Phosphorus Concentrations and Flow in Maize Wet-Milling Streams

Phosphorus Concentrations and Flow in Maize Wet-Milling Streams Kent D. Rausch, 1,2 Lutgarde M. Raskin, 3 Ronald L. Belyea, 4 Roderick M. Agbisit, 1 Becky J. Daugherty, 3 Thomas E. Clevenger, 5 and M. E. Tumbleson 1 ABSTRACT Cereal Chem. 82(4):431 435 Marketing of coproducts such as corn gluten meal (CGM) and corn gluten feed (CGF) is important to the maize wet-milling industry. High phosphorus concentrations could lead to limited markets for CGF due to its potential to increase phosphorus in animal wastes. The objective was to measure the concentration and flow of phosphorus in the wet-milling process and identify streams that could be altered. Samples were taken from 21 process streams of three facilities and the phosphorus content of each was determined. Flow of phosphorus was simulated using a computer model for a 2,700 tonne/day (105,000 bu/day) wet-milling plant. Phosphorus concentrations of streams varied from <10 mg/kg to >14,000 mg/kg. Phosphorus content of many streams differed significantly among facilities. Flow of phosphorus (kg/day) varied dramatically among streams. However light steepwater, light gluten, and process water streams (5,960, 3,080, and 970 kg/day, respectively) accounted for much of the phosphorus flow. Modification of these streams could reduce phosphorus content of coproducts. The high phosphorus content of either CGF or CGM could be reduced markedly if phosphorus was reduced in the appropriate streams. Wet milling is an important maize processing technology; it accounts for the processing of 51 million tonnes of maize or 22% of the U.S. crop annually (ERS 2003). In wet milling, the maize kernel is steeped and fractionated; starch, fiber, protein, and oil are separated and concentrated in specific process streams. Two major coproducts are generated (corn gluten feed [CGF] and corn gluten meal [CGM]). CGF is produced from mixing heavy steepwater with maize fiber (Fig. 1); it has high fiber and protein content and is fed mainly to ruminant animals. CGM is produced from dewatering of gluten; it contains high protein concentrations and is fed mainly to nonruminant animals. Income from marketing of CGF and CGM is important to the economic viability of the wet-milling industry because it partially offsets production costs. Factors that affect quality or marketability can have marked effects on value of these coproducts. One emerging issue is the phosphorus concentration of CGF ( 6 g/kg), which is high relative to ruminant requirements ( 3 g/kg). This could affect marketing in the near future. When ruminants consume diets containing elevated concentrations of phosphorus, the amount of phosphorus excreted in wastes is increased (Morse et al 1992). This is a concern because environmental regulations for land application of animal wastes are based partly on phosphorus concentration and are becoming more restrictive. Animal producers who access land with soils with high phosphorus concentrations or who have limited cropland for waste disposal may have to minimize use of high phosphorus feed ingredients such as CGF (Tamminga 1992; Van Horn et al 1996; Dou et al 2001; Rotz et al 2002; Spears et al 2003). A substantial (i.e., 50%) reduction in phosphorus concentration would lessen the potential environmental implications of using CGF in animal diets. To reduce phosphorus content, it is necessary to know the concentration and flow of phosphorus in wetmilling process streams. This information can be used to identify which processing streams have high phosphorus loads and potential strategies for phosphorus reduction. The objective was to determine concentrations and flows of phosphorus in wet-milling process streams. 1 Assistant professor, former graduate assistant, and professor, respectively, Department of Agricultural and Biological Engineering, University of Illinois at Urbana- Champaign, 1304 W. Pennsylvania Ave., Urbana, IL 61801, USA. 2 Corresponding author. Phone: 217-265-0697. E-mail: krausch@uiuc.edu 3 Former graduate assistant and associate professor, respectively, Department of Civil and Environmental Engineering, University of Illinois at Urbana-Champaign. 4 Professor, Department of Animal Sciences, University of Missouri, Columbia. 