Introduction

In order to understand fine emulsion products better, as it is manufactured in South Africa, we have to understand soya. I turned to Prof. Zeki Berk’s 1992 work when he was at Technion, Israel Institute of Technology, Haifa, Israel. The work was titled Technology of Production of Edible Flours and Protein Products from Soybeans.

What follows is a selection of quoted sections from his book relevant to our discussion on emulsion sausages and polony. By way of introduction, let me make a few notes about Zeki Berks (1931-2019) based on an obituary written by Prof. Sam Saguy (19.07.2019).

In our trade we stand on the shoulders of giants. I have done a review of the men and woman who brought about the current understanding and methods used in meat curing in my article, Fathers of Meat Curing. Here is another giant on whose shoulders we stand in our understanding of soy processing! It was thrilling to discover a work by a consummate professional and talented academic on such a subject. Right from the start, I was struck by the quality of his work. I looked for details on him and was saddened that he passed away last year. It is nevertheless a thrill to know that I am taught, as it were, by a man from beyond the grave, as we are often influenced by the lives of people who are no longer with us.

Prof. Saguy wrote about Zeki Berk, “For all his numerous students around
the world, he was an icon, beacon and the compass who taught and implemented basic and applied science, technology and devoted his life to education and excellence. In addition, he was also a person that symbolized more than probably anything else being kind, receptive, a great listener and above all redefining the meaning of a ‘Mensch’. His delightful and brilliant cooking skills as well as his amazing linguistics knowledge were extraordinary.”

“Zeki was a chemical engineer and food engineer and scientist with a long history of work in food engineering, including appointments as a professor at the Technion The Israeli Institute of Technology (IIT), M.I.T. and Agro- Paris, and as a consultant at UNIDO, FAO, the Industries Development Corporation, and Nestle. He was the recipient of the International Association of Food and Engineering Life Achievement Award (2011), and the first recipient of the Academic Life Time Achievement Award from the Food Industry Association-Manufacturers of Israel (2001). Prof. Berk
wrote/edited 7 books and numerous papers and reviews.”

I decided to quote his work here to add it to my own collection of invaluable resources for quick and easy access.

1. Utilization of Soy

Berk (1992) authored a comprehensive review of soy production, published in 1992. The data may be outdated, but his review is as relevant today as the day he wrote it.

The various uses of soy in 1986 are given below.

2. How Soya is Processed

– Roasting and Grinding Whole Soy

In my review of the health concerns associated with soya (Soya: Review of Health Concerns and Applications in the Meat Industry), I noted that roasting soya has been practised from as early as 300 BCE and milling it from before the Han dynasty (202 BC–220 AD). It is some of the oldest processing technology known for soy.

Berk (1992) writes that roasted whole soybeans and their flour are used as ingredients in China, Japan, Korea and Indonesia. I am very interested to know what this looks like in practical terms and will make this the subject of a future investigation. Besides ingredients, roasted whole soybeans and its flour are also used in traditional confectionery products and snacks in the same countries.

Another very interesting utilisation is that of immature whole green soybeans which are consumed as a vegetable. The use of the small plants has been practised from antiquity. By the time of the Han dynasty, it was already practised. Cooking the mature dry soybeans the way we do with other legumes ( such as navy beans, black beans, chickpeas or lentils) as was done in antiquity is seldom done even in the traditional areas of soybean consumption. “The reason for this may have been the persistent bitterness and “green beany taste” of soybeans, the low starch content, the relatively low water adsorption (swelling) capacity, long cooking time and poor digestibility.” Berk (1992)

– Soy: Oil Mill

“This option starts with the separation of the soybeans into two fractions: oil and meal. I deal with them side-by-side in the columns below. There are, basically, two process alternatives to achieve this purpose: pressing and solvent extraction. Each one of the fractions is then further processed to yield a multitude of products and by-products, with practically no waste. Since oil meal operations are often the starting point in the preparation of soybean protein products, they will be reviewed in some detail later in this article. The processes and products associated with the oil fraction will be described here in some detail. Soybean protein products which branch-off from the meal fraction will now be just mentioned for the sake of completeness and taken up in detail further on.” Berk (1992)

Full-Fat Soya

The application of heat removes the anti-nutritional factors of soy. As we have seen in my article, Soya: Review of Health Concerns and Applications in the Meat Industry, is not such a big factor for humans, but later in this article, we will show it has an immense impact on animal nutrition. A heating step before processing begins makes the oil also more accessible. Ottevanger Milling Engineers give the following process overview.

• Cleaning – at the start of the soybean processing, it is important to remove stones with a destoner, metal parts with a magnet and small grit & fines with a vibrating sieve.
• Crushing – a crusher will crush the bean in 4-8 particles, leaving the skin and crushed soybean. The hulls are removed from the crushed pieces through a wind sifter.
• Temperature – the crushed soybeans are brought up to temperature by adding steam in a conditioner. A toaster is used to keep the crushed soybeans at temperature for a longer period of time.
• Expansion – we use the expander for the expansion of the crushed and conditioned soybean into full-fat soy.
• Steam – the application of steam on the conditioner, toaster and expander is used to heat up and keep the product warm in order to improve gelatinization.
• Cooling – after expansion the product will be cooled to bring the product back to an ambient temperature.

This process is used to create full-fat soya, but you can see the important step of steam application right at the start of the process. (Ottevanger Milling Engineers)

3. OIL-MILL OPERATIONS

Operation principles

Continuous pressing by means of expellers (also known as screw presses) is a widely applied process for the extraction of oil from oilseeds and nuts. It replaces the historical method for the batchwise extraction of oil by mechanical or hydraulic pressing. The expeller (seen below) consists of a screw (or worm), rotating inside a cylindrical cage (barrel). The material to be pressed is fed between the screw and the barrel and propelled by the rotating screw in a direction parallel to the axis. The configuration of the screw and its shaft is such that the material is progressively compressed as it moves on, towards the discharge end of the cylinder. The compression effect can be achieved, for example, by decreasing the clearance between the screw shaft and the cage (progressive or step-wise increase of the shaft diameter) or by reducing the length of the screw flight in the direction of the axial movement. The gradually increasing pressure releases the oil which flows out of the press through the slots provided on the periphery of the barrel, while the press-cake continues to move in the direction of the shaft, towards a discharge gate installed at the other extremity of the machine.

Before entering the expeller, the oilseeds must be cleaned, dehulled (optional), flaked, cooked and dried. Flaking facilitates oil release in the press by decreasing the distance that the oil will have to travel to reach the particle surface. Cooking in the presence of moisture is essential for the denaturation of the proteins and, to some degree, for the coalescence of the oil droplets. Cooking plasticizes the flakes, renders them less brittle and thus reduces the extent of flake disintegration as a result of shear in the press. Extensive flake disintegration would reduce oil yield and produce a crude oil with a high content of fine solid particles (foots). After cooking, excess moisture is removed in order to avoid the formation of muddy emulsions in the press. Cooking is usually achieved by mixing the flakes with live steam. Additional heat may be provided by indirect steam, while thoroughly mixing the mass.

3-1-2 Advantages and disadvantages of the expeller process

Expellers can be used with almost any kind of oilseeds and nuts. Therefore, in a multi-purpose plant built to process different types of raw materials and not only soybeans, the expeller process may prove advantageous. The process is relatively simple and not capital-intensive. While the smallest solvent extraction plant would have a processing capacity of 100-200 tons per day, expellers are available for much smaller capacities, from a few tons per day and up.

The main disadvantage of the screw-press process is its relatively low yield of oil recovery. Even the most powerful presses cannot reduce the level of residual oil in the press-cake below 3 to 5%. In the case of oil-rich seeds such as sesame or peanuts, this may still be acceptable. Furthermore, most of the oil left in the cake can be recovered by a stage of solvent extraction. Such two-stage processes (pre-press/solvent extraction) are now widely applied. In the case of soybeans, however, a 5% residual oil level in the cake represents an oil loss of about 25%. Solvent extraction of the cake would not be economical, because of the bulk of material which must be processed. Pre-press/solvent extraction processes are, therefore, not applied to soybeans.

The commercial value of the meal is usually higher than the income from sales of the corresponding quantity of oil. The quality of the meal is therefore a factor of particular importance in the selection of a processing method for soybeans. In this respect, the expeller process has several disadvantages. The first is the poor storage stability of the press-cake, due to its high oil content. Furthermore,the extreme temperatures prevailing in the expeller may impair the nutritive value of the meal protein, mainly by reducing the biological availability of the amino acid lysine. At any rate, expeller press-cake is not suitable for applications requiring a meal with high protein solubility.

3-1-3 Equipment

Unlike solvent extraction equipment which is supplied by a relatively small number of manufacturers, screw presses with a widely varying degree of sophistication are available from a multitude of sources. Yet, considerable technical improvement and advanced features can be found in the models offered by the leading manufacturers. Such features include: multi-stage pressing to increase oil yield, better feed rate control, water cooled barrel and shaft, ease of maintenance and repair, improvements in the drive and transmission, sanitary construction, safety features etc. Most press manufacturers also supply cooker-dryer units, designed to operate with the press. Cooker-dryers may be horizontal (jacketed screw conveyor type), but the most common types consist of vertical stacks of round chambers (rings) equipped with paddle stirrers.

This design is indeed very common in operations for heat treating oilseed material and will be encountered in flake conditioners, desolventizers, meal dryers and coolers.

3-2 The solvent extraction process

3-2-1 Operation principles

A flow diagram describing the solvent extraction process for soybeans is given in the figure below. The process consists of the following stages:

a- Receiving and storage of soybeans.
b- Preparation of the raw material for extraction.
c- Solvent extraction.
d- Recovery of the solvent from the extract (micella).
e- Desolventizing/toasting of the meal.

3-2-2 Receiving and storage of soybeans

Nowadays, soybeans are received at the factory, almost exclusively in bulk, by truck or rail. They are weighed, unloaded and conveyed to the main storage silos. The size of the silos depend on the frequency of reception and the availability of other storage facilities in the region. Normally the main storage volume should correspond to the raw materials needed for a few months of operation at full capacity.

Pneumatic conveying is used in large installations while mechanical conveyors and elevators are more common in smaller plants. It is extremely important to maintain good sanitary conditions on and around the receiving areas and especially, to protect the seeds from contact with moisture. The receiving area, which consists of outdoors installations with a fair amount of movement of people and vehicles, tends to be one of the most critical parts of the factory, from the sanitation point of view.

As soybeans are purchased by grade, it is necessary to draw representative samples for quality evaluation from each lot at the point of reception. The samples are analyzed for moisture, foreign materials, colour, broken beans etc. in order to determine the compliance of the lot with the specified grade criteria. It is also advisable to determine oil and protein content, free fatty acids and other quality factors for the sake of proper bookkeeping, even if these criteria are not part of the standard grading and pricing system.

The typical storage facility in soybean oil plants is the vertical cylindrical silo. In recent years the conventional concrete silo is being replaced by steel silos of different types. A recent innovation in this area is a silo construction method based on the use of a steel strip wound in the form of a continuous spiral, each winding being fastened to the next one by crimping. The steel strip is supplied as compact coils, thus reducing the cost of transportation of bulky pre-fabricated constructions. One of the advantages of the metal silos is the speed of erection.

