Physical properties of defatted and extruded black soldier fly (Hermetia illucens) larvae-based aqua-feed using a twin-screw extruder

A twin-screw hot extrusion processing study was performed using defatted black soldier fly larvae (DBSFL) meal as a complete substitute for Danish fishmeal (DFM) in the production of aqua-feed. The defatting process of full-fat black soldier fly larvae (BSFL) reduced its crude fat content by 60%. This was mixed together with other ingredients (soybean meal, cassava flour, vitamin and mineral mix, vitamin C, and anti-mold) to formulate a fish diet containing 40% crude protein. The moisture content of the formulated mixture was 300 g/kg prior to extrusion. The extruder feeder motor speed, main motor speed and cutter motor speed were set at 210, 300 and 700 rpm respectively. Results show a significant increase (P < 0.05) in the crude protein (CP) content of DBSFL meal when compared with full-fat BSFL meal with values of 51.83 ± 0.03 and 47.44 ± 0.2 respectively. However, CP values of the DBSFL meal were significantly (P < 0.05) lower than those obtained for DFM with values of 51.83 ± 0.03 and 72.13 ± 0.18 respectively. The physical properties of extruded floating pellets showed significant differences (P < 0.05) between DFM-based extrudates and DBSFL-based extrudates in terms of expansion ratio (ER), sinking velocity (SV), Unit density (UD), bulk density (BD), and pellet durability. However, no significant differences (P > 0.05) were observed in terms of floatability, moisture content and water absorption index (WAI). It was concluded that, the extrusion of DBSFL-based pellets compared favourably with DFM-based extruded pellets thus, the processing of full-fat BSFL into DBSFL meal makes the latter suitable for extrusion purposes and appropriate in the production of good quality floating feed pellets desired by fish farmers and suitable for feeding fish.


Introduction
In recent years, fish farmers have embraced the use of extruded floating fish feed as a more suitable type of feed for farmed fish other than sinking feed.This is only a mirror reflection of the receptiveness of farmed fish to floating feed technology.The benefits of this technology are numerous in aqua-feed production and include, the possibility of wide application, high productivity, energy efficiency, and high quality of the resulting product [1].The extrusion process also enhances the digestibility of starch through starch gelatinization, melting, fragmentation, denaturation of anti-nutritional factors, ingredient sterilization, and structural changes in protein and amino acid [2,3].These however, did not come without the enduring challenges of getting a suitable and sustainable alternative to fishmeal. 1 3 The increasing rate of global demand for fishmeal, coupled with the depletion of wild fish stocks has elevated the cost of fishmeal, and consequently, diet prices [4].Therefore, the utilization of alternative protein sources can play a crucial role in achieving a sustainable and profitable aquaculture industry.
A familiar alternative to fishmeal is the use of insects.Evidence has suggested that insect meal has higher nutritive values, greater feed conversion efficiency, smaller environmental footprint, and relatively lower price than Fishmeal and other alternative protein sources [5][6][7].Among the most commonly used insects are the culture and utilization of the Black Soldier Fly Larvae (BSFL) in fish feed.This is a technique that offers the additional benefit of organic waste utilization and provides a potential solution for safe manure management [8].BSFL is proven to be nutritionally very rich.Its larvae contain about 40-45% protein, 30-35% fat, 11-15% ash, 4.8-5.1% calcium, and 0.6% phosphorous, as well as a range of excellent amino acids and minerals [9].Several authors have successfully fed full-fat BSFL to different fish species like rainbow trout, catfish and tilapia and domestic animals like swine and poultry [10][11][12][13].
The desire to produce extruded and floating insect-based fish diets has been constrained by the use of full-fat insects which are most common for use in large quantities especially BSF larvae.Defatting of these full-fat insects becomes very cumbersome thus fish farmers are left with the only option of directly using such insects as live feed in their full-fat state to feed farmed fish or as ingredients in sinking feed blends without extrusion.It is also worthy of note that most extruders used by aqua-feed producers show very little affinity to the high lipid content of blends.These necessitated this research.To our knowledge, there is a dearth of information on the defatting process of BSFL, their use in the production of extruded and floating insect-based fish diet and a comprehensive study of their extrudate's physical properties.The aim of this study was to investigate the physical properties of defatted and extruded black soldier fly (Hermetia illucens) larvae-based aqua-feed using a twin-screw extruder.

