M. PEISKER
ADM Specialty Ingredients (Europe)
The Netherlands
Most of the technological processes in compound feed manufacturing impact on nutritive value and hygienic status of the resulting feed and sometimes act synergistically, (e.g. expansion plus pelleting on salmonella decontamination). The nutritive effect is mainly exerted on feed digestibility and efficiency of energy utilisation. The return on energy invested in feed processing (electrical and steam energy), is positive for the majority of processes when balancing against the gain in available feed energy.
In terms of dietary protein (amino acids) and feed additives like vitamins, enzymes and probiotics, special attention must be paid when applying thermal processes. Only a feed/purpose specific selection of treatment processes and fine-tuned process parameter setting brings optimal results in a scenario with sometimes conflicting objectives: maximum improvement of nutritive value combined with optimal decontamination of pathogenic germs. Whereas an improved nutritional status of the processed feed is durable, precautionary measures must be taken to maintain its hygienic status.
Introduction
Domestication and breeding of animals has improved animal performance in an unprecedented manner. The requirements in terms of nutrients have increased accordingly. Feedstuffs as offered by nature cannot meet these requirements, even if they are offered in a variety, without further processing. The coverage of dietary energy needs poses a particular challenge in feed formulation and subsequent manufacture. In the process of making dietary nutrients available for metabolic purposes the digestion process is a major influencing factor.
Digestive capacities of animals have remained largely unaffected despite centuries of breeding and selection (Wenk, 1982). Therefore, the preparation of feedstuffs before ingestion to improve digestibility became a major focus in nutrition research. On the other hand, the production units for livestock have become larger driven by economy of scale. This has heightened the importance of hygienic regimes including feed hygiene to prevent disease outbreaks and securing hygiene and feed ingredients of animal origin.
This paper highlights some key processing steps in compound feed processing and their impact on nutritive value and hygienic status. Table 1 gives an overview of the major technological processes for making poultry feed. It can be stated that short and long conditioning have very little effect on nutritive value. Grinding and crumbling have no hygienic effect. Conditioning with subsequent pelleting has some and expansion plus pelleting has pronounced effects on both nutritive value and hygienic status. Toasting is reserved for the treatment of single ingredients like soybeans, peas, beans, canola and others. It is used to eliminate or reduce the content of anti-nutritional factors in these ingredients, before they can be used in compound feed. From that angle this process is very significant in terms of improving the nutritional value of these ingredients, which otherwise could be used only with limitations. The toasting process is not primarily applied for microbial decontamination, however, the processing conditions would allow for this.
Grinding
Poultry have a short digestive tract and therefore the digestibility must be high; but is capable of grinding entire grain in the gizzard. Nir and Hillel (1994) found a correlation between particle size and weight of gizzard and duodenum in 21days-old broilers (Table 2). Also the gizzard pH-value is significantly lower when the feed particles are coarser. Feeding partly entire grains thus is seen as a possibility to prevent intestinal infections in poultry. Kamphues et al. (2005) made similar observations in weaning piglets when testing the effect of potassium diformate on Salmonella in faeces. The number of animals excreting Salmonella and the duration of excretion was significantly lower, when the feed was coarse ground (control diet: 32% > 1.0 mm; 26% < 0.4 mm; test diet: 58% > 1.0 mm; 10% < 0.4 mm). The authors concluded that Salmonella shedding is related to feed structure.
In broiler feeding, pelleted feed is the preferred choice whereas in layers its mash feed. Finely ground mash increases the time for feed intake and reduces feather picking (Walser, 1997). Particle size structure should be uniform to prevent nutrient imbalances by selective feed intake. An optimal mash structure can be achieved with a combination of expansion and crumbling process. It also must be mentioned that the level of fineness of mash impacts on subsequent processes like pelleting (throughput, pellet quality). Grinding of feed ingredients is a pre¬requisite of mixing different ingredients and achieving a low coefficient and variation percentage of nutrients in the feed mash. 