3-2-3 Preparation for extraction

This stage comprises drying, tempering, cleaning, classification (optional), cracking, dehulling (optional), conditioning and flaking. A flow diagram for the conventional preparation of soybeans prior to solvent extraction is given in below.

a- Drying: If the soybeans are to be dehulled before extraction, they must be dried to a moisture content below 10% in order to facilitate separation of the hulls. This is achieved in vertical gas or oil fired forced circulation driers. If the natural moisture content of the beans is 10% or less, or if dehulling is not practised, drying as a preparation step can be omitted.

b – Tempering: After cooling, the dried soybeans are stored in bins for 2 to 5 days, in order to allow for moisture equilibration by diffusion. This is called tempering. The tempering bins, which are usually outdoors silos of the vertical type, also serve as working bins (day bins), to secure uninterrupted feeding of the plant. As all the subsequent steps of processing are continuous, it is necessary to monitor the flow of soybeans from the working bins to the processing plant, in accordance with the planned processing capacity. This is done by means of automatic balances installed at the feed-end of the line.

c – Cleaning: The soybeans are subjected to a number of cleaning operations throughout the process. Tramp iron is removed by magnetic separators. In moderate capacity installations, these can be magnets attached to conveyors or chutes carrying a stream of beans. For larger plants, revolving drum type magnets which permit continuous removal of tramp iron from magnet surface are used. Both permanent magnets and electromagnets can be used. Permanent magnets have the advantage of being practically maintenance-free. Furthermore, they do not consume electrical power. Since the beans may become re-contaminated with stray iron (loose nuts and bolts, nails etc.) as they pass through the machinery, magnetic cleaning is not a one-time operation but must be repeated several times along the line. It is therefore advisable to install magnetic separators at the entrance of each machine where the presence of metal particles may cause serious damage (cracking mills, flaking machines etc.)

Stones, sand, dust and other foreign materials are usually removed by conventional seed cleaners. Typically, the seed cleaner consists of a two-deck vibrating screen. The upper screen retains the stones and other coarse materials but allows whole soybeans to fall through. The lower screen retains the soybeans but lets finer particles such as sand to pass through. Light trash, free hull particles and dust are removed by aspiration and trapped in cyclones.

d – Classification: The purpose of this operation is to separate split beans from whole beans. This step is optional and it is applied only if the meal is to be processed for human consumption. Classification is carried out by a simple sifting operation.

e – Cracking: The purpose of this operation is to break the seeds into smaller particles in preparation for flaking. If the beans have been dried to 10% moisture and tempered as described above, cracking also loosens the hulls and permits their separation by aspiration. Ideally, the seeds should be broken to 4 to 6 pieces of fairly uniform size. Production of fines should be minimized. Cracking machines consist of pairs of counter-rotating, corrugated rolls. One roll in each pair rotates faster than the other, to provide the shearing effect necessary to break the seed. Roll diameter is in the order of 25 cm. Roll length depends on the capacity. Two or three pairs of rolls are provided, mounted one on top of the other. A vibrating conveyor secures feeding of the mill at a uniform rate. The corrugations on the upper pair of rolls are coarser and deeper than those on the lower pairs.

A vibrating screen is provided at the exit from the mill. This is where the stream of broken particles is separated into hulls (removed by aspiration for further processing), oversize particles (returned to the cracking mill), meats of the correct size (sent to conditioning and flaking) and fines (usually mixed with the meats for conditioning).

The surface of cracking rolls is subject to considerable wear. After a certain service period, it may be necessary to renew the corrugations (refluting). Good quality rolls may be refluted several times before it becomes necessary to replace them.

f – Conditioning: The purpose of this operation is to increase the plasticity of the meats, in preparation for flaking. The conditioner is similar to the cooker described in connection with expellers. It can be a horizontal screw conveyor type heated reactor or a vertical stacked cooker. Heat can be provided by indirect steam or by direct steam injection, the latter being used to increase the moisture content when necessary. The meats are heated to 65-70^oC and the moisture content is brought to 10.5-11%. At this point the plasticity of the meats is such that they can be flattened by pressure in the flaker, without breaking.

g – Flaking: Flaking machines consist of a pair of horizontal counter-rotating smooth steel rolls. Typical roll sizes are in the range of 60-80 cm. in diameter. The rolls are pressed one against the other by means of heavy springs or by controlled hydraulic systems. Conditioned soybean cotyledon particles are fed between the rolls and they are flattened as the rolls rotate one against the other. The roll-to-roll pressure can be regulated and it determines the average thickness of the flakes. The main purpose of flaking is to increase the contact surface between the oilseed tissues and the solvent and to reduce the distance that the solvent and the extract will have to travel in the process of extraction. It is also believed that flaking disrupts the oilseed cells to some degree and thus makes the oil droplets more available for solvent extraction. Typical values for flake thickness are in the range of 0.2 to 0.35 millimetres.

Flaking rolls require maintenance as they wear considerably. To maintain the smoothness of the surfaces and to secure good contact between the rolls at every point, the rolls are reground from time to time. This requires expertise and accurate machines. In order to compensate for uneven thermal expansion, the rolls are manufactured not as perfect cylinders but with a slightly curved profile, thinner at both ends and thicker in the middle. Furthermore, the wear is usually not uniformly distributed and tends to be more extensive at the middle. Some manufacturers supply grinding devices which allow the roll ends to be reground without removal of the rolls.

h: Alternative processes: The processes described above are conventional oil-mill operations. Recently, improved processes have been suggested for individual steps or for the whole seed preparation line.

The ” Hot Dehulling (Popping) System “, offered by Buhler-Miag Ltd. makes use of a “shock treatment” to loosen the hulls.

Soybeans with a moisture content of about 13% are preheated to 60^oC, then contacted with a stream of hot air in a fluidized bed unit. This treatment causes popping of the hull. Now the seeds are split in half by impact and the hulls are separated by air. The dehulled split beans are further cracked and flaked. The main advantage of the process is its lower energy consumption since the multiple heating and cooling, drying and humidification steps of conventional dehulling are obviated. The short duration of the heat treatment step prevents extensive protein denaturation. The reduction in NSI (Nitrogen Solubility Index) is claimed to be essentially the same as in conventional dehulled flake preparation. A process flow diagram for the Hot Dehulling System is given below.

The “Alcon Process” offered by Lurgi GmbH, consists of a series of operations installed between the conventional preparation line (right after the flaking mill) and the extractor. The flakes are humidified and heated in a conditioner, maintained at the desired moisture content and temperature for 15-20 minutes (tempering), then dried and cooled before being led to the extractor. This is, essentially, an agglomeration process, whereby the flakes are fused into more compact, porous granules. The following benefits are claimed:

a: The bulk density of the modified granules is by 50% higher than that of the original flakes (550 against 360 kg/m³). This results in a corresponding increase in extractor capacity.

b: The rate of percolation of micella or solvent through the granules is tripled. This results in improved extractor efficiency (see below).

c: Solvent retention in the spent granules is about 25%, while conventional spent flakes may retain as much as 35% solvent. As a result, desolventizer capacity is increased, oil yield is improved and energy is saved.

d: During preparation and extraction, certain enzymes reduce the hydratability of the phospholipids. The thermal treatment associated with the Alcon process inactivates these enzymes and improves the efficiency and yield of the oil degumming process.

e: Due to the thermal treatment mentioned above, meal toasting requirements are less severe.

In the Pellet method suggested by the FRENCH Oil Mill Machinery Company, the crushed material is extruded as pellets. The extruder which is called “the Enhanser Press”, is equipped with special ports for the injection of steam or water into the barrel. The mass is pressed through the holes on a die plate, expands as a result of the sudden evaporation of water and yields firm pellets with sufficient internal porosity but a bulk density higher than that of flakes. The advantages claimed are essentially the same as those of the other agglomeration processes.

A drawing showing the FRENCH Enhancer Press is given in the figure below.

3-2-4 Solvent extraction:

a – Basic principles of solvent extraction: The extraction of oil from oilseeds by means of non-polar solvents is, basically, a process of solid-liquid extraction. The transfer of oil from the solid to the surrounding oil-solvent solution (micella) may be divided into three steps:

* diffusion of the solvent into the solid
* dissolution of the oil droplets in the solvent
* diffusion of the oil from the solid particle to the surrounding liquid.

Due to the very high solubility of the oil in the commonly used solvents, the step of dissolution is not a rate limiting factor. The driving force in the diffusional processes is, obviously, the gradient of oil concentration in the direction of diffusion. Due to the relative inertness of the non-oil constituents of the oilseed, equilibrium is reached when the concentration of oil in the micella within the pores of the solid is equal to the concentration of oil in the free micella, outside the solid. These considerations lead to a number of practical conclusions:

* Since the rate-limiting process is diffusion, much can be gained by reducing the size of the solid particle. Yet, the raw material cannot be ground to a fine powder, because this would impair the flow of solvent around the particles and would make the separation of the micella from the spent solid extremely difficult. Instead, the oilseeds are rolled into thin flakes, as described in the previous paragraph, thus reducing one dimension to facilitate diffusion, without impairing too much the flow of solvent through the solid bed or contaminating the micella with an excessive quantity of fine solid particles. The effect of flake thickness on the efficiency of solvent extraction is demonstrated in the figure below.

* The rate of extraction can be increased considerably by increasing the temperature in the extractor. Higher temperature means higher solubility of the oil, higher diffusion coefficients and lower micella viscosity. In fact, it is customary to heat the solvent and the intermediate micella to the highest temperature which would still provide an acceptable level of safety.

* An open, porous structure of the solid material is preferable, because such a structure facilitates diffusion as well as percolation. A number of processes have been proposed for increasing the porosity of oilseeds before solvent extraction (See para. 3-2-3-h ).

* Although most of the resistance to mass transfer lies within the solid, the rate of extraction can be increased somewhat by providing agitation and free flow in the liquid phase around the solid particles. Too much agitation is to be avoided, in order to prevent extensive disintegration of the flakes.

* Since the concentration gradient is the factor responsible for moving the oil out of the solid, it is important to keep this gradient high, at each point within the extractor. This effect is obtained most economically by the principle of counter-current multistage extraction. The process is divided to a number of contact stages . Each stage comprises means for mixing the solid and the solvent phases and for separating the two streams after extraction has been achieved. In going from one stage to the next, the flakes and the solvent move in opposite directions. Thus, flakes with the lowest oil content are contacted with the leanest solvent, resulting in high oil yield and high driving force throughout the extractor. The principle of counter-current extraction is shown in Fig.15.

A detailed discussion of the theoretical basis for the design of multistage solid-liquid extraction processes is beyond the scope of the present work. We shall outline here only its principal practical consequences, as far as they provide useful criteria for the selection and operation of an extractor.

Two different methods can be used to bring the solvent to intimate contact with the oilseed material: percolation and flooding. In the percolation method, the solvent trickles through a thick bed of flakes without filling the void space completely. A film of solvent flows rather rapidly over the surface of the solid particles and efficiently removes the oil which has diffused from the inside to the surface. This mode of contact is preferable whenever the resistance to diffusion inside the flake is relatively low (thin flakes with large surface area, open tissue structure). In the flooding mode the solid particles are totally immersed in a slowly moving, continuous phase of solvent. Immersion works better with materials offering a greater internal resistance to oil transfer (thick particles, dense tissue structure).

The number of contact stages necessary to perform a given extraction operation depend on the following variables:

* Flakes/solvent ratio: If the quantity of solvent used to extract oil from one ton of flakes is increased, a smaller number of contact stages will be needed to achieve a given extraction job. However, the full micella resulting from the process would be less concentrated in oil, meaning that we would have to evaporate larger quantities of solvent for each ton of product, and hence, spend more on energy.

* Oil yield: If the number of stages is increased while all other variables are kept unchanged, the proportion of oil left in the spent flakes will be lower and therefore, the oil yield will be higher. The relationship between the number of stages and residual oil in the meal is shown in the figure below.

