Routine estimation of the Organic Matter content of soils by Loss on Ignition The organic matter content of soils is an important parameter in assessing the quality of a soil. It is formed by the breakdown of plant and animal matter and constitutes a complicated mixture of many different compounds. Therefore there is no analytical test which can provide an absolute measure of organic matter. There are several different laboratory techniques for estimating the organic matter content of soil based on various assumptions. This report compares two major analytical methods offered at NRM and re-evaluates our recommended technique for routine soil organic matter estimation in agricultural soils. Walkley Black (WB) The Walkley and Black (1934) technique uses a strong oxidising agent, potassium dichromate, to react with the organic matter in the soil. Chromium (VI) is converted to chromium (III) which is measured spectrophotometrically and provides an indication of the amount of organic matter present. This assumes complete oxidation of all organic matter present, which may not be the case for all samples. Loss on Ignition (LOI) The organic matter content is burned off under controlled conditions in a furnace and the organic matter determined gravimetrically. Various temperatures are used by different laboratories which give variable results. Which method? The colorimetric WB method has traditionally been the standard method recommended for estimating the organic matter content of agricultural soil samples. However, there are serious issues with the routine use of this procedure. Organic Matter tests based on the Walkely-Black methodology use Dichromate as one of the reagents. Dichromates are Highly Toxic and Class-1 Carcinogens which pose a serious risk to health. There has been an increase in the number of accidents involving this substance, mainly due to the large number of samples requiring testing, and therefore the large number of flasks which are having to be manipulated. The test uses relatively large amounts of Dichromate and therefore produces large amounts of waste Dichromate which is becoming increasingly expensive to dispose of. The Health and Safety risk to laboratory personnel has been a key driver in re-examining the standard organic matter method offered.
The alternative cost-effective method is LOI but there have been some historic concerns on the accuracy and precision of the results for this technique. Combustion may be an under or over estimate depending on what else is present. Structural water in clay lattices and carbonates in calcareous soils can potentially be volatilised and contribute to the LOI result. The method is also perceived to give variable results due to errors introduced through sample handling and the conditions within the muffle furnace. (Goldin, 1987). NRM have developed and validated a new LOI method to address these issues using the latest Thermo Gravimetric Analyser (TGA) equipment. The following sections outine this development. Ignition Temperature Research has been carried out into different ignition temperatures for LOI as the choice of final temperature is critical in removing the organic matter present but minimising the thermal breakdown of carbonates. Ben-Dor and Banin (1989) found temperatures between 400 and 430 C did not show a significant bias for calcareous soils. Davies (1974) demonstrated that the presence of calcium carbonate in soils did not cause errors in LOI results at 430 C. Ball (1964) suggested using lower temperatures (less than 400 C) would reduce the extent of structural water losses from clays but this has been found to lead to incomplete ashing and very long analysis times (Howard and Howard, 1990). Our investigations* showed that calcium carbonate did not ignite at 430 C (table 1.) so there should be no significant contribution to the LOI result where chalk is present. Using temperatures lower than this was found to produce an under-estimate of the organic matter content. A temperature of 430 C was selected to provide a good correlation with our existing LOI method. Sample LOI % CACO 1 0.006 CACO 2-0.025 CACO 3 0.006 CACO 4 0.028 CACO 5 0.016 CACO 6 0.006 CACO 7-0.004 CACO 8-0.005 CACO 9 0.006 CACO 10 0.005 CACO 11 0.01 CACO 12 0.011 Mean 0.005 Std Deviation 0.012 Table 1. LOI results for Analar grade Calcium Carbonate *Samples of Analar grade Calcium carbonate where analysed in the TGA at 430 C
The comparison data of WB/LOI for calcareous soils did produce the strongest correlation (correlation coeficient 0.958) out of all the soil types with a gradient very close to 1:1 (figure 1.) with no significant positive bias. This helps prove that there is no significat contribution to LOi from chalk present in the soil. Walkley-Black vs Loss On Ignition 15 Data 1:1 Line Linear (Data) 12 9 LOI y = 1.1899x - 0.565 Correlation = 0.958 6 3 0 0 3 6 9 12 15 WB Figure 1. OM Walkey Black vs Loss on Ignition for calcareous soils Variability Inconsistencies in the analytical method can cause variation in the results produced. Various factors have been studied - position in the furnace, sample size, container size, furnace loading, heating / cooling times (Schulte et al. 1991). A greater level of consistency and control of operating conditions is recommended (Hoskins, 2002). The advantage of the TGA instrumentation over the traditional muffle furnace is that as well as heating the samples it also takes the weight readings. This reduces sample handling errors, ensures full weight loss has been achieved (according to criteria set) and re-absorption of moisture from the atmosphere is not a problem. The sample carousel rotates the crucibles to ensure even heat is applied to each which eliminates sample positioning errors. Residual Moisture It has been found that it is essential to drive off the residual moisture in the soil before taking an initial weight as incomplete drying has been found to lead to significant errors (Hoskins, 2002). An initial oven drying stage and preventing the ashed sample from re-gaining moisture from the atmosphere before a final weight is taken is recommended.