5 Professor, Department of Civil Engineering, University of Missouri, Columbia. DOI: 10.1094/ CC-82-0431 2005 AACC International, Inc. MATERIALS AND METHODS Three maize wet-milling plants located in the midwestern United States collaborated in the study by providing samples and information regarding processing streams. All plants used regular dent maize obtained from commodity markets. Samples were obtained from 21 processing streams in each plant. These included maize, process water, process water after sulfur dioxide (SO 2 ) addition, light steepwater, heavy steepwater, steepwater condensate, steeped maize, wet germ, wet fiber, pressed germ, pressed fiber, light gluten, heavy gluten, gluten cake, starch slurry, dry germ, CGF, CGM, wastewater, fresh water, and final effluent (Fig. 1). Maize was sampled as it entered the steep tanks. Samples of process water and process water with SO 2 were taken before and after addition of SO 2 and before addition to the steep tanks. Light steepwater was obtained from sample ports on lines leading from the steep tanks. Heavy steepwater and steepwater condensate were obtained from ports leading from the evaporators used to concentrate steepwater. Steeped maize samples were obtained from sample points located just before the first grind mills. Wet germ and wet fiber were taken from dewatering screens before pressing of germ and fiber. Pressed germ and pressed fiber samples were obtained from respective sampling ports located just downstream from the dewatering presses. Light gluten and heavy gluten were sampled near gluten-thickener centrifuges; gluten cake was sampled directly from vacuum belt filters. Starch slurry samples were taken downstream from the starch-washing hydrocyclones. Dried germ, CGF, and CGM were sampled at ports just downstream from the respective drying operations. The wastewater sample was untreated raw waste sampled at a port before the waste treatment system. Samples were taken during four sampling periods for one plant and during three periods for the other two plants. During each sampling period, three sets of samples were obtained. Each sample set consisted of 21 samples; three sample sets were taken over a seven-day period and at least 24 hr apart. Samples were taken only when operating conditions were considered to be stable. As soon as samples were obtained from process lines, they were placed in sealed containers, frozen, and shipped on dry ice to the University of Illinois for analyses. Phosphorus contents were determined using Approved Method 40-56 (AACC International 2000). Flow rates of process streams (kg/day) were estimated for a 2,700 tonne per day (105,000 bushel/day) wet-milling plant using values found in the literature (Blanchard 1999) and a simulation model developed by the University of Illinois and the Eastern Regional Research Center (USDA-ARS, Wyndmoor, PA, USA). For this simulation, all process flows were held constant and wastewater and final effluent flow rates were assumed to be equal Vol. 82, No. 4, 2005 431

to the steepwater condensate flow rates, a simplifying assumption for purposes of comparing the amount of phosphorus discharged as waste from the wet-milling process. Other waste streams joined the steepwater condensate stream but phosphorus concentrations were not measured. Phosphorus flows for each plant were estimated from the model using the phosphorus concentrations of streams for each plant. The data were analyzed for effects of plant, period, and plant-by-period interactions using a general linear model (SAS Institute, Cary, NC, USA). Means were separated when main effects were significant using the least square means procedure. RESULTS AND DISCUSSION There were significant effects of plant on phosphorus content of most samples but there were few significant period or periodby-plant interactions. Therefore, discussion will be limited to plant effects. Phosphorus concentrations of the streams varied markedly among plants and samples (Table I). Five streams had low phosphorus (<10 mg/kg) concentrations (steepwater condensate, starch slurry, fresh water, wastewater, and final effluent). Some of this variability could be due to variation in the total solids content TABLE I Phosphorus Concentrations of Samples from Maize Wet-Milling Plants (mg/kg) a Process Stream Plant A Plant B Plant C Standard Error Maize 2,186 2,260 2,034 213 Process water 589a 302b 154c 