* Percolation: The quantity of solvent or micella retained within the capillaries and pores of the solid after drainage is called “bound extract” or “bound solvent”. This quantity depends on the properties of the flakes and solvent as well as the drainage conditions. Easy percolation of the solvent through the solid bed leaves less extract in the capillaries after drainage and results therefore, in a reduction of the number of contact stages needed. Proper preparation and handling of the flakes are important to ensure high percolation rate.

b- Choice of solvents:

An ideal solvent for the extraction of oil from soybeans should possess the following properties:

* Good solubility of the oil.
* Poor solubility of non-oil components.
* High volatility (i.e. low boiling point), so that complete removal of the solvent from the micella and the meal by evaporation is feasible and easy.
* Yet, the boiling point should not be too low, so that extraction can be carried out at a somewhat high temperature to facilitate mass transfer.
* Low viscosity.
* Low latent heat of evaporation, so that less energy is needed for solvent recovery.
* Low specific heat, so that less energy is needed for keeping the solvent ant the micella warm.
* The solvent should be chemically inert to oil and other components of the soybean.
* Absolute absence of toxicity and carcinogenicity, for the solvent and its residues.
* Non-inflammable, non-explosive.
* Non-corrosive
* Commercial availability in large quantities and low cost. 

Unfortunately, the ideal solvent possessing all these properties does not exist. Most of the requirements, with the notable exception of flammability and explosiveness, are met by low-boiling hydrocarbon fractions obtained from petroleum. A typical commercial solvent for oil extraction would have a boiling point range (distillation range) of 65 to 70^oC and would consist mainly of six-carbon alkanes, hence the name “hexane“by which these solvents are commonly known in the U.S.A.. “Hexane ” solvents for the extraction of edible oil must comply with strict quality specifications. The quality parameters which make up the specifications usually include: boiling (distillation) range, maximum non-volatile residue, flash point,maximum sulphur, maximum cyclic hydrocarbons, colour and specific gravity.

The main shortcoming of light hydrocarbon solvents is their flammability and the explosiveness of mixtures of their vapours and air. Safety considerations gave led to the enforcement of special standards for buildings and installations in solvent extraction plants. All the electrical installations have to be explosion-proof. The discharge end of all vents have to be equipped with refrigerated condensers to minimize escape of solvent vapours to the atmosphere. Very strict safety measures are taken to prevent the hazard of sparks in and around the plant. All these add to the high cost of erection and operation of solvent extraction plants.Even so, accidents are not uncommon.

The continuous search for alternative solvents is, therefore understandable. One such solvent, trichloroethylene, was in commercial use for a short period in the early 1940’s, but had to be abandoned when it was discovered that the meal prepared in this way was toxic to animals. Another alternative approach makes use of “supercritical extraction” with liquid carbon dioxide under high pressure. Although technically feasible, supercritical extraction of soybean oil is not commercially viable at present, due to the high cost of the equipment and the relatively poor oil dissolving capacity of carbon dioxide near its critical point. Alcohols constitute yet another class of potential solvents for oil extraction. Water-free (absolute) low aliphatic alcohols such as ethanol and isopropanol are fairly good solvents for oils at high temperature but the solubility of oils in these solvents decreases drastically as the temperature is lowered. This high dependence of solubility on temperature is precisely the principle on which alcohol extraction processes are based. Extraction takes place at high temperature. The micella is then cooled. Saturation occurs and excess oil separates as a distinct phase which can be recovered by centrifugation. The solvent is reheated and sent back to the extractor. These alcohols are less flammable then hexane, but precautions are still necessary. Despite considerable research efforts to develop alternative solvent systems, extraction with light hydrocarbons continues to be, practically, the only commercial solvent extraction process for soybean oil.

c- Types of extractors:

Solvent extractors are of three types: batch, semi-continuous and continuous.

In batch processes, a certain quantity of flakes is contacted with a certain volume of fresh solvent. The micella is drained off, distilled and the solvent is recirculated through the extractor until the residual oil content in the batch of flakes is reduced to the desired level. Batch extractors as industrial units are now obsolete. Laboratory and pilot plant size extractors are still used for experimentation and instruction purposes.

Semi-continuous systems consist of several batch extractors connected in series. The solvent or micella flows from one extractor to the next one in the series. The material in the first extractor is the most exhausted, since it has been treated with fresh solvent. After a while, the second extractor is made “head” of the series and connected to the fresh solvent line. The spent flakes are discharged from the first extractor, which is then filled with a batch of fresh flakes and is connected to the system as the “tail” unit, and so on.

Semi-continuous systems of the type described above are seldom used for the solvent extraction of soybeans. However, the same principle is applied in one of the widely known solvent extraction systems for other oilseeds: the FRENCH Stationary Basket Extractor.

The FRENCH extractor is essentially a vertical cylindrical vessel, divided into a number of tall vertical sections or “baskets” by radial walls. The baskets are stationary. Solvent or micella is fed at the top of the basket and percolates through the deep bed of solids. Using a system of moving micella showers, the oilseed material is contacted with micella at decreasing oil content, and finally with fresh solvent, thus achieving countercurrent extraction, without moving the solid bed. In its recent version, the FRENCH stationary basket extractor is equipped with a rotating basket bottom, to achieve automatic discharge of the baskets at the correct time and to render the extractor nearly continuous. The capacities of units supplied since 1975 for soybean oil extraction, range from 100 to 3000 tons per day.

In continuous extraction, both the oilseeds and the solvent are fed into the extractor continuously. The different available types are characterized by their geometrical configuration and the method by which solids and solvents are moved one in relation to the other, in counter-current fashion. The most prominent types will be described in the next paragraphs.

* Belt extractors_ the DE SMET extractor: This extractor, offered by the Belgian De Smet Company and its subsidiaries in many countries, was developed in 1946 by J.A. De Smet at the “Nouvelles Huileries Anversoises” oil mill in Belgium. According to the company, since then over 450 plants using the DE SMET process have been built in various parts of the world.

A drawing describing the DE SMET Extractor is given in the figure below. The extractor consists of a horizontal, sealed vessel in which a slowly moving screen belt is installed. Flaked soybeans are fed on the belt by means of a feeding hopper. A damper attached to the hopper outlet acts as a feed regulating valve and maintains the solids bed on the belt at constant height. This height can be adjusted according to the expected rate of percolation of the micella through the bed. Difficult percolation is compensated for by lowering bed height. For properly flaked soybeans, the height of the flake bed at the head end of the extractor is normally 6 to 8 feet (180 to 240 cm.). The throughput rate of the extractor is adjusted by changing the belt speed. There are no dividing baffles on the belt and the solid bed is one continuous mass. Yet the extractor is divided to distinct extraction stages by the way in which the micella stream is advanced. The solvent is introduced at the spent flake discharge end ( i.e. at the end opposite to the flake feeding side of the extractor ). It is sprayed on the flakes, percolates through the bed, giving the spent flakes a last wash and removing some oil. The resulting dilute micella is collected in a sectional hopper underneath the belt, from which it is pumped and sprayed again on the flakes at the next section in the direction opposite to belt movement. This process of micella collection, pumping and spraying at the next section is repeated until the micella leaves the hopper at the head-end of the extractor, carrying the highest concentration of oil (heavy micella). The screen is washed with heavy micella at the head-end, just before the entrance of fresh flakes, and then again with fresh solvent, right after the discharge of spent flakes.Washing of the screen is essential to prevent clogging. Washing with full micella at the feed-end provides surface lubrication and prevents adhesion of the flakes to the surface of the screen. The entire extractor vessel is maintained at a slight negative pressure so as to prevent leakage of solvent vapours to the atmosphere.

According to the manufacturers, DE SMET extraction plants have been built for capacities ranging from 25 to 3000 tons of raw material per day. Solvent losses are 0.07% to 0.3% and the residual oil content of the extracted material is 0.25% to 0.6%.

* Moving basket extractors: In this class of extractors, the flakes do not constitute a continuous mass but are filled into separate, delimited elements (baskets) with perforated bottoms for draining. The baskets can be moved vertically (bucket elevator extractors), horizontally ( frame belt and sliding cell extractors), or can be rotated around a vertical axis (carrousel extractors). Vertical bucket-chain extractors are among the first industrial solvent extractors constructed for continuous operation. Many are still in operation but they are less frequently found in more recent installations.

In the horizontal moving basket extractors manufactured by the LURGI Company, the “basket” or “cell” is formed by an endless bucket belt and a separate perforated bottom. The bottom can be fixed perforated plates on which the bucket separations slide (sliding cell design) or screen belt conveyors moving with the buckets. Both types are shown below.

Another type of horizontal basket extractor, featuring tilting baskets or trays, is manufactured by the HLS Company Ltd. The operation principle of the T.O.M. (Turning Over of Material) HLS extractor is shown in Fig.20. Each basket in the extractor can be flooded, permitting immersion and percolation in the same extractor. In order to overcome the problem of the formation of a dense surface layer of compressed fines, the trays or baskets are inverted at the end of the conveying chain. The material falls to the basket or tray below. The impermeable surface layer is broken and the oilseed material undergoes mixing in the process of its transfer from one level to the other. Extraction continues as the material moves, in reversed direction, on the lower (return) side of the conveyor. Thus, unlike most horizontal extractors, in the HLS Extractor the inlet for fresh raw material and the outlet for the spent flakes are on the same end of the shell.

* Carrousel extractors somewhat resemble the cylindrical FRENCH extractor described above, but here, the “baskets” rotate around the axis of the cylinder while the solvent/micella circuitry is fixed. The construction principle of the Carrousel Extractor, manufactured by EXTRACTIONSTECHNIK GmbH, is shown in below. The following description of the extractor and its operation is from an article by Dr. Ing. Wolfgang Kehse:

” The extractor consists of a single-part rotor with an inner and outer cylindrical wall. The ring-shape interspace is divided by radially arranged conical partition walls into a number of chambers (10 to 20) . It is slowly rotated usually by chain drive, the larger gear rim of which is placed round the rotor. Smaller extractors may be directly driven by a central shaft. These rotation speeds vary from one rotation in 20 minutes up to one rotation in 4 to 5 hours, and are adjustable. The rotor rotates above a slitted bottom with only a few millimetres’ gap. This slitted bottom is constructed of profiled rods with a trapezoidal cross-section. This profile causes the slits which are at their surface about 0.8 mm wide to become wider further down. The specific advantage of this slitted bottom, however, is that the slits are exactly concentric with the rotor shaft. The raw material is filled into the chambers and thus form a compact layer which can reach a height of from 0.5 to 2.5 meters, depending on the material to be extracted. The height of the rotor corresponds to this. Therefore, a free space of about 200 mm above the layer remains, which is filled with liquid solvent during the time that the chamber is being sprayed with solvent.

Depending on the required time for extraction, the material is moved at a speed of 1-10 mm/sec., over the concentric slits in the bottom. Because the slits are arranged parallel to the direction of movement of the material, no mechanical forces apart from the sliding resistance are exerted on the extraction material and subsequently no plugging of the slits can occur. While moving over the slitted bottom, the bed of material is percolated by micella of different concentrations, beginning with the end-micella having the highest concentration immediately after feeding of the solid material up to the pure solvent at the end of its passage. The micella passes through the bed of material and the slitted bottom and is the collected in chambers separated by weirs in the lower part of the extractor. From there it is pumped back onto the bed of material. The discharge of the extracted solid material is effected through the slitted bottom by a hole as wide as a rotor chamber and allowing the contents to drop down into a discharge chute where it is moved on for further processing by a screw conveyor.
The partition walls of the chambers are conically widening downward so that any sticking of the chamber contents is impossible.”

According to the manufacturer, Carrousel Extractors are available in capacities from 20 up to 4000 tons per 24 hours. The largest extractor (4000 tpd.) has a nominal diameter of 15 m.

