Investigation of IRRI Test Tube Mill Operating Parameters

Transcription

1 RICE QUALITY AND PROCESSING Investigation of IRRI Test Tube Mill Operating Parameters R.C. Bautista, T.J. Siebenmorgen, and R.M. Burgos ABSTRACT The objective of this study was to optimize the operating parameters of the International Rice Research Institute (IRRI, Los Banos, The Philippines) Test Tube rice mill. Operating parameters including brown rice mass (BRM), milling duration (MD), tube oscillation frequency (OF), and test tube head space (THS) were investigated with the brown rice mass-loss percentage (MLP) as the response parameter. Two rice cultivars, Bengal (medium-grain), and Cypress (long-grain), at approximately % MC were used. Results indicated that milling with the IRRI Test Tube mill was significantly affected by MD and OF. The results provide a guide to users, particularly rice breeders and physiologists, in the selection of proper mill design and operation for different amounts of brown rice sample. INTRODUCTION Rice samples from breeding lines, ranging from approximately 2 to g or even smaller quantities obtained from a single or multiple panicles, are often produced in research with a need to be milled in order to estimate milling quality. While there are an enormous number of small-quantity samples being generated from experimental test plots and breeding lines with this need, there is no accepted laboratory mill that adequately provides an indication of milling quality. Such a mill would be extremely beneficial to rice breeders, cereal chemists, and physiologists. In a recent effort to meet the above need, an evaluation of milling performance of the Kett Pearlest Polisher and the IRRI Test Tube mill was made relative to the milling behavior of the McGill No. 2 lab mill. This study indicated the potential of the IRRI Test 343

2 AAES Research Series 504 Tube mill for milling small quantities of brown rice (Bautista and Siebenmorgen, 2002). The IRRI mill, however, required further improvement to reduce the duration of milling; currently, the IRRI Test Tube mill requires more than two hours to mill samples to remove the bran from brown rice. The IRRI mill was developed to mill very small masses of brown rice. It accomplishes milling by vigorously oscillating tubes containing brown rice mixed with an abrasive compound (aluminum oxide) that acts as a milling aid. The mill essentially comprises two boxes oscillated by an eccentric shaft, with each box holding as many as 35 -ml test tubes. Each test tube typically contains approximately 3 g of brown rice and about 4 g of aluminum oxide. The milling duration required to remove bran varies with rice variety and moisture content (MC), which follows the findings of Andrews et al. () using a McGill No. 2 mill. Milled rice is separated from the bran and aluminum oxide by using a sieving screen. The mill was developed with the intention of milling rice from greenhouse or breeding studies in which very small amounts of rice are produced yet an indication of milling quality is desired. The study conducted herein investigated whether the operation of the IRRI Test Tube mill could be optimized in order to achieve shorter milling durations for milling small quantity samples of medium- (Bengal) and long-grain (Cypress) brown rice cultivars. The American Association of Cereal Chemists (AACC)-established % brown rice MLP criterion was used as the response parameter to indicate that sufficient bran removal had occurred to produce well-milled rice. MATERIALS AND METHODS Figure 1 shows the experimental design for evaluating the operation of the IRRI Test Tube mill. Using the brown rice MLP as a response parameter, the following parameters were evaluated: brown rice mass (BRM), milling duration (MD), oscillation frequency (OF), and tube head space (THS). Four tube oscillation frequencies [0, 140, 140, and 2040 cycles per minute (cpm)] were used. Three tube sizes (, 1, and 1 ml) were selected based on a preliminary investigation comprising tube sizes from to 20 ml. Brown rice sample masses of 2, 3, 4, and 5 g were used in the ml tube;,, and g in the 1 ml tube; and,, and g in the 1 ml tube. Tube head-space volumes were computed for each tube size given the corresponding masses of brown rice and aluminum oxide for both cultivars. The following bulk densities were used to calculate the THS: 0. g/cc for Bengal brown rice, 0.3 g/cc for Cypress brown rice, and 3.g/cc for aluminum oxide (Lide, 1). Brown rice bulk densities were measured using the method used by Candole et al. (2001). Milling duration varied from to 25 min. The mass of aluminum oxide, with grit size # 4 (Duralum special white, Washington Mills Electro Minerals, Inc.), was fixed at 5 g for all experiments. The response variable used was the brown rice MLP, which was computed as follows: Brown rice mass loss (MLP), % = Mass of brown rice Mass of milled rice (head rice + brokens) Mass of brown rice X 0% 344