Conditioning and pelleting
Before pelleting, mash feed needs a certain degree of steam conditioning. Figure 1 shows different options for conditioning prior to pelleting. All processes require the use of a short-term conditioner for steam and water addition. Pelleting agglomerates smaller feed particles with the help of mechanical pressure, moisture and heat to larger particles. This tends to improve animal performance due to less feed wastage, no selective feeding, and improved palatability and starch gelatinization. Numerous trials have shown better daily gain and feed conversion broiler feeds.
Average daily gain and feed conversion ratio is improved by 5-8% and 3-5% respectively, when feeding a pelleted diet with zero fines versus a mash diet. Pelleting has an effect on the hygienic status of the feed; however, a short-term conditioner alone, or in connection with a pelleting press, will not be sufficient for decontamination of pathogenic microorganisms. Already Friedrich (1979), Hacking et al. (1978) and Pietzsch (1985) have stated that pelleting is not sufficient for safe decontamination as relatively large numbers of microorganisms remain in the finished feed. Sufficient decontamination with pelleting as the final processing step is achieved, when using a hygienizer with horizontal retention screw or vertical shaft (similar to long term conditioner or ripener), assuring minimum retention times of three minutes. It is suggested to increase retention time at the technically highest possible moisture level; however, limits apply for the proper functioning of the pellet press (< 16% moisture). 
Expansion
Expansion of mash feed is an established thermal processing technology that is widely used for broiler diets to improve pellet quality, include higher liquid and fat levels and increase operational productivity by increasing pelleting line capacity, and to enhance the flexibility of ingredients usage and animal performance (Wilson et al., 1998). This technology simultaneously leads to achieve a hygienic status, described below as “commercial sterility”. 
Table 3. Influence of different treatments of broiler diets on performance
Table 3 shows results from broiler trials comparing different treatments of broiler feed. Most studies document an improvement in body weight gain and feed conversion ratio; however, depending on the selection of ingredients, enzyme supplementation to expanded diets proves to be obligatory to achieve this. Gauer (2002) reported no differences in FCR, when energy level in a corn-soy broiler diet (ME 3200 kcal) was decreased by 3.1% (Figure 2). 
Separate treatment of single ingredients can prove beneficial for some but not all ingredients (Table 4).
Table 5 shows a significant effect of expanding on the digestibility of the dietary fat fraction, leading to an increase in metabolizable energy. The expansion of the corn alone did not improve digestibility or energy value of the diet. This supports the results in Table 4. 
The results show that expansion is not acting on all ingredients in the same manner. In corn-soy diets an increase in dietary energy improves feed conversion ratio. In wheat or barley based diets, the addition of enzymes is recommended to reduce the gut viscosity that is inevitably increased by expansion of these ingredients. Only then a positive effect on body weight gain and feed efficiency is achieved.
Energy return
Table 6 shows ranges of expenditure for steam and electrical energy in major feed manufacturing processes. The values are dependent on selected feed ingredients; their moisture content and the technical design and wear and tear impact on equipment. The expenditure of technical energy is balanced against the gain in feed energy (ME) in Table 7.
The increase in metabolisable energy for pelleting and expanding and pelleting has been measured in several broiler trials. In pelleting, energy expenditure (57 kWh/t) and gain in ME (56 kWh/t) are about the same. Thus for pelleting the technical energy expenditure is recovered by the increase in metabolisable energy of 1 – 1.5%. In expanding and expanding plus pelleting a net gain in feed energy versus technical energy is observed. About 73 kWh/t are expended in return for 122 kWh/t extra feed energy. As trials have shown, the ME increase in corn-soy-based broiler diets can exceed 4% (Table 5). 