39.5 Process water + SO 2 389a 325a 235b 32.8 Steeped maize 660a 914b 654a 68.6 Light steepwater 3,831a 5,242b 3,342a 295 Heavy steepwater 15,636 14,594 13,553 818 Steepwater condensate 5 4 5 Wet germ 915a 1,154b 710c 88.4 Pressed germ 1,320 1,393 1,200 145 Dry germ 3,135a 2,673b 2,057c 300 Wet fiber 125a 314b 175a 26.8 Pressed fiber 311a 397b 254ac 26.2 Corn gluten feed 6,065a 5,179b 4,958c 589 Light gluten 458a 731b 513a 49.7 Heavy gluten 767a 1,372b 1,028c 114 Gluten cake 1,803a 2,160b 1,577c 233 Corn gluten meal 2,951a 7,185b 2,500c 404 Starch slurry bdl bdl bdl Fresh water bdl bdl bdl Wastewater 4a 6ab 9b 1.4 Final effluent 6a 4b 3b 0.7 a Values followed by the same letter in the same row are not significantly different (P < 0.01). Below detectable limits (bdl); does not apply ( ). Fig. 1. Maize wet-milling process showing sample locations. 432 CEREAL CHEMISTRY

of individual streams. Eight streams had phosphorus concentrations of 100 1,000 mg/kg (process water, steepwater condensate, wet germ, wet fiber, pressed fiber, light gluten, and heavy gluten). The remaining streams (maize, light and heavy steepwater, pressed germ, dried germ, CGF, gluten cake, and CGM) had phosphorus concentrations of 1,000 15,600 mg/kg. The phosphorus concentration of maize ( 2,000 to 2,300 mg of P/kg) was not different among plants (Table I). Phosphorus content of process water decreased significantly from Plant A to Plant B to Plant C (589, 302, and 154 mg of P/kg, respectively). Process water with SO 2 had similar phosphorus concentrations for Plants A and B, and both were greater than for Plant C. The phosphorus concentrations of steeped maize and light steepwater were higher for Plant B than for Plants A and C. However, phosphorus concentration of heavy steepwater was not different among plants ( 13,500 to 15,600 mg of P/kg). The differences among plants in phosphorus concentrations of process water, steeped maize, and light steepwater, when there were no differences in the initial feedstock (maize), suggests that plants differed in the amount of process water or phosphorus concentration of process water recycled and used in steeping. Lack of differences among plants in phosphorus concentration of heavy steepwater despite differences in the phosphorus content of the parent stream (light steepwater) indicated that water was removed to a similar end point (probably because of the difficulty of removing water from heavy steepwater), resulting in similar phosphorus concentrations. Plant B had the highest concentration of phosphorus in wet germ, dried germ, wet fiber, and pressed fiber. However, Plant A had the highest phosphorus concentration in CGF. Corn gluten feed is produced by combining two streams, heavy steepwater and pressed fiber (Fig. 1). Precise control over the proportions of the two streams is difficult to achieve consistently, and the proportions can vary to achieve the desired final protein content in CGF (Rausch et al 2003). This could account for at least some of the variation among plants in the phosphorus concentration of CGF samples. In addition, there was variation among plants in the phosphorus concentrations of heavy steepwater and pressed fiber (Table I), providing an additional source of variation. Plant B had the highest phosphorus concentration in light gluten, heavy gluten, gluten cake, and CGM. This may be because 1) the phosphorus concentration of steeped maize was higher, or 2) this plant did not remove as much water (and phosphorus) at the starch separator, gluten thickener, or gluten filter. To better understand the quantities of phosphorus in a process plant and identify potential areas for improved process stream characteristics, the total quantity of phosphorus is needed. When flow rates of streams and concentration data are available, the quantity of phosphorus carried in a stream (flow data) can be estimated (flow rate concentration). Flow rate data are very useful because they provide estimates of the absolute quantities carried in process streams (Fig. 2). Flow data can be used also to indicate the importance of each stream in the phosphorus balance of processing plants and can be used to identify important control points for potential manipulation of the phosphorus content of different streams. Table II contains mean phosphorus concentrations (across plants) of process stream, daily flow rates of each stream, and absolute quantities of phosphorus (P) carried in each stream per day. These data indicate that most streams (12 of 21) carried relatively low quantities of phosphorus (<1,000 kg of P/day). Of the remaining nine streams, six carried 1,000 3,200 kg/day and three carried 3,200 6,400 kg of P/day. Fig. 2. Flows of phosphorus (kg of P/day) in the wet-milling process with a capacity of 2,700 tonnes/day. Below detectable limits (bdl). Vol. 82, No. 4, 2005 433