Harvest of BSFL and purchase of DFM
BSF larvae were harvested from the BSFL production unit of MagProtein factory a newly established BSFL production factory situated at Epe, Logos State Nigeria.Harvested larvae were stored in a clean ice-chest cooler and transported to the fish Nutrition laboratory section of the Nigerian Institute for Oceanography and Marine Research, Lagos for processing, while Danish fishmeal (DFM) was purchased from SOLASE Group of Companies.SOLACE is a major wholesale distributor of imported fishmeal in Lagos, Nigeria.

Processing of full-fat BSFL samples into defatted BSFL meal
Harvested BSF larvae (Fig. 1) were rinsed thoroughly with clean water, blanched and then oven-dried at 32 °C for three days (72 h) using a laboratory dryer (Germany FP: 240 Model).Dried BSFL were subjected to a de-fattening process using an automatic oil extractor machine (Model: ZF-868 China).The oil extraction method involved feeding dry BSFL samples into a barrel through the hopper feeder situated at the top of the machine.In the barrel, the dry insect enters the automatic extruder-like barrel where oil is squeezed out leaving a dry extrudate which is collected in a clean tray at the end of the barrel.The extrudates; coarse Defatted Black Soldier Fly Larvae (DBSFL) meal were then milled using the laboratory milling machine (Model: HK-860 Germany) and passed through a 0.05 mm aperture sieve.

Chemical analysis
The proximate composition of DFM and BSFL were determined in triplicates according to [14] at the Central Laboratory of the Nigerian Institute for Oceanography and Marine Research, Lagos, Nigeria.Crude protein was determined by AOAC method 984.13, crude fat by AOAC method 920.39 and crude ash by AOAC method 942.05.Crude fibre was determined according to ISO 6865.2000(ISO, 17025/2005), while carbohydrate content was determined by difference in Table 1.
Mineral composition after wet digestion was determined using a flame Atomic Absorption Spectrophotometer AAS (Spectraa AA 220, Varian Inc., Australia).Calcium (Ca), magnesium (Mg), zinc (Zn), iron (Fe) and copper (Cu) were determined by flame AAS.Sodium and potassium were measured by flame photometer and phosphorus by UV-Vis spectrophotometer.Lanthanum was used to compensate for ionization interferences in the analysis of Ca and Mg (Table 2).

Formulation and incorporation of DFM and DBSFL meal in blends
Based on the proximate composition of ingredients, the required quantities to formulate two iso-nitrogenous (40% crude protein) diets were calculated using Microsoft Excel® function spread-sheet in Windows 2010 (Microsoft Corporation, Redmond, WA, USA) and weighed using a top-loading balance (Mettler Toledo PB8001 London).
The amount of water needed to achieve the desired moisture content was calculated, weighed, and added to the other ingredients (soybean meal, cassava flour, vitamin and mineral pre-mix, vitamin C and anti-mold) in a circular bowl (20L capacity).These were mixed manually by hand for 2 min, then transferred to an automatic mini horizontal laboratory mixer (model UDHM-25 KG; Kaifeng, China) and mixed further for 5 min at 26 rpm.The composition of the blend is presented in Table 1.

Extrusion of DFM-based pellets and DBSFL-based pellets
The diet mixture was extruded using a laboratory-scale automatic double screw extruder machine (Model: DSE32 Lab Extruder) with the first, second and third barrel heater temperatures set at 115 °C, 120 °C and 130 °C respectively.The extruder motor, feeder motor and cutter motor were also set at 210 rpm, 300 rpm and 700 rpm respectively.The temperature inside the barrel and speed of the twin-screw was controlled by a digital computerized monitor which is connected to the extruder.The extruder was connected to a three-phase 12.5 HP motor.After the extrusion process, the extrudates (floating DFM-and DBSFL-based pellets) were then coated with vegetable oil using an oil-coating machine.
The oil-coated extruded feeds were 2 mm in diameter suitable for farmed fish fingerlings.Extruded diets were allowed to cool and dried under ambient conditions for 6 h before packaging.Packaging was done using a polyethylene bag.