3-2-5 Post-extraction operations

Two streams leave the solvent extraction stage: an oil-rich fluid extract (full micella) and solvent-laden spent flakes. The next operations have the objective of removing and recovering the solvent from each one the two streams.

a: Micella distillation: Full micella contains typically 30% oil. Thus, for every ton of crude oil some 2.5 tons of solvent must be removed by distillation. Most manufacturers of solvent extractors also offer micella distillation systems. The characteristics of a good micella distillation system are: good energy economy, minimal heat damage to the crude oil and its components, minimal solvent losses , efficient removal of the last traces of solvent from the oil and, of course, good operation safety. The modes of solvent vaporization include flash evaporation, vacuum distillation and steam stripping.

b: Meal desolventizing: The spent flakes carry with them about 35% solvent. The removal and recovery of this portion of the solvent is also one of the most critical operations in oil mill practice, since it determines, to a large extent, the quality of the meal and its derivatives.

In desolventizing-toasting (DT) applied in the production of soybean oil meal for animal feeding, the time-temperature-moisture profile of the process permits, in addition to solvent removal, a heat treatment sufficient to inactivate the undesirable enzymes and inhibitors and to improve the palatability of the meal to animals (toasting). The most common type of desolventizer-toaster consists of a vertical cylindrical stack of compartments or “pans”. Each compartment is fitted with stirrers or racks attached to a central vertical shaft. Spent flakes are fed at the top of the desolventizer-toaster. The pan floors are equipped with adjustable-speed rotating valve, to permit downward movement of the material , through the pans, at the desirable rate. Two methods of heating are used: direct steam heating and indirect steam heating. For heating with indirect steam, the pans are equipped with double bottoms acting as steam jackets. For direct steam heating, hot live steam is injected into the mass through spargers. The rotating stirrers spread the material and provide the necessary mixing action. Direct steam is used for three reasons:

* The transfer of heat from the heated surface of the pan floor to the oilseed material is slow and difficult, especially after a considerable proportion of the solvent has been removed and no fluid medium is available for heat transfer. In this case, direct contact between the solid material and condensing steam is a more efficient method of heating. Condensation of the steam adds moisture to the flakes.

* The added moisture facilitates the protein denaturation reactions leading to the inactivation of trypsin inhibitor. It is also believed that the toasting effect accomplished by the combined action of heat and moisture enhances the palatability of the meal to animals.

* The steam distillation effect is necessary in order to remove last traces of solvent from the meal.

The various models of vertical stack type DT’s differ in the sequence of direct/indirect heating zones and several other features. In the FRENCH DT shown below, the top pans are indirect steam heated. They constitute the pre-desolventizing zone. The bottom pans are direct steam heated and they serve as the toasting/stripping zone. The meal coming out of this DT has about 18% moisture and a temperature of about 105^oC. It has to be dried and cooled. A separate dryer/cooler (DC) is used for this purpose (see below).

The DE SMET DT shown in below has 4 to 10 pans with steam-heated bottoms. The apparatus is maintained at a slight negative pressure.

The meal dryer-cooler (DC) is similar to the DT in construction but much shorter. Ambient air is used to dry and cool the meal before storage or bagging. The construction of a self-standing DC unit, offered by FRENCH, is shown below.

The DT and DC units can also be combined into one piece of equipment. Most manufacturers of desolventizing equipment also offer combined DTDC units. The operating principle of such a system, sold by LURGI is shown below.

A photograph of a 1200 ton per day desolventizer-toaster-dryer is given in below.

While desolventizing-toasting is the standard method for the manufacture of soybean oil meal for animal feeding, this process is not suitable for the production of “white flakes”, i.e. meal with minimum protein denaturation. As it can be seen below, protein denaturation ( expressed as the reduction in Nitrogen Solubility Index, NSI) by treatment with live steam is very rapid. White flakes, which are the starting material for the production of soybean protein isolates, most concentrates and texturized products, must have a high NSI value.

The best method of desolventizing for the production of white flakes is flash desolventizing (FD). In this process, the solvent laden spent flakes coming out from the extractor are fluidized in a stream of superheated solvent vapours. The superheat of the vapour provides the energy for the evaporation of solvent from the flakes. The turbulent nature of the flake-vapour flow permits extremely rap[id heat and mass transfer. Protein denaturation is minimized, mainly because of the short heating time. A short stripping stage may be necessary to complete solvent removal and rapid cooling is a must for preventing undue reduction of NDI. The flow-diagram of a flash desolventizing system is shown below.

4. EDIBLE SOYBEAN FLOURS AND GRITS

4.1 Introduction

Flours and grits are the simplest of all edible soybean protein products. The extent of processing which goes into their production is minimal. The cost of extra processing, starting with the dehulled clean beans (for full-fat flour) or with dehulled white flakes (for defatted flour), has been estimated at 60 to 100 U.S.$ per ton. The total cost of the product, in the bag, at the production site, would then be less than $400 per ton. Recently (January 1991) a leading supplier in the U.S. has quoted soybean flour at $14.00/cwt. ( approximately $308 per metric ton), ex-factory. This makes soybean flour one of the most economical sources of edible protein. Speciality flours, produced in smaller quantities, may be more expensive.

The annual production of edible soybean flours and grits increased from some 60,000 tons in 1960 to about 2,000,000 tons today.

The production of edible soybean flours and grits may take place either as an independent industrial activity or as a natural sequel of oil-mill operations. In fact, many oil-mills, recently erected in various parts of the world, feature production lines or departments for edible products, in addition to the usual oil and meal lines. The principal differences between processing for meal and processing for edible flour are in the quality of the raw material, the need for dehulling and the more rigorous control of the sanitary conditions of the plant and the process. Frequently, oil-mill operators prefer to produce only edible products or only meal, in alternate fashion rather than simultaneously.

4.2 Defintions, composition and quality parameters

4-2-1 Definition and classification of edible soy flours and grits

Soy flours are products obtained by finely grinding full-fat dehulled soybeans or defatted flakes made from dehulled soybeans. To be called soy flour, at least 97% of the product must pass through a 100-mesh standard screen. (A 100-mesh screen has 100 openings per inch.)

Soy grits have essentially the same composition as flour, but coarser granulation. They are usually classified into three groups, according to particle size:

Coarse 10 to 20 mesh
Medium 20 to 40 mesh
Fine 40 to 80 mesh

Circle and Smith (1972) have pointed out that the name soy flour may be misleading, since its composition is totally different from that of the popular product commonly known as flour, i.e. wheat flour. They suggested alternative names such as “defatted soy solids” (as non-fat milk solids) or “soy powder” or “soy pulverate”.

Edible soy flours are made from dehulled beans, hence their relatively low crude fibre and high protein content.

Soy flours (or grits) are classified according to their lipid content as follows:

* Defatted soy flour, obtained from solvent extracted flakes, contains less than 1% oil.

* Full-fat soy flour, made from unextracted,dehulled beans, contains about 18% to 20% oil.

* Low fat soy flour, made by adding back some oil to defatted soy flour. Lipid content varies according to specifications, usually between 4.5% and 9%.The most common range is between 5% and 6%.

* High fat soy flour, produced by adding back soybean oil to defatted flour, usually at the level of 15%.

* Lecithinated soy flour, made by adding soybean lecithin to defatted, low fat or high fat soy flours in order to increase their dispersibility and impart emulsifying properties.. Lecithin content varies according to specifications, usually up to 15%.

Commercial soy flours and grits are further classified according to their Nitrogen Solubility Index (NSI), or their Nitrogen Dispersibility Index (NDI). It will be recalled that these parameters indicate the extent of protein denaturation and hence the intensity of heat treatment which has been applied to the starting material. Flours made from “white flakes” have NSI values of about 80%, while those made from toasted flakes show NSI levels of 10 to 20%. Other grades are available over the entire range of intermediate NSI values.The specification of a specific value of NSI reflects , in fact, a compromise between the need to maintain the functional properties of the soy proteins or some enzyme activity, and the desire to inactivate anti-nutritional factors and eliminate the beany taste, all in function of the end use.

4-2-2 Composition

The typical composition of different types of soy flours is given in Table 4-1. The basic composition of soybeans is added for comparison. Since the moisture content of the products may vary during storage, the percentage figures for protein, fat, fibre and ash are given on a moisture-free basis. A typical level of moisture content is also shown.

4-2-3 Quality standards

In addition to the identity standards and definitions mentioned above, quality standards have been formulated by official agencies (e.g. FAO/WHO/UNICEF Protein Advisory Group). Trade specifications usually exceed the official standards. The quality parameters which constitute a specification usually include:

a- Composition:
Protein	a minimum value
Fat	a maximum value for defatted flour a range for others
Lecithin	a range for lecithinated flours
Crude fibre	a maximum value
Ash	a maximum value
Moisture	a maximum value
b- Physical parameters:
Granulation	as mesh number or particle size distribution.
c- Microbiology:
Total plate count	a maximum value
Coliforms	a maximum value
Salmonella	a maximum value (usually 0)
d- Heat treatment history:
Protein solubility	as NSI, NDI, PSI or PDI
Tryps ininhibitor activity
Urease activity
Lipoxidase activity	for enzyme-active flour.
Available lysine
e- Sensory parameters:
Colour
Taste
Odour
f- Defects:
Insect parts	a maximum value or total absence
Foreign material	”              ”                     “
Black specks	”              ”                     “
g- Commercial:
Packaging, delivery etc.

4.3 Full fat soy flour and grits

4-3-1 Production processes

a- Oil-mill related industrial production process: The process for the production of full-fat soy flour and grits as a side line of large scale oil-mill operation is relatively simple. It consists of three major steps: dehulling, heat treatment and milling.

Cleaned, grade A yellow soybeans are dried, tempered, classified to separate split beans, cracked and dehulled by aspiration. These operations are essentially similar to the seed preparation steps of an oil-mill, from raw material silos up to the obtention of dehulled meats, and have been discussed in detail previously (Section 3-2-3, a to e).

The dehulled meats coming out of the vibrating screen are now subjected to humid heat, to achieve the specified product NSI value. This is conveniently done in a vertical conditioner with direct and indirect steam heating sections.

This step is obviously omitted if the final product is to be unheated (enzyme active) flour. The last sections of the conditioner are used to dry the meats to a moisture content below 10%.

The properly conditioned and dried meats are cooled and then finely ground. Hammer mills, pin mills, impact turbo mills and similar pulverizers are used to grind the meats so that not more than 3% of the product will be retained by a 100-mesh screen. In practice , full fat soy flour is difficult to screen on such fine sieves, due to particle agglomeration. Air classification systems which separate the fine product and recirculate the coarse fraction through the mill are more adequate than screen sifters.

b- Alternative processes: In the framework of the efforts to promote direct consumption of soybeans in the less industrialized parts of the world, methods for the preparation of full-fat soybean flour with a minimal amount of processing have been developed. These methods permit production of flours independently of the oil industry.

One such process has been described by Mustakas et al.(1967) In this village scale production method, the soybeans are soaked in water, then cooked in boiling water, air dried, cracked by hand, winnowed to separate the hulls and finally hand ground in a mortar or any other grinding device available.