These calculations consider only the nutritional improvement by enhancing the digestibility at equal feed intakes. Better economic feed conversion ratios by reduced feed losses (dust, fines, uneaten feed) will further improve the value of these technological processes. A total energy balance should include energy for transportation and storage processes of feed. High-density feeds resulting from pelleting after expansion are advantageous in terms of transport and storage space and feed intake in broilers. 
Hygienic feed preparation
a) Decontamination
Sterility cannot be achieved in feedstuffs; however, a so-called "commercial sterility" (Asquith, 2002) is possible. This means that pathogenic microorganisms have been eliminated (coliform bacteria, Salmonella, moulds, etc.). A higher, but still incomplete level of sterility can be reached at temperatures of 130°C, pressures of 3 bar and treatment times of 20 minutes, as practiced when sterilizing meat-meal (autoclaving). This technology is reserved for special applications because valuable components, such as amino acids and vitamins, are damaged at these temperature and pressure conditions.
Appropriate processing technology depends on proper judgement of how these organisms live, grow and die.
Salmonella, for example, have the highest growth rate at a temperature of 35-38°C. But, they can also grow at temperatures between 5-50°C if the ambient conditions are optimal. This depends mainly on the moisture content. Salmonella can only reproduce at an aw-value (available water) of >0.92, which is not found in normal feedstuffs with ~13% moisture as it requires a moisture level of >25%. The aw -value is the part of the water in feedstuffs, which is not bound to other substances but is completely available to microorganisms. This value varies between 0 (anhydrous substance) and 1 (pure water). It indicates the equilibrium moisture, adjusted between the sample and the relative air humidity. Salmonella and coliform bacteria need an aw value of >0.92 for growth, while moulds need >0.8. The aw value of cereals with 17% moisture is ~0.8. To avoid spoilage, the cereals needs drying to <16% moisture. The temperature range allowing growth of Salmonella, other bacteria and moulds is in the range of 5-55°C. When heating the product to more than 60°C, microorganisms stop growing and die. They do not die abruptly, but according to a logarithmic function.
Apart from aw values and temperatures, the mortality rate also depends on the pH value.
Salmonella and other bacteria have good growth conditions in the pH range of 7.0 to 8.5, moulds from 5.0 to 7.0. The reduction of pH value by adding organic acids can be used to decontaminate feeds. For complete decontamination, the addition of about 2-4% of organic acids is required. The costs involved are much higher when compared with thermal treatment; also, special attention is needed for the selection of acid-resistant equipment. In dry feed at ambient temperatures, Salmonella cannot reproduce, but they do survive by downscaling their energy metabolism. When the aw -value increases, the energy metabolism rekindles. Activated, moist salmonella can be eliminated much easier than non-activated Salmonella. For this reason, hydrothermal treatment should always be coupled with an increase in moisture.
Salmonella are not uniformly distributed in the feed, but are found in spots throughout the mixture. Therefore, it is difficult to "locate" them and to account for them in samples. For this reason coliform bacteria counts are used as a measure for the decontamination effect of different treatments since they exhibit the same heat resistance as Salmonella and are ubiquitously and uniformly distributed in feedstuffs. Some organisms, such as spores of aerobic bacteria and anaerobic sporoform bacteria (Clostridium perfringens), cannot be eliminated readily by thermal processes. The same applies for toxins formed by moulds. Figure 3 displays the correlation between temperature and time for Salmonella decontamination at normal product moisture in meat meals (Beumer, 1992). Heidenreich (2002) showed that moisture content of about 14% in expanding at 105 °C is needed for a significant reduction in total germ counts (Figure 4). Israelsen et al. (1996) and König (1994) aimed at similar conclusions (Figure 5). In expander processing prior to pelleting the expander peak temperature must clearly exceed 100°C to result in total elimination of Salmonella. In expanders and extruders, next to temperature and moisture, the sudden pressure drop at the outlet is a key element in killing bacterial cells, causing the living cells to burst (Peisker, 1991). Without pelleting, the expander temperature must reach 110°C for total Salmonella elimination. 