Much (5,960 kg of P/day or 0.86) of the 6,920 kg of phosphorus entering the steeping process (phosphorus from maize and process water) was removed in the light steepwater stream. Most phosphorus in light steepwater was recovered in heavy steepwater (4,840 kg of P/day or 0.81). When light steepwater was evaporated, 1,080 kg of P/day was unaccounted for; the fate was not clear and may have been due to errors in estimation of flow rates or measurement of phosphorus concentrations. Heavy steepwater was added to pressed fiber to form CGF (Fig. 2). Because pressed fiber contained low concentrations of phosphorus, much of the phosphorus in CGF originated in heavy steepwater. Another stream that carried relatively large quantities of phosphorus was the light gluten stream (3,080 kg of P/day). A considerable amount of phosphorus in light gluten was recovered in the heavy gluten (1,460 kg of P/day); however, much of the phosphorus in light gluten was recovered in the process water leaving the gluten thickener (1,020 kg of P/day). Overall, 66% of phosphorus entering the wet-milling process (5,900 kg of P/day) could be accounted for in the CGF, CGM, and dried germ coproduct streams (2,410, 970, and 540, respectively) exiting the process. Inability to account for 34% of the phosphorus could be due to sampling errors or simplifications during process simulation, as all three plants were assumed to have similar flow rates. More accurate flow rates, specific to each plant, would likely improve accuracy of the simulation and accounting of phosphorus but they could not be obtained due to the proprietary nature of flow rate data. Three process streams carry most of the phosphorus in maize wet milling. These include light/heavy steepwater stream, light/ heavy gluten streams, and process water as it enters the steep tanks. Additional processing of either of these streams could significantly affect the phosphorus flow through the wet-milling process as well as the phosphorus content of the end products (CGF and CGM). Modification of either of these streams has specific and different implications. Changes to the light/heavy steepwater stream to remove phosphorus could reduce phosphorus content of CGF markedly; this would make CGF more marketable because it would reduce the phosphorus contribution to diets and animal wastes. However, modification of the steepwater streams would not affect the other two high-phosphorus streams. From a practical standpoint, light steepwater should be filtered more readily than Process Stream TABLE II Mean Concentrations, Flow Rates, and Quantities of Phosphorus in Process Streams a Phosphorus Concentration (mg/kg, db) Total Flow Rate (kg/day 1,000) Phosphorus Flow Rate (kg/day) Maize 2,164 2,727 5,901 Process water 344 2,960 1,018 Process water + SO 2 317 2,964 940 Steeped maize 749 4,223 3,163 Light steepwater 4,249 1,403 5,961 Heavy steepwater 14,591 332 4,844 Steepwater condensate 6 1,070 6 Wet germ 926 1,324 1,226 Pressed germ 1,304 345 450 Dried germ 2,602 206 536 Wet fiber 210 3,468 728 Pressed fiber 318 606 193 Corn gluten feed 5,375 448 2,408 Light gluten 588 5,229 3,075 Heavy gluten 1,058 1,376 1,456 Gluten cake 1,827 514 939 Corn gluten meal 4,198 230 966 Starch slurry bdl 3,648 Fresh water 5 3,423 17 Wastewater 6 1,070 6 Final effluent 6 1,070 6 a Below detectable limits (bdl); does not apply ( ). heavy steepwater. Therefore, processing light steepwater is more logical than processing of heavy steepwater. If gluten streams were modified to reduce phosphorus content, it is likely that the phosphorus content of both process water and CGM would be reduced markedly. From a practical standpoint, light gluten can be processed more effectively with microfiltration than can heavy gluten (Rausch et al 2002; Thompson et al unpublished). Therefore, the light gluten stream is more