A more industrialized version of the process (Mustakas et al.(1970) is similar in most aspects to the large scale production process described above, except for the step of heat treatment. In this process the flour is submitted to a continuous high temperature-short time humid heat treatment, using an extruder-cooker. The dehulled meats are first equilibrated with moisture in a direct steam fed conditioner/ tempering bin, then cooked under pressure in a continuous extruder/cooker. The extrudate is cooled and ground as usual. The HTST treatment eliminates the beany flavour and produces a light, open structured flour. A slightly different process, also centred around extrusion cooking, known as the WENGER PROCESS, is available from the Wenger Mixer Manufacturing Co. More recently, low-cost extruders have been made available for the less sophisticated extrusion-cooking applications. “Low-cost” may mean $5,000, compared to $100,000 for a regular extrusion system. Such low-cost extruders have been used for the preparation of full-fat soy flour. According to Lorenz et al.(1980), the total investment needed for a 550 kg./hr plant was (1980) about $120,000 including building and land. The cost of production, including raw materials, packaging materials and overhead was $223 per ton. Extruded full-fat soybean flour was being produced with a low-cost extruder in Mexico in 1980.

In extrusion cooking, the material reaches temperatures in the order of 150^oC. At such high temperatures, destruction of urease activity is no longer a credible indicator for the inactivation of trypsin inhibitor, which must be monitored directly.

The BUHLER PROCESS developed by Buhler Co. in Switzerland, is based on very fine grinding and fast heating. The resulting powder has been suggested as an alternative for soymilk solids.

A process, based on pre-germinated beans has been described by Suberbie et al.(1981). The beans are soaked in water for 3 hours and allowed to germinate. At the end of the germination period, the soybeans are steamed, dried to 6% moisture, dehulled and ground in a cooled hammer mill. Germination resulted in flavour and odour improvement. Milling capacity was impaired by germination. Pre-germinated full-fat soybean flour has been produced commercially in Mexico.

4-3-2 Utilization

The principal use of full-fat soybean flour, as well as re-fatted and lecithinated flours, is in the bakery industry. Two types of flour are used: enzyme-active and enzyme-inactive.

Enzyme-active full-fat soybean flour is prepared without heat treatment and has a high NSI value around 80%. It is used in bakery products (white bread and rolls), mainly for its lipoxidase activity. Lipoxidase catalyses oxidative bleaching of the carotenoid pigments in wheat flour. Enzyme-active soybean flour is a valuable “natural” flour bleaching agent, especially where the use of chemical bleaching agents has been prohibited. Lipoxidase activity is also beneficial to the mechanical properties of the dough. Since the soybean product is added in relatively small quantities (up to 0.5% on flour basis in bread and buns in the U.S.A.) the beany flavour of unheated soybeans is not a limiting factor. Usually, enzyme-active full-fat soy flour is not sold as such, but rather in mixtures containing other ingredients such as cornflour.

With the development of successful flash desolventizing systems which permit desolventizing without appreciable enzyme inactivation, defatted enzyme-active flours have largely replaced the full-fat product, especially in the U.S.A.

Enzyme-inactivated (heated) full-fat soybean flours, alone or with re-fatted and lecithinated soy flours, is mainly used in the heavier types of cake batters, such as sponge cake and pound cake. It contributes to the richness of the cake while increasing the proportion of water that can be added to the mix. Due to their oil and phospholipid content, these flours exert egg and shortening sparing effects and act as emulsifiers. In these formulae, soybean flours are used at the level of 3-5%, based on flour weight. Full-fat or lecithinated soy flour with high nitrogen solubility (NSI of 80%) has been found to improve eating quality and reduce fat absorption in doughnuts.

4.4 Defatted soy flours and grits

4-4-1 Production processes

The processes for the manufacture of raw or heated dehulled solvent extracted flakes have been described in section 3.

Usually, all the flakes made for edible products are flash-desolventized, then carefully steam-heated to the desired NSI value.

The final milling is critical and energy-consuming. Although identity standards require milling to 97% minus 100-mesh, specialty flours (such as those used as milk solids replacement in infant formulae) are ground to a finer particle size.

At such levels of fineness, the conventional hammer mill is practically useless. Impact turbo mills or high-speed pin mills have to be used.

4-4-2: Utilization:

a- Use in bakery and other cereal products:

Nutritionally, soybean protein is an excellent complement to lysine-limited cereal protein, hence the basis for the use of soy flour as an economical protein supplement in bread, tortillas, pasta and other cereal products. Supplementation of bread and other cereal staples with defatted soy flour has been promoted in a number of countries, and even enforced in some. The use of defatted soy flour in bread does not create any appreciable technological or quality problems, as long as less than 10% of the wheat flour has been replaced by soy flour. At higher replacement levels, up to 15%, loaf volume and crumb texture may be impaired. Baking quality can be recovered, however, by means of some adjustments such as higher yeast level, use of lecithin and other emulsifying agents etc.

Another bakery related potential use of soy flour in combination with cereals is in the production of the so-called “composite flours.” These are mixtures of flours, starches and other ingredients, supposed to replace wheat flour, totally or partially, in bakery products. Extensive research projects aimed at the development of such flours have been sponsored by international and national development agencies in the last 20 years or so. The main reason for developing composite flours is to relieve the economy of countries where wheat is not grown, from the burden of importing this commodity. Other reasons include the production of alternative baking flours for people who cannot tolerate wheat products (e.g. coeliac disease patients).

Considerable quantities of soy flour (1.5 to 2% on flour weight basis) are used in bakery products, particularly in white bread, as a replacement for nonfat milk solids. In this application, soy flour (and sometimes soy protein concentrate) is used in combination with whey solids. Milk replacer blends, consisting mainly of defatted soy flour, whey solids, caseinates and other nutritional or functional ingredients are available at protein content levels of 20% to 40%.

In many applications, especially in the U.S.A. and Europe, the largest quantity of soy flours is used in bakery products, not for nutritional reasons but rather for their functional characteristics.

Enzyme-active defatted flour is used as a bleaching and dough improving agent as discussed in the previous section dealing with full-fat flours.The characteristics of such a flour (SOYBAR, made by Solbar Hatzor Ltd.), as reported by the manufacturer, are given below, as an example:

Product description: Enzyme active defatted soy flour, derived from high quality, dehulled soybeans. Has a mild flavour and aroma profile and a light cream colour.

Characteristics: Highly dispersible in water. Has excellent water binding properties.

Specifications:
Protein (as is) 50% min.
Moisture	10% max.
Crude fibre	4% max.
Ash	6.5% max.
Fat	1.0% max.
Particle size	95% less than 74 microns
Standard plate count	50,000/g. max.
Salmonella in 200g.	Negative
E. Coli in 1g.	Negative

Packaging: 20 kgs. net weight, in multi-ply, valve-pack, kraft paper bags with polyliner.

Defatted soy flours with 50-75% protein dispersibility are extensively used in bakery products. They increase the water absorption capacity of flours in bread dough and cake batters. In cakes,they improve film forming and even distribution of air cells. As a result, even cake texture and more tender crumb structure are achieved. In hard cookies, soy flour improves machining. In all these products, soy flour is used at the level of 2-5%.

More thoroughly toasted flours and grits are used to impart a pleasant nutty flavour to whole-grain and multi-grain specialty breads.

An important application of defatted soybean flour and grits in combination with cereals is in the production of nutritionally balanced all-purpose food blends, distributed to under-nourished populations or in cases of food shortage emergencies. The best known of these blends are: CMS (corn-milk-soy), developed in the U.S.A. by the Northern Regional Research Centre in cooperation with the American Corn Millers Federation, National Institute of Health and AID, CS (corn-soy) and WS (wheat-soy). More than 1.5 million tons of CMS have been distributed between 1966 and 1979. An “instant” CMS has been also developed. CMS can be used in soups, gruels, porridges etc. typically, CMS contains 17.5% defatted soy flour, 15% non-fat milk solids, about 60% corn. CS contains 22% soy flour and 71% corn. WS has 20% soy, 53% wheat bulgur and 20% wheat protein concentrate. Another well-known blend is INCAPARINA, developed by the Instituto de Nutricion de Centro America y Panama (INCAP), to fight children malnutrition. The oilseed protein source in the original formula of INCAP was cottonseed, but it has been replaced by soybean flour.

b- As a raw material for further processing: White flakes and defatted soy flour with a high protein solubility serve as the starting raw material for the manufacture of most protein concentrates, isolated soy protein and extrusion- texturized soy flour. They are also used, alone or in combination with whole soybeans , as a starting material for the production of soy sauce.

5. SOYBEAN PROTEIN CONCENTRATES (SPC)

5.1 Introduction

Edible soybean protein concentrates are relatively new products. Their availability as commercial products dates from 1959. In the last 30 years or so, these versatile products have become important ingredients, well accepted by many food industries. In many applications, they simply replace soy flours. In others, they have specific functions which cannot be performed by soy flours.

Historically, the need for the development of soybean protein concentrates stemmed primarily from two considerations: to increase protein concentration and to improve flavour.

It is very difficult to avoid the occurrence of the green-beany flavour of soybeans in untoasted full-fat or defatted soy flour, prepared in the conventional way. Beany flavour is one of the major objectionable characteristics, limiting the use of conventional soy flours. One of the objectives of the further processing of flours into concentrates is to extract the particular components which are responsible for the bitterness and beany taste.

As shown in the previous chapter, the maximum level of protein content in soy flour, even after nearly complete removal of hulls and oil, is about 55% (moisture-free basis). In certain applications, such as in meat products, a soybean protein ingredient with a higher percentage of protein is often preferable.

Soybean protein concentrates normally cost 2 to 2.5 times more than defatted soy flour. Considering the relative protein contents of these two products, the cost per unit weight of protein is about 80% higher in the concentrate.

The starting material for the production of soy protein concentrates is dehulled, defatted soybean meal with high protein solubility (white flakes). The concentration of protein is increased by removing most of the soluble non-protein constituents. These constituents are primarily soluble carbohydrates (mono, di and oligosaccharides), but also some low molecular weight nitrogenous substances and minerals. Normally, 750 kilograms of soybean protein concentrate are obtained from one metric ton of defatted soybean flakes.

There are three major methods for extracting these components in a selective manner, without solubilizing the major protein fractions. These are not different methods for manufacturing the same product, but each method produces a different type of concentrate, with distinct characteristics and specific uses. These methods are known as:

* The aqueous alcohol wash process
* The acid wash process
* Heat denaturation/water wash process

5.2 Defintion, compostion, types

The Association of American Feed Control Officials, Inc. (AAFCO), specifies soy protein concentrates as follows:

” 84.12: Soy Protein Concentrate is prepared from high-quality sound, clean, dehulled soybean seeds by removing most of the oil and water-soluble non-protein constituents and must contain not less than 70% protein on a moisture-free basis.” ( from the ’89 Soya Bluebook.)

Following is the composition of a typical food-grade soy protein concentrate ( SOLCON, made by Solbar Hatzor Ltd.) as specified by the manufacturer:

Protein (mfb) .	70% min.
Moisture	8% max.
Crude fibre	4.5% max.
Ash	7% max
Particle size	95 % < 150 microns
Fat	1% max
Standard plate count	15,000/g. max
Salmonella in 200 g.	Negative
E. Coli in 1 g	Negative

As explained above, there are three basic types of soy protein concentrates, distinguished according to the method used for extraction of the non-protein solubles. All three types have basically the following proximate composition, on a moisture-free basis:

Protein (Nx6.25)	70%
Insoluble carbohydrates	20%
Ash	5%to 8%
Lipids	1%

Soy protein concentrates are further characterized by their protein solubility index. Soy proteins are rendered insoluble by each of the three extraction processes. However, it is possible to increase the solubility of the protein in the concentrate by further processing, for example by neutralization of acid-washed concentrate with alkali. Concentrates made by heat denaturation/water leaching processes are irreversibly denatured and darker in colour. Alcohol-wash concentrate has a low NSI value (10 to 15%) due to denaturation of the protein by the aqueous alcohol. The molecular changes in the proteins caused by alcohols are, however, different from those resulting from heat denaturation. Thus, alcohol-wash concentrate retains most of the functional properties (slurry viscosity, emulsification power etc.) despite its low protein solubility as determined by the standard NSI or NDI tests.