b) Prevention of recontamination
Recontamination is a true hazard in feed production. A multitude of steps must be taken to secure the hygienic status achieved by feed processing. The addition of organic acids, such as propionic or formic acid (~ 0.5%) is generally accepted to protect a clean feed mixture after hygienic treatment against recontamination. As a consequence, the design of the cooling area for pellets, the finished product sector, the bulk transport to the farm and the storage of the feed in the farm silos must be taken into account for maintaining hygienic status of finished compound feeds. Figure 6 shows a typical “recontamination pattern” from the feed plant to the farm silo when no chemical stabilisers such as organic acids are added. Moulds were completely eliminated and total germ count considerably reduced after thermal treatment. The weak spot at the feed plant level are the ducts from the pellet press to the cooler and the cooler itself. Such equipment in particular should be addressed in HACCP¬ programs. Also unloading pits, dust filters, elevator legs, bins and trucks must be checked on regular basis. In the farm silo the feed maybe completely re-contaminated due to feed residues and insufficient cleaning before recharge.
Fully integrated poultry companies can address this and establish a successful Salmonella control system, comprising raw materials, feed processing (broiler, parent stock and layer feed), transport systems and farms. For example, the German poultry integrator “Wiesenhof” has adopted a policy enabling them to offer guaranteed Salmonella free hatched chicken, broiler meat, table and breeding eggs (Gill, 2002). However, in the commercial feed industry sector similar policies can be agreed upon contractually and several “tracking and tracing” systems have been developed by private or governmental entities and are quickly gaining importance in the industry. 
References
Armstrong, H. (1993). Feed Mix, 1: 24-27
Asquith, R. (2002). Proceedings 5th International KAHL Symposium, Reinbek, Germany
Beumer, H. (1992). Mühle + Mischfuttertechnik, 129: 639-645
Friedrich, W. (1979). German Veterinary Weekly Review, 86: 399-405
Gauer, R.V. (2002). Proceedings 5th International KAHL Symposium, Reinbek, Germany
Gill, C. (2002). Feed International, 09: 30-32
Hacking, W.C., Mitchell, W.R. and Carlson, H.C. (1978). Canadian Journal of Comparative Medicine, 42: 392-399
Heidenreich, E. (2002). Proceedings 5th International KAHL Symposium, Reinbek, Germany
Israelsen, M., Busk, J., Virsöe, M. and Hansen I.D. (1996). Feed Magazine, 12:589-595
Jackson, D.A. and Bollengier, S. (1994). Proceedings 3rd International KAHL Symposium, Reinbek, Germany
Kamphues, J., Papenbrock, S., Bruening, G., Amtsberg, G. And Verspohl, J. (2005). Proceedings of Society of Nutrition Physiology, 14: 68
König, H.G. (1994). Dissertation, Freie Universitaet Berlin, Veterinary Faculty
Liebert, F. (2002). Proceedings 5th International KAHL Symposium, Reinbek, Germany
Nir, I. and Hillel, R. (1994). Poultry Science,73:781-791
Nissinen, V.J., Peisker, M. and Liebert, F. (1993). Feed Magazine, 9:364-367
Peisker, M. (1991). Die Mühle und Mischfuttertechnik, 128: 315-319
Pietzsch, O. (1985). Central Library of Bacteriology and Hygiene, 180: 282-298
Smith, P.A., Friman, J.D. and Dale, N.M. (1995). Poultry Science, 74 Suppl. 1: 145
Walser, P.T. (1997). Dissertation, Eidgenoessische Technische Hochschule Zurich
Wenk, C. (1982). Proceedings Seminar Haustiergenetik, Ed. Geldermann and Stranzinger
Wilson, K.J.,Beyer, R.S., Froetschner, J.R. and Behnke, K.C. (1998). Poultry Science, 77 (Suppl. 1):41
From Proceedings of the “18th Australian Poultry Science Symposium”, New South Wales, Australia.
