Tests were conducted comparing the precision of analysis Sampling and analysis of prepared cane for its ash content for 50 g samples with that obtained for 100 g samplesThe is described and evaluated. The results show the sampling and test procedure was to mix a sample of prepared cane, and analytical procedures to be without bias and the precision of a from this to withdraw ten pairs of 50 g and 100 g sub-samples. testforaconsignment o f c a n e h a s b e e n f o u n d t o b e f 0,32 The sub-samples were then ashed in axordance with the units of ash % cane, the coefficient of variation being 8,8 per procedure described in 2 (e) below. The results are shown in cent. Use of the ash I;/, cane data in arriving at an estimate Table 1. of the level of soil in the cane is discussed. TABLE 1 Comparison of ash values obtained by ashing 50 g and 100 g samples Introduction respectively The level of soil in cane delivered to some factories and the 50 g sample 100 g sample Difference resultant wear to milling plant and disruption to operations (Ash %dry sample) (Ash % dry sample) 50 g 100 g has prompted investigation into the development of a method for assessing the soil content of cane consignments. To this 7,21 7,11 +0,10 6,49 7,11 -0,62 end the Sugar Milling, Research Institute and Huletts Sugar 6,59 6,79 -0,20 Limited1 have been inbestigating an ashing method as a means 7,72 7,22 +0,50 of estimating soil levels in prepared cane. With this procedure 7,62 6,95 +0,67 the soil content of a cane sample is taken as the ash residue 6,79 7,43 -0,64 6,59 7,11 -0,52 obtained from ashing the sample under specified conditions, 6,38 7,11 -0,73 minus an ash amount deemed to be the ash content of the cane 6,69 6,79 -0,lO alone. Work on this latter aspect, viz., the assessment of the 6,59 6,95 -0,36 ash content of the cane alone, has still to be completed. Mean 6,87 7,06 -0,19 As part of the overall project the Sugar Industry Central Board, at the request of the South African Sugar Technologists' SD f 0,48 f 0,20 +0,49 Factory Control Advisory Committee, has planned and concv ducted tests to determine the accuracy and precision of cane 7,o 2 8 SE f 0,15 sampling procedures in terms of ash content assessment. n 10 10 10 The results of these investigations are reported below. SD = Standard deviation Analytical CV = Coefficient of variation n = Number of analyses (a) General SE = Standard error Essentially the procedure is to ash a sample of prepared Cane From the results shown in Table 1 it is seen that there is no to constant mass in a muffle furnace at 850°C. significant difference in accuracy between the two sample (b) Ashing container sizes. However, the precision of the 100 g sample is superior Initially stainless steel dishes (SS 316), with perforated lids, to that of the 50 g sample. were used for ashing of samples. Dimensions were 150 mm (d) Ashing time long x 80 mm wide x 30 1nm deep. It was, however, found that As mentioned above an ashing time of 2 hours is required these dishes underwent thermal degradation at the elevated to obtain constant mass when using a 100 g sample in the temperature used (850°C), and serious errors were introduced fused silica basin. Earlier work with the stainless steel dishes in weighing. which were 30 mm deep as opposed to 62 mm deep for the Fused silica basins of 400 cm3 capacity (Gallenkamp silica basins showed that constant mass could be attained at Catalogue No. B 11) are now used for ashing. Dimensions are 45 minutes for the 100 g sample. Clearly the greater depth of 145 mm diameter and 62 mm deep. Perforated lids used with the silica basin is resulting in a less efficient air flow over the these basins are from SS 316 and although the lids cane sample and a more suitably shaped container should see undergo thermal degradation, this is of no consequence as the ashint3 times well below 1 hour for the 100 g sample. lids do not feature in any of the weighing stages. The lids are (e) Procedure perforated with 2 mm holes having a pitch of 17,5 mm and fit ~ ~ t ~ : Each basin must be "cured" prior to ash determinathe basins loosely. tions by repeating 10 cycles of heating for 2 hours then cooling. The silica basins initially show a loss in mass, on average hi^ is done to eliminate the initial mass loss as mentioned 0,06 g per analysis, but after approximately 10 ashings stability in (b) above. is attained. (i) Weigh 100 g of sample, f 0,01 g, into a clean, dry and (c) Sample size pre-weighed basin. Generally the larger the sample the greater the precision of (ii) Cover the basin with the lid and put in a well vented analysis but with the ashing container in use at present larger mufflesfurnace, set at 850' C, for 30 minutes. samples also result in longer ashing times. For example, a 50 g (iii) Remove the lid from the basin and leaving the door of sample requires a total ashing time of 45 minutes to achieve the furnace slightly open (about 10 mm), to facilitate constant mass whereas a 100 g sample requires 120 minutes. air circulation, heat for a further 90 minutes. roceedings of The South African Sugar Technologists' Association June 1977 (iv) Replace the lid on the basin, remove basin from the ' furnace and cool for 2 minutes on an asbestos sheet before transferring to a desiccator for 90 minutes. (v) Place the basin, with lid, on the balance, then remove the lid and weigh the basin plus ash. (vi) If the ash % is expressed on sample, the result can now be calculated directly, but if ash % is expressed on a dry sample basis a moisture determination must be performed on a subsample taken from the original at the same time when the sample for ashing is taken. (f) Accuracy ' Lionnet and Wagenerl have tested the response of the method to varying amounts of sand of known ash content. The procedure was to mix a prepared cane sample and then sub divide into five sub-samples. One sub-sample was ashed as is, while known masses of sand (98 % ash content) were added to each of the others and the sub-samples then ashed. The results are given in Table 2. It is seen that the method responds directly to the variations in ash present in the sub-samples. (g ) Precision Analytical precision was determined by analysing 60 samples in duplicate. Each cane sample consisted of about 3 kg of cane which had been taken from the cane sampling point at the mill and subsequently prepared in two batches (each of 1,5 kg) in the sample shredder2. In the laboratory the two batches were thoroughly mixed on the bench and two 100 g samples (A and B) withdrawn for ashing. Results of the 60 comparisons are given in Table 3. From the standard deviation of differences the standard deviation of a test is obtained as follows: (i) In terms of ash % cane Standard deviation of a test = f 0,18 + d-2= f 0,13 ash % cane (ii) In terms of ash % dry matter Standard deviation of a test = f 0,61 + 47 = f 0,43 ash % dry matter (h) Comparison of the "basin" and "crucible" methods The procedure described in (e) above, but using 50 g of sample instead of 100 g of sample with commensurate reduction in ashing time from 120 minutes to 45 minutes was compared with an alternate procedure used in the past by the Sugar Milling Research Institute. The proposed method will be referred to as the "basin" method whilst the alternative Sugar Milling Research Institute TABLE 2 Mass ash analysed versus mass ash calculated Mass cane (1,47 ash %) (g) . / 50,34 / 54,42 1 57,14 1 56,46 1 61,22 Sand added (98% ash) (g). . 1 0 1 2,5 1 5,O 1 7,s 1 12,s Mass ash (g) Analysed . . . . Calculated . . . . .