The dispersibility and functionality of alcohol-wash concentrates can be increased by steam injection or jet-cooking and improved further by high-shear homogenization. (Soy Protein Council 1987).

Much of the characteristic beany flavour is also usually removed by the extraction process. Soybean protein concentrates are relatively bland. The flatus-producing oligosaccharides of soybean flour, raffinose and stachyose, are also efficiently removed by the solvents used in the production of concentrates.

Soy protein concentrates are marketed in various forms: granular, flour and spray dried. In addition, texturized concentrates are also available. These texturized products will be discussed later.

Since some low molecular weight proteins are also extracted along with the sugars, the amino acid composition of the concentrates may differ slightly from that of the original flour. (Table 5-1).

Table 5.1 Amino acid composition of SCP and soy flour (grams per 16g. nitrogen)

5.3 Production processes

5-3-1 The aqueous alcohol wash process

The process is based on the ability of aqueous solutions of lower aliphatic alcohols (methanol, ethanol and isopropyl alcohol) to extract the soluble sugar fraction of defatted soy flour without solubilizing its proteins. The optimal concentration of alcohol for this process is about 60% by weight.

The theory of solvent extraction (see para. 3-2-4) is applicable to the extraction of defatted soy flour with aqueous alcohol.

Starting with defatted white flakes as raw material, the process consists of the following steps: Liquid-solid extraction, removal and recovery of the solvent from the liquid extract, removal and recovery of the solvent from the extracted flakes, drying and grinding of the flakes.

a- Solid-liquid extraction: This can be carried out batchwise or continuously. Continuous extraction is justified for relatively large scale operations. According to Campbell et al.(1985), continuous processes are employed for plants with typical capacities over 5,000 tons per year. Unlike oilseed crushing industries, smaller plants are not uncommon in this branch. The batch process is, therefore, rather widely applied. The methods and types of equipment used are essentially similar to those encountered in oil extraction plants: horizontal belt and basket extractors, stationary and rotary cell extractors etc. In the case of alcohol extraction, the solvents are quite volatile and flammable. Adequate precautions for the prevention of fire and explosion are necessary.

The reason for using high-NSI white flakes as the starting material is not necessarily related to the objective of obtaining a product with high protein solubility.( As explained above, this would not help anyhow , due to the different type of protein denaturation caused by the alcohol.) The principal reason for preferring this type of raw material is due to the fact that the percentage of extractable soluble sugars in white flakes is higher than in toasted meal. Toasting renders the sugars less soluble by binding them to proteins (Maillard reaction) or by caramelization. As a result of this type of condensation reactions, the sugars are no longer extractable by the solvent and they remain in the product, lowering the protein concentration in it. Furthermore, the darker colour of concentrates made from overheated meal is also objectionable, and their nutritional value is lower (lower lysine availability.)

b- Removal and recovery of the solvent from the liquid extract: The alcohols are removed from the liquid extract by evaporation and rectified by distillation. They are then brought to the proper concentration and recycled through the extractor. The distillation residue is an aqueous solution of the sugars and other solubles. It is concentrated to the consistency of honey and sold as “soy molasses”. Typically, soy molasses contain 50% total soluble solids. These solids consist of carbohydrates (60%), proteins and other nitrogenous substances (10%), minerals (10%), fats and lipoids (20%). It is mainly used as a caloric ingredient and as a binding agent in animal feeds.

c- Desolventizing the solids: After extraction, the solvent saturated flakes are desolventized . The methods are essentially the same as for the removal of hexane from soybean meal flakes. Flash desolventizing, using superheated vapours of the alcohol-water mixture can be applied to protein concentrates. Any excess water left in the flakes after desolventizing is removed by hot air drying.

d- Grinding: The methods and equipment used to grind soy protein concentrate flakes are essentially the same as those employed in the production of soy flours (see Section 4-3-1).

5-3-2 The acid-wash process

This process is based on the pH-dependence of the solubility of soybean proteins, discussed in Section 1-6-2. It will be recalled that the majority of soybean proteins exhibit minimum solubility at pH 4.2 to 4.5 (isoelectric region). Therefore, it is possible to extract the sugars, without solubilizing the majority of the proteins, using, as a solvent, water to which an acid has been added so as to keep the pH at the isoelectric region.

The acid-wash process has the obvious advantage of using a non-flammable, non-explosive, non-toxic and inexpensive solvent: water. To a certain extent, this is also the disadvantage of the process. Separation of the solid from the solvent is more difficult and less complete, due to the fact that the flakes absorb considerable quantities of water and swell. Gravity draining is not suitable for efficient solid-extract separation. Rotary vacuum filters or decanting centrifuges must be used instead.

A batch process using horizontal decanting centrifuges is shown in Fig. 29. Defatted soy flakes or flour are mixed with acidified water in an agitated vessel. The slurry is then fed to the decanter centrifuge which separates the extracted solids from the extract (whey). The solids are discharged continuously at approximately 30% dry matter content. The solids can be dried at this stage, to yield an “isoelectric” concentrate of low protein solubility. If a more functionally active, neutral concentrate is desired, the isoelectric solid cake is resuspended in water and the acidity is neutralized. A second step of centrifugal separation gives a cake of neutral concentrate with a protein content of 75% on dry matter basis. This cake also retains about 70% water, by weight.

The cake is usually wet-milled to a fine slurry and spray dried. The protein solubility of the neutralized product is quite high, giving NSI values above 60%, provided that white flakes were used as the starting material.

The liquid extract containing sugars, minerals, the protein fractions which are soluble at pH 4.5, and other soluble components is usually known as “whey”, in analogy to the process of cheese making. Unlike cheese whey, however, soy whey has no use and must be discarded as waste. The reasons for not using soy whey for animal feeding will be discussed in the next chapter, dealing with isolated soybean protein.

5-3-3 Heat denaturation/ water extraction process

In this process, the proteins of defatted soy meal are first rendered insoluble by thermal denaturation, using humid heat. The heat-treated meal is extracted with hot water, which dissolves the sugars.

5.4 Utilization

5-4-1 Basic considerations

Just as with soy flours, soy protein concentrates are used in food products for their nutritional characteristics or for their functional properties or for both.

Nutritionally, the attractive features of concentrates include their high protein content, the near-absence of anti-tryptic and other anti-nutritional factors, the absence of flatulence and the substantial “dietary fibre” content. The nutritional value of the protein in the concentrates of different types, expressed as Protein Efficiency Ratio (PER) is slightly lower than that of soy flour protein. (Table 5-2). This is probably due to the slight fractionation effect of the extraction process, mentioned above.

Table 5.2 Per * value of soy protein products

(*) The PER values corrected to: casein = 2.5
Source: Soy Protein Council (1987)

The most important functional characteristics of soy protein concentrates are: water binding (water adsorption) capacity, fat binding capacity and emulsification properties.

5-4-2 Use in bakery products

Unless higher protein fortification levels are necessary, there is no special reason for using soy protein concentrates in bakery products. Nutritionally and functionally, soy flours do the same job, more economically.

5-4-3 Meat products

This area probably represents the most important application of soy protein concentrates in the food industry. SCP is used mostly in comminuted meat, poultry and fish products ( patties, emulsion type sausages, fish sticks etc.) to increase water ant fat retention. The nutritional contribution of soy protein in low-meat, high-fat, low-cost products may also be significant. Typical usage levels, on moisture-free basis, are: 5-10% in patties, 2-8% in chili, 2-12% in meatballs, 3.5% max. in sausages, 5-10% in fish sticks. (Campbell et al. 1985).

5-4-4 Other uses

Soybean protein concentrates have been used as stabilized dispersions in milk-like beverages and simulated dairy products such as sour cream analog. Campbell et al.(1985) present a formula for a milk-like beverage, suggested by A.E. Staley Mfg. Co., producers of the soy protein concentrate and the corn syrup solids components in the formula. The formula and directions for the preparation of the beverage are given below:

Formula for “Soy Concentrate Milk”:

Soy protein concentrate	6.0 %
Sucrose	0.6 %
Corn syrup solids	2.0 %
Fat	3.0 %
Mono-and di-glycerides	0.1 %
Salt	0.05 %
Water	88.25 %

The SCP is hydrated with water in a high-shear mixer, then all other ingredients, except the fat are added and mixed thoroughly. The mixture is heated to 65-70^oC. The fat (apparently a hydrogenated, well deodorized oil) and flavouring agents are added. The mixture is homogenized, cooled and packaged.

Non-dairy coffee whiteners can also be made, using the same principle, but different ingredients and proportions.

6. ISOLATED SOYBEAN PROTEIN (ISP)

6.1 Introduction

Isolated soybean proteins, or soybean protein isolates as they are also called, are the most concentrated form of commercially available soybean protein products. They contain over 90% protein, on a moisture free basis.

Soy protein isolates have been known and produced for industrial purposes, mainly as adhesives for the paper coating industry, well before World War II. ISP’s for food use, however, have been developed only in the early fifties.

The basic principles of ISP production are simple. Using defatted soy flour or flakes as the starting material, the protein is first solubilized in water. The solution is separated from the solid residue. Finally, the protein is precipitated from the solution, separated and dried. In the production of ISP for food use, in contrast to ISP for industrial use, care is taken to minimize chemical modification of the proteins during processing. Obviously,the sanitary requirements are also much more demanding.

Being almost pure protein, ISP can be made to be practically free of objectionable odour, flavour, colour, anti-nutritional factors and flatulence. Furthermore, the high protein concentration provides maximum formulation flexibility when ISP’s are incorporated into food products. These and other advantages have been the source of highly optimistic forecasts regarding the widespread use of ISP. Although the volume of production increased and although several production facilities have been erected in the U.S.A., Europe, Japan, India and Brazil, the tonnage figures are far from those predicted when food grade ISP was first marketed.

The principal reasons for this situation are the relatively high production cost (see below), nutritional and regulatory limitations, the inability of ISP-based texturized products to compete with texturized soy flour and texturized SPC, and finally, the competition of other abundant “isolated proteins”, particularly casein and caseinates. Nevertheless, it should be noted that many novel isolated proteins, such as those obtained from cottonseed, peanuts, fish, squid etc. have been much less successful than ISP. Many of these did not reach the stage of commercial production.

Although actual trade figures are not disclosed, the growth in sales of concentrates and isolates is said to be, at present, stronger than that of flours.

ISP can be further modified and processed into more sophisticated products. These include: spun fibres from ISP as an ingredient for muscle food analogs, proteinates and enzyme modified ISP.

The cost of isolated soybean proteins is five to seven times higher than that of defatted soy flour. On an equal protein weight basis the cost ratio of these two products is nearly 3:1. The main reasons for the added cost will become evident from the description of the manufacturing methods for ISP.

6.2 Defintion, composition, types

The specification of the Association of American Feed Control Officials, Inc. (AAFCO) defines ISP is as follows:

“Soy Protein Isolate is the major proteinaceous fraction of soybeans prepared from dehulled soybeans by removing the majority of non-protein components and must contain not less than 90% protein on a moisture-free basis.” (from ’90 Soya Bluebook).

There are no official standard definitions or specifications for the various types of isolates. ISP is bought and sold on the basis of specifications formulated by the manufacturer or the user.

The typical composition of an isolated soy protein is shown in Table 6-1.