3 B.R. Wells Rice Research Studies 2002 Treatment combinations of BRM, MD, OF, and THS were tested for Bengal and Cypress brown rice. All treatment combinations were replicated. The experimental procedure consisted of placing the desired masses of brown rice and aluminum oxide in a tube and closing with a stopper. A pair of wooden tube holders was used for the experiment, each with four holes drilled on one face to accommodate a maximum of eight tubes for each run. Each tube holder was secured on the mill s oscillating arm by a winged lock nut. The frequency of tube oscillation for each experiment was determined by adjusting the diameter of a variable-diameter pulley ( to mm) and checked with a tachometer. A timer was used to monitor the milling duration. After milling, the samples were allowed to cool inside the tubes prior to opening and separating the milled rice from the bran and aluminum oxide using a # 20 screen (Central Scientific Co., Chicago, IL) and weighed. Whole-kernel yield (WKY), which is defined as the ratio of the mass of milled kernels with length at least 0.5 of the original kernel length after milling (head rice) to the mass of original rough rice, was measured for each experiment. The mass of rough rice was estimated by dividing the initial brown rice mass by 20%, which is the assumed average mass percentage of rough rice that is hulls. Whole kernel yield is to be distinguished from HRY as used by those in the commercial milling industry in that WKY consists of kernels 0.5 or more of the original kernel length that may or may not have been milled to an acceptable degree of milling, which in this study is determined by the % MLP. The head rice from the IRRI Test Tube mill setting may be well-milled or could be below the required degree of milling, depending on the experimental milling treatment. Thus, the WKY may have values much higher than would be encountered for HRYs. To obtain WKY, head rice was separated from brokens by using a sizing machine (Grainman, Model 1-5-0, Grain Machinery Manufacturing Corp., Miami, FL). The following equation applies: Whole kernel yield = Mass of head rice Estimated mass of rough rice RESULTS AND DISCUSSION Mass Loss Percentage Response Results of multivariate response analyses (JMP ver. 5, Sall et al., 2002) are shown in tables 1 and 2 for Bengal and Cypress, respectively. For both rice cultivars, the independent milling parameters of BRM, MD, OF, and THS were analyzed for individual and cross-parameter effects on MLP. Significant effects on brown rice MLP were found due to MD, OF, and interactions among MBR, MD, OF, and THS (p<0.0001). Figures 2 and 3 show the brown rice MLP for the various treatment combinations for Bengal and Cypress. Bengal generally had lower brown rice MLPs compared to Cypress. These results agree with the findings of Bautista and Siebenmorgen (2002) where a medium-grain Bengal had less bran removed (lower MLP) than a long-grain Drew using the IRRI Test Tube mill. The MLP response for Bengal and Cypress resembled a parabolic function for brown rice masses of sample sizes 2, 3, 4, and 5 g 345

4 AAES Research Series 504 brown rice used in the ml tube. Mass loss percentage for Bengal and Cypress decreased with increases in sample masses for,,,,, and g at all oscillation frequencies. These results indicated that treatment combinations can be optimized for the IRRI Test Tube mill for certain BRM that meets the % MLP level for Bengal and Cypress, as indicated by a model in Table 3. Test Tube Volume Effects on Mass Loss Percentage The results indicated the importance of the brown rice mass used for a given tube size. This implies that there is a critical relationship between the occupied volume (brown rice and aluminum oxide) and the THS for greatest bran removal. The THS provides room in which rice kernels can move back and forth and tumble as the tube is oscillated by the reciprocating motion of the tube holder. Tube head-space can vary slightly according to rice cultivar, since rice bulk density can vary with cultivar. Whole-Kernel Yields Whole-kernel yields were calculated for Bengal and Cypress for the different milling treatment combinations. Whole kernel yield was inversely related to MD and OF wherein WKY decreased with increases in MD and OF. For different tube sizes and sample masses, the trend in WKY was similar at all oscillation frequencies. Between cultivars, Bengal had, in most cases, lower WKY than Cypress. SIGNIFICANCE OF FINDINGS This study was able to determine the effects of milling parameters on MLP for the IRRI Test Tube mill. Significant effects on MLP were found due to MD, OF, and interaction among BRM, MD, OF, and THS. A parabolic relationship indicates optimal BRMs for the given tube sizes and oscillation frequencies that meet the % MLP level for Bengal and Cypress. The results of this study should also be useful in operating similar mills when milling different small masses of brown rice. ACKNOWLEDGMENTS The authors are very grateful to the Arkansas Rice Research and Promotion Board and the University of Arkansas Rice Processing Program Industry Alliance Group for the financial support of this project and to Washington Mills Electro Minerals, Inc (Buffalo, NY) for providing the aluminum oxide for the experiments. 34

