School of the Environment and Natural Resources, Bangor University, Gwynedd LL57 2UW, UK.
Bioresour Technol. 2010 Jul;101(14):5431-6. doi: 10.1016/j.biortech.2010.02.033. Epub 2010 Mar 23.
Chemical properties have been used as a way of following the composting process and compost maturity, however, their analysis is very time consuming as each must be separately determined. By developing a more rapid method to predict these properties, time and cost would be saved. This study investigates the use of Fourier Transform mid-Infrared Spectroscopy (FT-IR) for this purpose. FT-IR spectra and measured values of several chemical properties from a variety of compost mixtures were used to produce calibrated models using partial least-squares regression analysis which predicted the known chemical properties. These models displayed a range of accuracies that for most properties was more than sufficient to follow at least broad dynamic changes associated with maturation. The best calibrations were achieved for total C, total N, LOI, lignin, and cellulose with r(2) values within the range 56-77%. Some degree of calibration was achieved for available-P and NH(4)(+)-N, with r(2) values of between 40% and 57%. No useful calibration could be achieved for NO(3)(-) or pH.
化学性质已被用作跟踪堆肥过程和堆肥成熟度的一种方法，然而，由于必须分别测定每个性质，因此其分析非常耗时。通过开发更快速的方法来预测这些性质，可以节省时间和成本。本研究调查了傅里叶变换中红外光谱（FT-IR）在这方面的用途。FT-IR 光谱和来自各种堆肥混合物的几种化学性质的实测值用于使用偏最小二乘回归分析来生成校准模型，该模型预测了已知的化学性质。这些模型显示出一系列的准确性，对于大多数性质来说，足以跟踪与成熟相关的至少广泛的动态变化。对于总 C、总 N、LOI、木质素和纤维素，校准效果最佳，r(2) 值在 56%至 77%之间。对于有效 P 和 NH(4)(+)-N，也实现了一定程度的校准，r(2) 值在 40%至 57%之间。无法为 NO(3)(-)或 pH 值获得有用的校准。