The TGA is able to carry out an initial moisture content determination at a selected temperature program. The conditions for the initial moisture removal stage were found to be crucial and optimised at a temperature of 120 C which is held until the weight readings for each sample are within 0.2%. Method Validation Soil samples from a North American Proficiency Testing Program were tested a number of times and results were within acceptable limits. The table below displays the data for this validation. The number in bold is the NAPT declared OM level for each sample. These results revealed both the low variability of results from the TGA and the accuracy of the instrument and method ( average relative bias 4.95%). Batch Date Soil 2012-111 Soil 2012-112 Soil 2012-113 Soil 2012-114 (%) (%) (%) (%) 2.520 1.910 5.100 1.080 1 14/02/13 2.601 1.844 4.924 0.970 1 14/02/13 2.590 1.802 4.871 0.921 2 15/02/13 2.574 1.888 5.083 0.933 2 15/02/13 2.580 1.864 4.932 0.900 3 26/02/13 2.569 1.828 5.000 0.844 3 26/02/13 2.597 1.801 4.822 0.841 4 27/02/13 2.589 1.769 4.817 0.924 4 27/02/13 2.599 1.935 4.974 0.848 5 28/02/13 2.556 1.833 4.830 0.901 5 28/02/13 2.600 1.866 4.793 0.931 6 28/02/13 2.670 1.824 4.883 0.897 6 28/02/13 2.659 1.908 5.100 0.909 7 01/03/13 2.626 1.795 4.840 0.881 7 01/03/13 2.636 1.912 5.068 0.867 8 01/03/13 2.613 1.862 4.946 0.948 8 01/03/13 2.629 1.868 4.875 0.901 Average 2.606 1.850 4.922 0.901 Std Deviation 0.032 0.050 0.106 0.038 Bias (%) 3.4-3.1-3.5-16.6 Precision (%) 1.2 2.7 2.1 4.2 Table 2. Determination of Organic Matter by LOI(TGA) in Soils - Validation data for 2012 North America PT testing program LOI and WB comparison Comparison of organic matter results by the colorimetric WB and TGA LOI method over a wide concentration range (approx 2 60%) showed an overall correlation coefficient of 0.95 (n=317). In general, the values produced were higher for LOI than WB (by an average of around 10%) but there was not a consistent bias and several samples gave a lower LOI result. Overall, the difference in the result produced did not appear to be significant in assessing the level of soil organic matter (low / medium / high / very high).
Walkley-Black vs Loss On Ignition 60 50 Data 1:1 Line Linear (Data) 40 LOI 30 y = 0.9796x + 1.6788 Correlation = 0.953 20 10 0 0 10 20 30 40 50 60 WB Figure 2. OM Walkey Black vs Loss on Ignition for all soils Where there was a large difference in the results (greater than +/- 2%) it appeared that sample homogeneity was a key factor. It was observed that several samples had visible organic matter (grass etc) present. The 2mm soil grind is used for both these tests but the amount taken is very different. Only 0.5g is used for the WB procedure compared to 5ml (approx 5g depending on soil density) for the LOI technique. A larger sub-sample is better for obtaining a representative analysis portion, especially for highly organic samples which may not be as homogeneous. Variability for Clay Soils From further investigation on the WB/LOI comparison data for clay soils (figure 3) the correlation was not as strong but still significant (correlation coefficient 0.862) and shows that a difference in result due to the method for clay soils would not affect the assessment of the soil in terms of its banding of organic matter content (low / medium / high / very high). Samples where there has been a positive bias in the LOI result in comparison to the WB results may indicate a measureable volatalisation of mineral and water structures within the soil particulates however from the comparison data this does prove to be an insignificant addition to the LOI result.
Walkley-Black vs Loss On Ignition 30 25 Data 1:1 Line Linear (Data) 20 LOI 15 y = 0.9038x + 1.893 Correlation = 0.862 10 5 0 0 5 10 15 20 25 30 WB Figure 3. OM Walkey Black vs Loss on Ignition for clay soils There have been various investigations attempting to determine a correction factor between organic matter by LOI and WB methods. Konare et al (2010) found it may be possible for a set of soils / sediments from a specific geographic region where information about the sample types is well known but this is not widely applicable. Howard and Howard (1990) concluded that differences in the nature of the organic matter and the clay content of soils meant that establishing a single conversion factor could lead to significant errors. Conclusion The determination of organic matter in soil is technique specific and there is no absolute correct answer. The organic matter content of a soil should always be defined in terms of the analytical method used. Organic matter by WB may have previously been considered to be the best available costeffective option but with health and safety risks this is no longer the case. NRM are confident our redeveloped LOI method by TGA offers a safe, reliable and fit for purpose assessment of the organic matter content for all soil types including chalky soils.
References: Ball, D.E. (1964) Loss-on-ignition as an estimate of organic matter and organic carbon in non-calcareous soils. J Soil Science 15: 84-92. Ben-Dor, E. and Banin, A. (1989) Determination of organic matter content in arid zone soils using a simple loss-on-ignition method. Communications in Soil Science and Plant Analysis 20: 1675-1695. Davies, B.E. (1974) Loss-on-ignition as an estimate of soil organic matter. Soil Science Society of America. Proceedings, 38:150-151. Hoskins, B. (2002) Organic Matter by Loss on Ignition. University of Maine. Howard, P.J.A. and Howard, D.M. (1990) Use of organic carbon and loss-on-ignition to estimate soil organic matter in different soil types and horizons. Biol Fertil Soils 9: 306-310. Konare, H., Yost, R.S., Doumbia, M., McCarty, G.W., Jarju, A., Kabian, R. (2010) Loss on ignition: Measuring soil organic carbon in soils of the Sachel, West Africa. African Journal of Agricultural Research, Vol 5(22) 3088-3095. Schulte, E.E., Kaufmann, C., Peters, J.B. (1991) The influence of sample size and heating time on soil weight loss-on-ignition. Comm. Soil Science and Plant Analysis 22: 159-168. Walkley, A. & Black, I.A. (1934): An examination of the Degtjareff method for determining soil organic matter, and a proposed modification of the chromic acid filtration method. (Soil Sci., 37, pp 29-38). Goldin, A. (1987): Reassessing the use of loss-on-ignition for estimating organic matter content in noncalcareous soils. Comm Soil Science and Plant Analysis 18: 1111-1116