Measurement of extrudate physical properties
The physical properties of extruded DFM-and DBSFL-based pellets were carried out.The moisture content (% db), unit density (kg mG 3 ), bulk density (kg mG 3 ), expansion ratio (%), floatability, sinking velocity (m sG 1 ), water absorption index and water solubility index were measured.

Moisture content
2 g each of extruded DFM-and DBSFL-based diets were dried in an oven at 104 °C for 24 h.Clean crucibles were also dried in an oven at 100 °C for about 30 min, cooled in a desiccator and accurately weighed (W 1 ).Two grams each of wellhomogenized sample were weighed into the crucible (W 2 ).The crucible containing the dry samples were then dried in an oven at 104 ± 2 °C to a constant weight (W 3 ) for about 24 h.

Unit density
Each extrudate sample was grinded into powdery form.The volumetric flask was then placed on an analytical balance and the weight zeroed.The samples were then transferred using a funnel into the volumetric flask (100 or 50 ml capacity) to about 50 ml of the mark and the volume (V) was recorded.The sample was thereafter weighed to get a new mass.Calculation

Bulk density
This was carried out using the procedure of [15].100 g quantity of each extrudate sample was transferred into an already weighed measuring cylinder (W 1 ).For the packed bulk density determination, the extrudate sample was gently tapped to eliminate spaces between the extrudate and the level was noted to be the volume of the sample and then weighed (W 2 ).No tapping was made in the case of loosed bulk density and the level was also noted to be the volume of the sample and then weighed.The study was conducted in triplicate.Calculation g∕ml or g∕dm 3

Expansion ratio
The expansion ratio was determined as described by [16].The diameter of the extrudates were measured with a Vernier caliper (Digimatic Series No. 293, Mitutoyo Co., Tokyo, Japan) and then divided by the diameter of the die nozzle (2.90 mm).

Floatability
Ten randomly selected pellets from each sample extrudate were placed in a 250 ml beakers containing 200 ml of distilled water at room temperature.Pellets were allowed to float for 20 min.The average number of pellets that were found floating after the allotted time of 20 min were recorded.Floatability of pellets was calculated as the number of floating pellets after 20 min divided by the total number of pellets initially introduced in the water multiplied by 100 [16].

Sinking velocity
The sinking velocity was measured using the method developed by [17] and was determined by monitoring the time taken for an extrudate to reach the bottom of a 2000 ml measuring cylinder filled with distilled water.Distance travelled for the time taken gave the sinking velocity (m s −1 ).

Water absorption index and water solubility index
10 g of each sample extrudate were ground into fine powders using a laboratory milling machine (Model: HK-860).The grounded extrudate (2.5 g) was suspended in distilled water (30 mL) in a tarred 60 mL centrifuge tube.The suspension was stirred intermittently and centrifuged at 3000×g for 10 min.The supernatant was decanted into a tarred aluminum cup and dried at 135 °C for 2 h.30.The weight of the gel remaining in the centrifuge tube was measured.The water absorption index and water solubility index (WSI, %) were calculated using the formula below.
where Wg is the weight of gel (g), and Wds is the weight of dry sample (g).
where Wss is the weight of dry solids of supernatant (g), and Wds is the weight of dry sample (g).

Statistical analysis
All data were measured in triplicates.Microsoft Excel version 2010 and SPSS version.20.0 software was used to analyze the collected data.A Two-Sample T-test was performed to identify the significant differences between treatments.

Results and discussion
The results of the proximate composition of DFM, full-fat BSFL and DBSFL meals are presented in; Table 2. Mineral composition and extruder calibration in Tables 3 and 4 respectively.Extrudate physical properties of DFM-based extruded pellets and DBSFL-based extruded pellets are presented in Table 5.
After the de-fattening process using the automatic oil extractor machine (Model: ZF-868) fat contents in DBSFL meal were reduced by 60% of the initial value.

Mineral composition of DFM and DBSFL
The mean mineral content in mg/100 g are presented in Table 2. Values are means plus or minus standard deviation.