Table 6.1 Typical composition of ISP
(Moisture-free basis)

Source: Kolar et al. (1985)

The conventional procedure for ISP production is based on protein solubilization at neutral or slightly alkaline pH, and precipitation by acidification to the isoelectric region, near pH 4.5. The resulting product is “isoelectric ISP”. It has low solubility in water and limited functional activity. Different “proteinates” can be produced by resuspending isoelectric ISP in water, neutralizing with different bases and spray-drying the resulting solution or suspension. According to the base used for neutralization sodium, potassium, ammonium or calcium “proteinates” are produced. The first three are highly soluble in water, producing solutions with very high viscosities, foaming, emulsification and gel-forming properties. Calcium proteinate has low solubility. Low-solubility (inert) ISP’s are used where the formulation calls for a high level of protein incorporation without excessive viscosity of other functional contributions.

Since spray-drying is the common drying method in the production of ISP, the primary physical form of ISP in commerce, is that of fine powders. Structured forms, such as granules, spun fibres and other fibrous forms are made by further processing. These forms will be discussed in a separate chapter, dealing with texturized products.

6.3 Production processes

6-3-1 The conventional process

This is the process commonly described in the literature and suggested by suppliers of equipment and complete plants. Exact processing conditions and the type of equipment used may vary from plant to plant.

An outline of the process is given in Fig.30.

a- Starting material: Dehulled, defatted, edible grade white flakes or meal with the highest possible protein solubility index are used. Although the rate of protein extraction from finely ground flour would be faster, flakes permit easier separation after extraction. In batch extraction, particle size has no effect on protein extraction yield, if extraction time is over 30 minutes.

b- Protein extraction: The flakes are mixed with the extraction medium in agitated, heated vessels. The extraction medium is water to which an alkali such as sodium hydroxide, lime, ammonia or tri-basic sodium phosphate has been added, so as to bring the Ph to neutral to slightly alkaline reaction. Under these conditions, the majority of the proteins go into solution. The sugars and other soluble substances are also dissolved.

* Alkalinity: More protein can be extracted at higher pH. However, the extracted proteins may undergo undesirable chemical modifications in strongly alkaline solutions. These include protein denaturation and chemical changes in amino acids. Excessively high pH also favours protein-carbohydrate interaction (Maillard reaction) which results in the formation of dark pigments and in loss of nutritive value. Furthermore, proteins precipitated from highly alkaline media tend to retain too much water, and do not settle well. In practice, the range between pH 7.5 and pH 9.0 is most commonly preferred.

One of the chemical reactions of amino acids in alkaline media has attracted particular attention. That is the destruction of the amino acid cystine, with the formation of dehydroalanine. In addition to the nutritional implications resulting from the loss of cystine, there might be also a toxicological aspect to consider. Dehydroalanine can react with free epsilon-amino groups of lysine, to produce lysinoalanine. This compound has been found to cause kidney lesions in rats under certain experimental conditions. The toxicity of lysinoalanine for man is still an open question.

* Extraction time: The course of nitrogen extraction from white flakes , using 0.03 molar calcium hydroxide as extractant is shown in Fig. 31. The amount of nitrogen extracted under these conditions increased steadily during the first 30 minutes and reached a nearly constant level after 45 minutes. The extraction time in industrial operation is, probably, in the order of 1 hour.

* Temperature: Protein extraction yield is considerably increased by raising the temperature, up to 80°C.

* Solid/liquid ratio: Protein extraction yield is improved as the quantity of liquid medium used to extract a given weight of flakes is increased. After extraction and separation by filtration or centrifugation, the extracted flakes retain a considerable proportion of extract, about 2.5 times the weight of solid. In single-stage batch extraction, if the more liquid is used for extraction, the protein concentration in the extract is lower and the quantity of protein associated with the retained portion of the extract is smaller. On the other hand, larger volumes of liquid have to be handled per unit weight of protein produced. This means larger extraction vessels, centrifuges etc. and a larger volume of “whey” for disposal.

The choice of a solid/liquid ratio for extraction is, therefore, a matter of economical optimization. The ratios used in industry range apparently between 1:10 and 1:20.

* Heat treatment history of the meal: The NSI value of the starting material is the most important factor affecting isolation yield. (Fig. 32)

* Agitation: As in any extraction operation, agitation increases the rate of protein solubilization. However, within the practical values of extraction time for batch operations (about one hour), little is gained by increasing the turbulence beyond that provided by moderate agitation. Furthermore, strong agitation causes excessive flake disintegration, increases the proportion of fine particles in the extract , rendering solid/liquid separation more difficult. Moderate agitation can be defined as any mixing operation that would keep the flakes in suspension within the extraction medium.

c- Solid-liquid separation after extraction: The extract contains considerable amount of fine particles of extracted flour, the elimination of which, prior to precipitation, is necessary in order to obtain a “curd” of acceptable purity.
Table 6-2 shows the effect of fine solids separation on the purity of the final product.

In industrial scale operation, it may prove convenient to carry out the extract clarification process in two steps: screening (vibrating screen, rotary screen or the like) to separate most of the solids, followed by centrifugal clarification of the extract. The wet solids can be pressed to remove as much entrapped extract as possible. All these operations can also be carried out in one step, using decanter centrifuges. A flow diagram of decanter-based process for the production of ISP is shown in Fig. 33.

d- Extract treatment: The clarified extract can be treated so as to remove certain impurities, thus improving the blandness, colour and nutritional quality and modifying the functional properties of the final product. Extract treatment may include: ion exchange to remove phytate and reduce the ash content, treatment with activated carbon to remove phenolic substances, ultrafiltration for concentration and removal of low molecular weight components etc. Although such processes have been suggested in the literature it is not known whether they are practised in the industrial production of ISP. The use of membrane processes for extract purification and concentration have been reported to be industrially applied in Europe and Japan. (Elias, 1979).

e- Precipitation: The protein is precipitated from the extract by bringing the pH down to the isoelectric region. The type of acid used or the temperature of precipitation do not affect the yield or purity of precipitated protein.

f- Separation and washing of the curd: The precipitated protein (curd) is separated from the supernatant (whey) by filtration or centrifugation. Desludger or decanter centrifuges can be used for this purpose. The curd must be washed in order to remove residues of whey solubles. This can be done by resuspending the curd in water and re-centrifuging, or continuously on a rotary or belt filter. Thorough washing is most important for the obtention of high purity ISP.

g- Drying: The usual method for drying the washed curd is spray-drying.

6-3-2 Problems in conventional processing

a: Process losses: The conventional process separates the soy solids into three fractions: extraction residue, curd (ISP) and whey.

Extraction residue (okara) is the insoluble solid material left behind after extraction and separated from the extract by filtration or decanting. It represents approximately 40% of the solids in the raw material and carries away 15% of the protein entering the process. It is usually pressed, dried and sold as a by-product of ISP manufacture. It can be used as a protein source for animal feeding rations or as a source of dietary fibre in human nutrition. It has been also used in food products for its exceptional water adsorbing capacity.

Whey is the liquid supernatant, after the protein is precipitated from the extract. It contains the sugars and the nitrogenous substances not precipitated by acidification.

Approximately 25% of the dry matter of the raw material and 10% of its nitrogen content is found in this fraction. Early investigations ( Hackler et al. 1963) indicated that soybean “whey” may be toxic to animals. This finding has been reconfirmed often since then. Furthermore, ISP whey is a highly diluted stream, containing 1 to 3% solids depending on the solvent:flake ratio used for extraction. Concentration and drying of ISP whey would be too costly. ISP whey is , therefore, a waste stream of the isolation process.

The curd is the precipitate obtained by acidification of the extract. After washing and drying, it becomes the final product: isoelectric ISP. It contains 75% of the protein of the starting material. Nearly 3 tons of defatted soybean are needed to produce one ton of protein isolate.

This low yield explains, to a large extent, the relatively high cost of ISP.

b: Quality: ISP obtained by the conventional process contains several types of impurities ( e.g. phytates and phenolic substances) which may somewhat impair its functional,sensory and nutritional quality. More complete dehulling of the beans , thorough extract clarification and repeated washing of the curd reduce the impurities but does not eliminate them completely.

6-3-3 Alternative processes

Several alternative processes for the isolation of soy protein have been reported in the literature. These include:

a: Solubilization of the soy proteins in the salt solutions (salting-in) followed by precipitation by dilution with water.

b: Precipitation from the extract at near-boiling temperature, using calcium salts ( as in the production of Tofu).

c: Ultrafiltration of the extract so as to remove the low molecular weight components of the whey , leaving a concentrated solution of protein which may be spray-dried.

d: Physical separation of the intact protein bodies from very finely ground soy flour by density fractionation (flotation).

e: Purification of the extract by ultrafiltration, filtration through activated carbon and ion exchange, in order to increase curd purity.

6.4 Utilization

6-4-1 Meat products

In this paragraph, only the use of non-texturized ISP and proteinates will be discussed. It should be remembered, however, that the major application of ISP in connection with meat and related product is based on the use of texturized ISP, in one form or another, to replace meat. This application will be dealt with later on.

In emulsion type sausages, such as frankfurters and bologna, ISP and proteinates are used for their moisture and fat binding properties and as emulsion stabilizers. Typical usage levels are 1% to 4% on a prehydrated basis. The use of ISP in these products permits reducing the proportion of expensive meat in the formulation, without reducing the protein content or sacrificing eating quality.

Methods for incorporating soy protein products into whole muscle meat have been developed recently. Isolated soybean protein is dispersed in specially formulated meat curing brines and injected into whole muscle using stitch pumps. It is also possible to incorporate the protein by surface application of the protein containing brine, followed by massaging or tumbling, as practised in the cured meat industry. Typical brine formulations contain salt, sugars, phosphates, nitrite and/or ascorbic acid.

6-4-2 Seafood products

The most important of application in this category is the use of ISP in fish sausage and surimi based restructured fish products in Japan. Surimi is extensively washed, minced fish flesh.

6-4-3 Cereal products

ISP is sometimes used instead of, or in combination with isolates and soy flour, in the formulation of milk replacer mixtures in bakery products. ISP has been used for protein fortification of pasta and specialty bread. In these applications, the high protein content and blandness of ISP are clear advantages.

6-4-4 Dairy-type products

Soybean protein isolates are used in non-dairy coffee whiteners, liquid whipped toppings, emulsified sour cream or cheese dressings, non dairy frozen deserts etc. The basis for these applications is, demand for non-non-dairy (all-vegetarian, cholesterol-free, allergen-free) food products, as well as economy.

Imitation cheeses have been produced from isolated soy proteins, with or without milk whey components. The types of cheeses which can be produced include soft, semi-soft, surface-cultured (imitation Camembert) and ripened hard cheeses.

6-4-5 Infant formulas

Infant formulas where milk solids have been replaced by soy products are well established commercial products. ISP is the preferred soy ingredient, because of its blandness, absence of flatus-producing sugars and negligible fibre content.The principal market for these products are lactose-intolerant babies. However, soy protein based dietetic formulas are finding increasing use in geriatric and post-operative feeding as well as in weight reduction programs.

6-4-6 Other uses

Partially hydrolysed soy proteins possess good foam stabilization properties and can be used as whipping agents in combination with egg albumen or whole eggs in confectionery products and deserts.

Isolated soybean protein has been shown to be an effective spray-drying aid in fruit purees. In this application, it can replace maltodextrins, with the advantage of contributing protein to the final product. A nutritious “shake” base was produced by spray-drying ripe banana puree containing up to 20% ISP on dry matter basis. (Mizrahi et al.,1967).

7. TEXTURED SOY PROTEIN PRODUCTS

7.1 Introduction

For many years, the newly developed soy protein products did not make much progress in occupying a central position in the global protein nutrition picture. The first processed soy protein products were mainly flours or powders which had to be “concealed” in existing foods such as bread, pasta or beverages. The objective of a great part of the research effort was to render these powders sufficiently flavourless and white, and to counteract any change in the accepted characteristics of the “host” food caused by the incorporation of soy protein products at nutritionally and economically significant levels. A breakthrough in the utilization volume occurred in the 1960s, when textured soy protein products of acceptable quality became increasingly available.