logical for processing than heavy gluten. However, because CGM is usually fed to growing animals with high dietary phosphorus requirements, the reduced phosphorus content may not result in a significant improvement in market value. The third high-phosphorus stream (process water) is a significant carrier of phosphorus. Reducing phosphorus content of process water at the steep tanks would reduce the phosphorus in steepwater and gluten streams. However, because phosphorus in process water probably is soluble, it may be difficult to remove without chemical treatment. CONCLUSIONS Phosphorus concentrations in maize did not vary among processing plants. Among plants, concentration of phosphorus in wetmilling process streams varied due to a variety of factors. There were differences in concentrations of phosphorus in coproducts (CGF, CGM) among the three plants. Based on simulated flow rates, phosphorus flows ranged from very low (<10 kg of P/day) up to 6,000 kg of P/day for a wet-milling plant (2,700 tonne/day capacity). Several process streams, light steepwater, light gluten, and process water carried large flows of phosphorus (5,960, 3,080, and 970 kg of P/day, respectively). A large portion of the phosphorus entering the simulated process could not be accounted for due to measurement and simulation errors. About one third of the phosphorus could not be accounted for, possibly due to sampling errors or simplifying assumptions of the process simulation. The wet-milling process stream most likely to benefit from phosphorus reduction was the light steepwater process stream. Reduction in the phosphorus concentration of this process stream could improve the market value of CGF. ACKNOWLEDGMENTS Funded in part by the Waste Management Research Center, Champaign, IL (DNR Contract No. HWR01168). We acknowledge the assistance of the Illinois State Geological Survey, Champaign, IL, and the Urbana Champaign Sanitary District, Urbana, IL. LITERATURE CITED AACC International. 2000. Approved Methods, 10th Ed. Method 40-56. The Association: St. Paul, MN. Agbisit, R. M., Daugherty, B. J., Belyea, R. L., Clevenger, T. E., Raskin, L. M., Tumbleson, M. E., and Rausch, K. D. 2002. Characterization and nutrient balance of corn wet milling streams. ASAE paper no. 026105. ASAE: St. Joseph, MI. APHA. 1992. Standard Methods for the Examination of Water and Wastewater. 18th Ed. A. E. Greenberg, L. S. Clesceri, and A. D. Eaton, eds. Am. Publ. Health. Assoc.: Washington, DC. Blanchard, P. H. 1999. Technology of Corn Wet Milling and Associated Processes. 2nd Ed. Elsevier Press: Amsterdam. Dou, Z., Galligan, D. T., and Ramberg, Jr., C. F. 2001. A survey of dairy farming in Pennsylvania: Nutrient management practices and implications. J. Dairy Sci. 84:966-973. ERS. 2003. Feed outlook report. United States Department of Agriculture, Economic Research Service. www.ers.usda.gov/publications Morse, D., Head, H. H., Wilcox, C. J., Van Horn, H. H., Hissem, C. D., and Harris, Jr., B. 1992. Effects of concentration of dietary phosphorus on amount and route of excretion. J. Dairy Sci. 75:3039-3049. Rausch, K. D., Thompson, C. I., Belyea, R. L., Clevenger, T. E., and Tumbleson, M. E. 2002. Characterization of gluten processing streams. Biores. Technol. 89:163-167. 434 CEREAL CHEMISTRY

Rausch, K. D., Thompson, C. I., Belyea, R. L., and Tumbleson, M. E. 2003. Characterization of light gluten and light steep water from a wet milling plant. Biores. Technol. 90:49-54. Rotz, C. A., Sharpley, A. N., Satter, L. D., Gburek, W. J., and Sanderson, M. A. 2002. Production and feeding strategies for phosphorus management on dairy farms. J. Dairy Sci. 85:3142-3153. Spears, R. A., Young, A. J., and Kohn, R. A. 2003. Whole-farm phosphorus balance on western dairy farms. J. Dairy Sci. 86:688-695. Tamminga, S. 1992. Nutrition management of dairy cows as a contribution to pollution control. J. Dairy Sci. 75:345-357. Thompson, C. I., Rausch, K. D., Belyea, R. L., and Tumbleson, M. E. Unpublished. Microfiltration of gluten processing streams from corn wet milling. Bioresource Technol. (in review). Van Horn, H. H., Newton, G. L., and Kunkle, W. E. 1996. Ruminant nutrition from an environmental perspective: Factors affecting wholefarm nutrient balance. J. Anim. Sci. 74:3082-3102. [Received April 10, 2004. Accepted March 23, 2005.] Vol. 82, No. 4, 2005 435