Extruder calibration and extrudate physical property analysis
Results of extruder calibration and extrudate physical properties (moisture, unit density, bulk density, expansion ratio, floatability, sinking velocity, water absorption index and water solubility index) are presented in Tables 4 and  5 respectively.The proximate composition of BSFL was considerably lower than those obtained for DFM in terms of crude protein and carbohydrate contents.Proximate composition values of BSFL before defatting and after defatting revealed better crude protein values, especially after defatting.The higher values of crude protein after defatting may be due to the reduced effect of crude fat dilution in a given quantity.The values of crude protein obtained after defatting are much higher than those reported by [18] who worked on black soldier fly full-fat larvae meal as an alternative to fish meal and fish oil in Siberian sturgeon nutrition, and [19][20][21][22].The higher fat and fibre contents of BSFL meal before de-fattening explains why in their crude form, are mostly unsuitable for use in extrusion purposes.
On the control panel of the extruder machine, temperature settings were adjusted to a constant level at the beginning of each treatment extrusion.However, during extrusion, temperatures within the different barrel zones increased slightly due to friction.This slight increase in temperature was adjusted automatically by the automated control panel to the locked-in temperature set.These temperature effects could be due to frictional heating and shear forces in the barrel during extrusion processing.This phenomenon is similar to the report of [23] who worked on melt extrusion processing parameter optimization.
Expansion ratio is one of the most important quality parameters for fish feed extrudates [4].The main driving force for extrudate expansion is moisture, pressure gradient, starch gelatinization and temperature during water evaporation upon exiting the die section of the extruder.This phenomenon causes a liquid-vapor phase change which creates air bubbles in the extrudate structure [24].In this study, it was observed that though the values of expansion ratio obtained for DBSFL-based extruded pellets were significantly lower than those obtained for DFM-based control pellets, they are however similar to those reported by [25].The use of cassava flour coupled with reduced lipid content of DBSFL-based blend at high extrusion temperature increased starch gelatinization of blend within the barrel, producing extrudates with a high expansion ratio and consequently reducing bulk density.These findings are similar to the report of [26] who worked on the effect of starch gelatinization on the physical properties of extruded wheat and corn-based products.This is also in line with the findings of [27] who worked on the nutritional composition and In Vitro starch digestibility of crackers supplemented with faba bean whole flour, starch concentrate, protein concentrate and protein isolate and affirmed that higher temperature causes more starch gelatinization as well as super-heated steam extrusion which consequently increases expansion ratio.According to [28], who studied the effect of moisture content on expansion ratio and submitted that, decreased moisture content increases the drag force and therefore exerts more pressure at the die resulting in a greater expansion ratio.This report is similar to the findings of this study as the higher moisture quantity of the blend was adequate for the proper extrusion of DBSFL-based floating pellets.
According to [29], sinking velocity gives an indication of how long the pellets would remain floating on water.In this study, the values of the sinking velocity of DBSFL-based extruded pellets within the first 4 h, 8 h and 12 h were significantly lower than those obtained for DFM-based control pellets.Although, during the first 0-10 h, DBSFL-based extruded pellets were still floating showing a strong buoyancy characteristic of pellets which could be due to the benefits of defatting BSFL, higher fiber content and starch gelatinization in this study.This is similar to the findings of [30] who worked on the physico-chemical properties of extruded aqua feed pellets containing black soldier fly (Hermetia illucens) larvae and adult cricket (Acheta domesticus) meals.This report is also in line with the findings of [31] who worked on the optimization of the process parameters for extruded commercial sinking fish feed with mixed plant protein sources and concluded that extrudate expansion influences the sinking characteristics of pellets.
Floatability defines whether an aqua-feed is to be utilized by bottom, mid-level or top feeders that feed on sinking, slow sinking or extruded floating feeds respectively [30].In the present study, starch content in the blend was supplied by cassava flour.The total substitution of DFM with DBSFL meal altered the composition of fat, and fibre contents.According to [32] the amount of fibre in a given feed blend during extrusion affects the moisture binding of the feed and hence modifies the behavior and viscosity of the melt as it flows and exits through the die.Furthermore, the high level of floatability observed in both DFM-based-and DBSFL-based extrudates as moisture content remained constant could also be attributed to the slightly high moisture and starch content of the blends which enhances gelatinization and produces floating pellets with over ninety-eight percent floatability for both DFM-based-and DBSFL-based extruded pellets.This finding is in agreement with the report of [33] who worked on the effect of the addition of high-protein hydrolyzed flour from Oncorhynchus mykiss by-products on the properties of an extruded feed and submitted that, excellent pore formation in extruded pellets is a factor of a proper gelatinization of starch granules during the extrusion process.Furthermore, the excellent percentage floatability of DBSFL-based extruded pellets in this study is a desirable characteristic needed during fish feeding.It is also worthy of note that most of the extrudates floated until they disintegrated after six hours.This finding is similar to the report of [34].
