % w/w. Based on the phenolic composition of the VPW extract, which includes (+)-catechin acid (24% w/w), gallic acid (16%w/w), (-)-epicatechin acid (14% w/w), caffeic acid (9% w/w), chlorogenic acid (7% w/w), others (30% w/w) [8], it is assumed that no valuable antioxidants are drag into the S11 exiting steam stream, since all of them have boiling point higher than 400 °C. Table S1: cont.

Flash and Separation of hydrochar from liquid VPW extract
Thus, the remaining concentrated-SWE extract is separated from the hydrochar by filtering in the presence of a pump of 4 bar discharge pressure, which demands 0.034 MWh/y electric energy. Two different pathways were considered for the hydrochar separated from concentrated liquid extract, namely: (i) hydrochar transport and landfill, assuming 20 km distance from the landfill and the use of a diesel fueled truck for hydrochar transportation (Table S5), and the (ii) introduction into the market as precursor for AC. This second option allows the "avoided production" of an equivalent amount of charcoal, providing environmental credits to the system, since the "system expansion" approach was adopted ( Figure 2a). For the modeling of these two scenarios, they have been used the reference processes included in the Ecoinvent 3.7 database "Wood ash mixture, pure {Europe without Switzerland} | treatment of wood ash mixture, pure, sanitary landfill| Conseq, U") (Table S4) and "Charcoal {GLO} | market for | Conseq, U", respectively (Table S2);

Concentrated VPW extract Evaporation
The remaining condensate SWE extract is evaporated in a thin film evaporator, operated in continuous under vacuum (0.01 bar) in the presence of a chiller used to lower the temperature of the incoming concentrated VPW extract to 33 . The refrigeration of ℃ outcoming stream from SWE reactor and the generation of vacuum requires additional 193.28 MWh/y electric energy, with the chiller representing more than 99% of this demand. The vacuum has the function of lowering the boiling point of water from 100 to 9.7 ℃ ℃ and to achieve then thermodynamic conditions which allow the recovery of the 174 MWh/y of heat produced during the water vapor flash (sub-process 7) for final evaporation. The design of the energy integration in the evaporation process allows to avoid the production of an equivalent amount of heat, based on the process "Heat, central or small scale, natural gas {Europe without Switzerland}, market for |Conseq, U" of Ecoinvent 3.7 database, since the heat required for the evaporation of the remaining concentrated SWE extract accounts is almost equivalent to the heat recovered from the flash. The condensate is then collected and sent to treatment according to the process included in Ecoinvent 3.7: "Wastewater, average {Europe without Switzerland} | treatment of wastewater, average, capacity 1E9l/year | Conseq, U");

Dried VPW extract Dissolution in benzyl alcohol
According to the laboratorial conditions tested, which were properly scaled-up, the use of 697.8 t/y of benzyl alcohol (ratio 0.02 kg extract :L alcohol benzyl ), was considered to dissolve the precipitated antioxidant extract and to obtain the final usable biodiesel additive.
. S12 Table S2: Inventory data of the VPW based antioxidant production process. All values are referred to the functional unit (1L biodiesel).        Table S6: Life Cycle Inventory assumptions made for the design of the synthetic BHT production process.

BHT synthesis
The LCI of the production process of BHT assumes that oxidative stability of biodiesel is achieved with 600 ppm of BHT addition and that 5,368 kg/y of BHT are required to equal the correspondent amount of VPW extract produced (Table S7). As first approximation, in absence of specific data in Ecoinvent 3.7 database, the estimate of the environmental impacts associated with the BHT production process was based on the data retrieved by the US Patent 2,428,745 [12] integrated with the experimental results obtained by Yadav et al [13], where the synthetic antioxidant BHT is obtained by alkylation of p-cresol with isobutylene in the presence of sulfated zirconia (S-ZrO 2 ) as superacid catalyst. The isobutylene amount required was assumed, as first approximation, 100% excess w/w of the amount calculated through the mass balance of the experimental results obtained by Yudav et al. [13]. Alkylation of p-cresol with isobutylene is a combination of series and parallel steps proceeding via an intermediate. The reaction products are the monoalkylated and dialkylated p-cresols and the oligomerized isobutylene leading to dimer, trimer, and tetramer: 9.6x10 3 g.mol.cm -3 p-cresol gives 1.4x10 3 g.mol.cm -3 2-tert-butyl-p-cresol; 7.5x10 3 g.mol.cm -3 ; 0.6 10 3 g.mol.cm -3 diisobutylene, 1.4 × g.mol/cm -3 triisobutylene, 0.6 g.mol.cm -3 tetraisobutylene in the presence of 3% (w/w) S-ZrO 2 at 348 K [13].

Catalyst production
According to the methodology suggested by Yadav et al. [13] the catalyst used for the pcresol synthesis is sulfated zirconia (S-ZrO 2 or Zr(SO 4 ) 2 ) . In the absence of a specific process in Ecoinvent 3.7 database, the assessment of the indirect environmental impacts associated to the production of this catalyst, it was created an ad hoc process, based on the following assumption:  zirconium dioxide (ZrO 2 ) is used as precursor for sulfated zirconia synthesis, since its industrial production process can be conservatory considered equal to that of zirconium oxychloride (ZrOCl 2 •8H 2 O) [14], which is referred to be raw material for catalyst production. The indirect impacts associated to ZrO 2 production processes were calculated based on the Ecoinvent 3.7 database: "Zircon, 50%, zirconium, {GLO} |, market for| Conseq, U". Proper quantity adjustments were made to guarantee the correct amount of ZrO 2 ;  the ammonia induced hydroxide precipitation from zirconium oxychloride was considered to be as in Eq. S6 [15]:  the impregnation of zirconium hydroxide was performed by immersing the dried zirconium hydroxide in 0.5M H 2 SO 4 in the ratio of 1:30 w/w;  the energy required for S groups calcination was calculated by enthalpy change from 25 to 650 , considering the constant pressure heat capacity of sulfuric ℃ ℃ acid of 1.380 J.(kg.K) -1 [16], of water of 4.190 J.(kg.K) -1 [16] and of zirconium of 270 J.(mol.K) -1 [16], which gives 2.293 kWh/y.

Waste
The amount produced of diisobutylene, triisobutylene, tetraisobutylene and 2-tert-butyl-pcresol are considered as hazardous liquid waste exiting in the system.

Energy
The energy required for reaching the reaction temperature was calculate as enthalpy change from 298 K to 348 K, considering the constant pressure heat capacity of p-cresol of 163 J/mol.K [16] and of isobutylene of 88.09 J/mol.K [16] applying the Eq. S7.
It accounts for 350 kWh/y. The final raw BHT obtained is washed with a NaOH 5% w/v solution, whose yield is considered to be 46% [12].