5 B.R. Wells Rice Research Studies 2002 LITERATURE CITED Andrews, S.B., T.J. Siebenmorgen, and A. Mauromostakos.. Evaluation of the McGill No. 2 miller. Cereal Chem. (1): Bautista, R.C. and T.J. Siebenemorgen Evaluation of laboratory mills for milling small samples of rice. Applied Engineering in Agriculture 1(5): Candole, B.L., T.J. Siebenmorgen, F.N. Lee, and R.D. Cartwright Effect of blast and sheath blight on physical properties of selected rice cultivars. Cereal Chem. (5): Lide, D.R. 1. Handbook of chemistry and physics, th edition, CRC Press, New York, pp Sall, J., A. Lehman, and L. Creighton JMP IN v. 5. SAS Institute, Inc. Cary, NC. Duxbury Press. Table 1. Multivariate response analyses indicating individual and cross-parameter tests of brown rice mass, milling duration, tube oscillation frequency, and tube head space for rice cultivar Bengal (at.% MC) using the IRRI Test Tube mill with five g aluminum oxide. Source DF Sum of squares F ratio Prob>F Brown rice mass (BRM) Milling duration (MD) < Oscillation frequency (OF) < Tube head space (THS) MBR*MD*OF*THS < Table 2. Multivariate response analyses indicating individual and cross-parameter effect tests of brown rice mass, milling duration, tube oscillation frequency, and tube head space for rice cultivar Cypress (at.5% MC) using the IRRI Test Tube mill with five g aluminum oxide. Source DF Sum of squares F ratio Prob>F Brown rice mass (BRM) Milling duration (MD) < Oscillation frequency (OF) < Tube head space (THS) MBR*MD*OF*THS < Table 3. Milling parameter constants for linear models to predict mass loss percentage for Bengal and Cypress using the IRRI Test Tube mill. Cultivar α ß 1 ß 2 ß 3 ß 4 R 2 Bengal Cypress

6 AAES Research Series 504 ml Tube Head space 5 g Aluminum oxide (Al2O3) as an abrasive milling aid 2, 3, 4, 5 g Brown rice Oscillation frequency Milling duration (min) (cycles per minute) 0, 20, , 20, ,, ,, 1 1 ml Tube Head space 5 g Aluminum oxide (Al2O3) as an abrasive milling aid,, g Oscillation frequency Milling duration (min) (cycles per minute) 0, 20, , 20, ,, ,, 1 Brown rice Head space 5 g aluminum oxide (Al2O3) as an abrasive milling aid,, g Brown rice 1 ml Tube Oscillation frequency Milling duration (min) (cycles per minute) 0, 20, , 20, ,, ,, 1 Fig. 1. Schematic of the experimental design used for evaluating oscillation frequency and milling duration of the IRRI Test Tube mill for milling the indicated amounts of brown rice of a medium- (Bengal) and a long-grain (Cypress) brown rice cultivar using three tube sizes. Aluminum oxide (Al 2 O 3 ) was used as milling aid. 34

7 B.R. Wells Rice Research Studies 2002 a) 0 cycles per minute ml tube 1 ml tube 1 ml tube b) 140 cycles per minute ml tube 1 ml tube 1 ml tube c) 140 cycles per minute ml tube 1 ml tube 1 ml tube Brown rice mass (g) d) 2040 cycles per minute ml tube 1 ml tube 1 ml tube 1 min 1 min 1 min Fig. 2. Brown rice mass loss percentage for Bengal brown rice (at.5% MC) milled with the IRRI Test Tube mill using oscillation frequencies: 1) 0 cpm, b) 140 cpm, c) 140 cpm, and 3) 2040 cpm at various milling durations, brown rice masses, and tube sizes. A desirable level of milling equivalent to % brown rice mass loss percentage is indicated by broken lines. Each data point is the average of two replications. 34

8 AAES Research Series 504 a) 0 cycles per minute ml tube 1 ml tube 1 ml tube b) 140 cycles per minute ml tube 1 ml tube 1 ml tube Brown rice mass loss (%) c) 140 cycles per minute ml tube 1 ml tube 1 ml tube Brown rice mass (g) d) 2040 cycles per minute ml tube 1 ml tube 1 ml tube 1 min 1 min min Fig. 3. Brown rice mass loss percentage for Cypress brown rice (at.1% MC) milled with the IRRI Test Tube mill using oscillation frequencies: a) 0 cpm, b) 140 cpm, c) 140 cpm, and d) 2040 cpm at various milling durations, brown rice masses, and tube sizes. A desirable level of milling equivalent to % brown rice mass loss percentage is indicated by broken lines. Each data point is the average of two replications. 350