Unit density is also an important quality parameter that determines whether the extrudate will sink or float [35].It plays a crucial role in the floatability of aqua-feeds and quantifies the density of a single extrudate [36].Unit density is also related to the expansion ratio, which in turn is affected by the temperature and moisture content of the feed blend during the extrusion process.This phenomenon is in line with the findings of this study.For many fish species, extruded floating feed is suitable since they tend to feed at the water surface thus, extrudates that sink to the bottom of the tank may not be eaten, and present potential feed loss and, over time, contamination of the water [37].Though the values of the unit density of DFM-based-and DBSFL-based extruded pellets are significantly different, the effects of their resultant interactions were however not substantive enough to cause a significant statistical difference in terms of floatability, water absorption and water solubility index of extruded pellets in this study.
The bulk density of any kind of processed materials plays a vital role in the cost estimation of the product.It is a property that describes the weight of an ingredient per unit volume and thus affects the transportation and unit costs.It is also an important factor to be considered when verifying the storage volume of transport vehicles, vessels and containers [4].Extrudates with higher bulk density require less space for storage and transportation systems which in turn reduces the associated costs.The bulk density obtained in this study for both DFM-based-and DBSFLbased extruded pellets is a factor of barrel temperature and adequate moisture content of the blend.Although values between DFM-based-and DBSFL-based extruded pellets were significantly different, they are however similar to those reported by [38] who worked on the effects of Clostridium autoethanogenum protein inclusion levels and processing parameters on the physical properties of low-starch extruded floating feed.Our findings further reveal a phenomenon that shows an inverse relationship between bulk density on the one hand and expansion ratio and floatability on the other hand.This is in agreement with the findings of [31] who reported that the bulk density of commercial sinking fish feed decreased as the moisture content increased.Nail et al. [39] also reported that, under high temperature and high moisture content conditions, the superheated water in the extruder promotes bubble formation and decreases the viscosity of the melt, resulting in a decrease in the bulk density which agrees with the findings of this study.
The water Absorption index refers to the amount of water absorbed by the extrudate during extrusion, while the water solubility Index determines the number of small molecules melted or dissolved in water during extrusion [40].During extrusion cooking, starch gelatinization, protein denaturation, modification of lipid, Maillard reactions, and complex formation among starch, protein, and lipids lead to significant changes in the chemical and physical characteristics of the extruded products.Jiang et al. [41] stated that, the rate of starch gelatinization and protein denaturation of extrudates determines the value of the water absorption index and water solubility index.In this study, the high moisture content of the melted blend resulted in a higher water absorption index and lower water solubility index of extrudates.This phenomenon may be due to the gelatinization of starch at a higher moisture content of melted blend rather than dextrinization reported by Ma et al. [38].Ma et al. [38] worked on the effects of Clostridium autoethanogenum protein inclusion levels and processing parameters on the physical properties of low-starch extruded floating feed at low moisture blends content.Yadav et al. [42] who reported that, extrusion processing at high barrel temperature and high blend moisture content increases the degree of starch gelatinization which leads to an increase in water absorption index and expansion of the product which is similar to the findings of this study.

Conclusion
In conclusion, the use of DBSFL as a complete substitute for DFM in the production of extruded floating aqua-feed was a success.The high moisture content of 30 g/100 g in the blend, the gelatinization of cassava starch and a temperature of 115 °C, 120 °C and 130 °C produced excellent quality floating pellets with regards to expansion ratio, floatability, sinking velocity, durability index, water solubility index, water absorption index and bulk density.

1 =
Moisture(%) =Weight of sample taken−weight of dried material Weight of sample taken × 100 loss weight due to drying Weight of sample taken × 100 1

Table 2
Proximate composition of DFM and BSFL Means ± standard deviation Mean values in the same row denoted by the same superscript are not significantly different at (ANOVA; P > 0.05) DFM Danish fishmeal, BSFL black soldier fly larvae, DBSFL defatted black soldier fly larvae

Table 3
Mineral composition of defatted BSFL sample