Applied to soy protein products, the terms “texturization or texturing” mean the development of a physical structure which will provide, when eaten, a sensation of eating meat. Meat “texture” is a complex concept comprising visual aspect (visible fibres), chewiness, elasticity, tenderness and juiciness. The principal physical elements of meat which create the texture complex are: the muscle fibres and the connective tissue.

A voluminous patent and research literature on vegetable protein texturization has accumulated.( See e.g.Gutcho, 1977). In fact, a meat analog based on wheat gluten was being used for institutional feeding already before the start of our century. A concept of a soy protein based chewy gel and processes for its production have been described in several patents in the late 1950s. ( e.g. Anson and Pader 1957). These inventions produced homogeneous, isotropic (unoriented, of equal structure in all directions) gels, which had only one of the elements of meat texture: chewiness. They had limited commercial success.

The more successful approaches to soy product texturization can be classified in two categories. The first approach tries to assemble a heterogeneous structure comprising a certain amount of protein fibres within a matrix of binding material. The fibres are produced by a “spinning” process, similar to that used for the production of synthetic fibres for the textile industry. The second approach converts the soy material into a hydratable, laminar, chewy mass without true fibres. Two different processes can be used to produce such a mass: thermoplastic extrusion and steam texturization.

It should be noted that the term “meat” is used here in the wide sense of “flesh food”, and includes not only red meat but also poultry, fish and seafood.

The starting material for spun fibres is isolated soybean protein. In contrast, extrusion or steam texturized soy products can be made from flour, concentrate or isolated protein.

7.2 Spun-fibre based texturization

The process for the production of textured soy products containing spun protein fibres was first described in a 1954 patent issued to Boyer. Since then many additions to and modifications of the basic concept have been suggested. The basic flow-diagram of the process is shown in Fig. 34.

The first part of the flow diagram describes the steps for the production of isoelectric isolated soybean protein. These steps can be omitted if commercial ISP is used as the starting material. A concentrated protein solution is prepared by adding alkali to the ISP slurry. The solution , containing approximately 20% protein at pH 12 to pH 13 is “aged” ( to permit unfolding of the protein molecules) until its viscosity rises to the consistency of honey (50,000 to 100,000 centipoise).This viscous concentrated protein solution is technically known as “dope“.

The next step is the transformation of the dope into distinct, stretched fibres (spinning) by coagulating fine jets of the solution in an acid bath.The “dope” is pumped into the coagulating bath through a spinneret, which is a plate with thousands of fine holes (about 75 microns in diameter). The bath contains a solution of phosphoric acid and salt, maintained at pH of about 2.5. As the jet of “dope” contacts the acid medium, the oriented protein molecules are suddenly coagulated and form a fibre. The fibres are picked up as a “tow” and stretched to enhance molecular orientation and increase fibre strength. Stretching reduces the diameter of the fibre well below that of the holes on the spinneret.

The tows of fibre pass through a step of washing, to remove excess acidity and salt. The subsequent operations depend on the final product. Soy protein fibres are only one ingredient of the meat-like structure. The other ingredients include fat, binders, colouring and flavouring additives etc. The nature of these ingredients, the proportion of fibres and their orientation in the binder matrix depend on the type of flesh food to be imitated. The binder matrix contains heat-coagulable components, commonly egg albumen and the final structure is usually stabilized by thermal setting.

Spun fibre-based textured soy products have been used as “total” meat analogues (i.e. to replace meat totally) and as meat extenders (i.e. to replace part of the meat in ground meat, patties etc.) Some of the products have been used in institutional feeding (hospitals) and in school lunch programs.

The main shortcoming of spun fibre type texturized products is their cost. In the first place, the process requires an expensive starting material: isolated soybean protein. Furthermore,t he process in itself is also costly, both in initial capital investment and in running expenses.

Today, there are very few producers of spun soy protein fibres and textured products containing them. The most successful spun fibre based meat analog has been the imitation bacon chip. This is a shelf-stable low-moisture product with the bite, chewiness and flavour of fried or roasted bacon bits and is used extensively in salads, snacks and garnishes. At present, however, this product too faces the competition of imitation bacon made by the less expensive extrusion texturization technique.

7.3 Extrusion texturization

Extrusion has been long used as a central unit operation in the plastic polymer industry. Their use for continuous pressure-cooking of flours and particulate feed materials has been advocated in the 1950s. A decade later, Mc.Anelly (1964) described a process for the production of spongy, elastic particles from soy flour. A mixture of defatted flour and water was extruded through a food grinder. The extruded strands were heat-set in an autoclave, chopped, leached with hot water and dried. Although this invention can be considered as the forerunner of the extrusion texturization processes, the breakthrough in this field was the disclosure of a continuous cooking-extrusion process, for which a patent was awarded to Atkinson in 1970. In this process, defatted soy flour containing a certain amount of water is passed through a high-pressure extruder-cooker to produce an expanded, porous, somewhat oriented structure described as “pleximellar”. Although devoid of true fibres, the product possessed the textural characteristics of chewiness and elasticity, and was deemed to imitate meat in this respect. Extrusion texturized soy flour soon became an established food ingredient known as TVP ( Textured Vegetable Protein ) or TSP (Textured Soy Protein).

The extruder consists basically of a sturdy screw or worm rotating inside a cylindrical barrel (Fig. 35). The barrel can be smooth or grooved. The screw configuration is such that the free volume delimited by one screw flight and the inside surface of the barrel decreases gradually as one goes from one end of the screw shaft to the other.

As a result of this configuration, the material is compressed as it is conveyed forward by the rotating screw. Screws having different compression ratios are used for different applications. The barrel is usually equipped with a number of sections of steam heated jackets or induction heating elements or cooling jackets. A narrow orifice or “die” is fitted at the exit end of the barrel. The shape of the die opening determines the shape of the extruded product.

Defatted soy flour with a high protein solubility index is first conditioned with live steam, before entering the extruder proper. Well controlled conditioning is essential for good texturization and product uniformity. The moisture content of the feed is very important. A moisture level of about 20-25% is used for texturization. The conditioned flour usually assumes the form of small spheres.

The flour-water mixture is next fed into the extruder and picked up by the screw. As it advances along the barrel, it is rapidly heated by the action of friction as well as the energy supplied by the heating elements around the barrel. The high pressures attained through the comression mechanism explained above permits heating to 150-180°C. This rapid “pressure cooking” process transforms the mass into a thermoplastic “melt”, hence the name of “themoplastic extrusion” by which the process isalso known. The directional shear forces causes some alignment of the high molecular weight component while the proteins undergo extensive heat denaturation. The sudden release of pressure causes instant evaporation of some of the water and “puffing”. The result is a porous, laminar structure. Puffing and therefore porosity can be controlled by monitoring melt temperature at the die. If a dense product is desired, the melt is cooled at the final section of the barrel, just before entering the die.

The extrudate is cut continuously by a rotating knife as it emerges from the die. It may be dried and sold as a shelf-stable product, or it can be hydrated, flavoured, mixed with other ingredients, shaped and marketed, usually, as a frozen food.

While texturizing the soy material, extrusion cooking also provides the heat treatment necessary to reduce the microbial load and to inactivate the trypsin inhibitor. It should be noted that, despite the high temperatures in the extruder, trypsin inhibitor inactivation may be incomplete, due to the relatively short processing time.

The so-called low-cost extruders which have been mentioned in connection with the continuous heat treatment of full fat soy flour or corn-soy-milk (CSM) food supplements are not suitable for texturization. These extruders work with low-moisture feeds and provide heat mainly by friction. The extrusion-cooking machines used for texturization are more sophisticated and expensive. Recently, double-screw food extruders have been replacing the older single-screw models in food processing applications. In double-screw extruders a considerable part of the mixing and friction-heating effect takes place between the screws. The shafts can be fitted with interchangeable screw elements, providing different processing profiles along the extruder.

Extrusion texturized soy flour has been called “the first generation TVP”. Being made of flour, it has the composition and flavour of heat treated soy flour. The flavour is intensified by retorting. It contains the sugars of soy flour and presents the problem of flatulence. Usage directions usually prescribe a reconstitution step of soaking in water and pressing to remove the soluble components. More recently, processes have been developed for the texturization of soy protein concentrates. Textured concentrates (second generation TVPs) are now widely available.

Table 7-1 compares the characteristics of texturized soy products, according to the starting materials from which they are made.

Since nothing is removed or added in extrusion texturization, the composition of texturized products, on a dry matter basis, is essentially the same as that of the starting material. Shelf-stable dry products are usually marketed at a moisture level of 8%. Texturized soy products made from concentrate do not need to be leached and can be used directly, after proper hydration.

7.4 Steam texturization

Several processes have been described in the patent literature for texturizing soy protein by thermal coagulation coupled with some form of shear induced orientation to provide a fibrous-like structure. In one of these processes, patented by Stromer and Beck (1973), moistened soy flour is fed continuously into a pressurized reactor where it meets high pressure steam (at about 7-8 atmospheres). The thick mass flows, under the action of pressure, through a cylindrical barrel the discharge end of which is open to the atmosphere. The process was sold to one of the leading manufacturers of soy protein products and commercially applied for some time. According to Snyder and Kwon (1987), it is no longer being used.

7.5 Utilization

7-5-1 Meat extenders

The principal use of texturized soy protein products is as a meat extender in comminuted meat product such as patties, fillings, meat sauces, meatballs etc. In such products, as much as 30% of the meat can be replaced by hydrated texturized soy products without loss of eating quality. The cost of textured soy flour is approximately 0.60 U.S. Dollars per kilogram. About 3.5 kilograms of hydrated base is obtained from each kilogram of textured flour. Thus, the cost of meat replacement is only 17 cents of a dollar for each kilogram of meat saved. Furthermore, textured soy products offer not only economic savings but also certain types of product improvement. Their ability to absorb water and fat results in increased product juiciness and permits the use of meat with higher fat content.

Ground beef extended with TVP has been used extensively in school lunch programs, with good results.

The property of TVP to withstand cooking in a retort (retortability, retort stability) is relevant to its use in canned luncheon meat, meat loaf and similar products.

7-5-2 Meat analogs

Chunks of extrusion texturized soy protein products and spun fibre based preparations are marketed as “imitation meat” or “meat analogs”. The market for these products was, at first, limited to the relatively small sector of vegetarians. Recently there is a marked trend to reduce the consumption of red meat, associated with the demand for low-cholesterol foods. At the same time, the industry has been successful in developing more attractive meat analogs made from rehydrated textured soy proteins, alone or in combination with wheat gluten. These products are marketed as flavoured, fully prepared, frozen ready-to eat entrées. The present marketing strategy for meat analogs is to present them to the public as new, high quality products, and not as inexpensive substitutes for meat. So far, this strategy seems to be successful. The market for these sophisticated (and by no means inexpensive) products is rapidly expanding, particularly in Western Europe.

7-5-3 Other applications

Imitation bacon bits based on texturized soy protein products have been mentioned earlier. The price range for this product is 1.50 to 2 U.S. Dollars per kilogram.

A pasta product containing texturized soy protein granules is being offered on the retail market, in addition to its use in institutional feeding.

Reference

Selected extracts from TECHNOLOGY OF PRODUCTION OF EDIBLE FLOURS AND PROTEIN PRODUCTS FROM SOYBEANS by Zeki Berk, Technion, Israel Institute of Technology, Haifa, Israel, FAO AGRICULTURAL SERVICES BULLETIN No. 97, Food and Agriculture Organization of the United Nations Rome 1992, M-81, ISBN 92-5-103118-5