Larsson William, Cedergren Anders
Analytical Chemistry, Department of Chemistry, Umeå University, SE-90187 Umeå, Sweden.
Talanta. 2005 Mar 15;65(5):1349-54. doi: 10.1016/j.talanta.2004.09.005.
Factors influencing the accuracy and precision for diaphragm-free Karl Fischer coulometric determinations of low mug-amounts of water have been studied using the Metrohm 756 (pulsed current) coulometer and eight different types of commercial coulometric reagents and some modifications of these. As in the case of diaphragm-free coulometric titration of large amounts of water, the positive errors, due to the formation of oxidizable reduction products (of sulfur dioxide) in the cathode reaction (besides hydrogen), were found to be minimized by the use of highest possible pulse current (in the range 100-400mA) in combination with the fastest possible titration rate. Most accurate (102-103%) and precise results (typical relative standard deviation 1.8%) were obtained for reagents containing very large concentrations of imidazole in combination with the presence of modifiers like hexanol, chloroform and propylene glycol (i.e. the HYDRA-POINT reagents). Similar results were obtained when this type of reagent was mixed 60/40 with xylene according to the ASTM recommendation for water determinations in petroleum products like crude and lubricating oils. Addition of decanol to this type of reagent mixture was found to reduce the influence from the oxidative reduction products significantly. A reduction of the error from 3.6% relative to 1.6% was achieved by addition of 9% (v/v) of decanol to a 60/40 reagent mixture of HYDRA-POINT Coulometric Gen (containing hexanol as modifier) and xylene. For larger concentrations of decanol the pulse current had to be lowered to 100mA and this might explain why no further improvement was observed. An additional attempt to minimize the interference by lowering the concentration of sulfur dioxide in the reagents gave no significant effect. However, by means of a home-built computer-controlled coulometric instrumentation based on continuous instead of pulsed current (including a large cathodic current density) it was possible to achieve recovery rates close to 100% for the best reagents investigated. The reason for this improvement is discussed.
使用万通756（脉冲电流）库仑仪以及八种不同类型的市售库仑试剂及其一些改进试剂，研究了影响低微克量水的无隔膜卡尔费休库仑测定准确性和精密度的因素。与大量水的无隔膜库仑滴定情况一样，发现通过使用尽可能高的脉冲电流（100 - 400mA范围）与尽可能快的滴定速率相结合，可将阴极反应中（除氢气外）由于（二氧化硫）可氧化还原产物形成而导致的正误差降至最低。对于含有非常高浓度咪唑并伴有己醇、氯仿和丙二醇等改性剂的试剂（即HYDRA - POINT试剂），可获得最准确（102 - 103%）和精密的结果（典型相对标准偏差1.8%）。当根据ASTM对原油和润滑油等石油产品中水测定的建议，将此类试剂与二甲苯按60/40混合时，也获得了类似结果。发现向此类试剂混合物中添加癸醇可显著降低氧化还原产物的影响。向含有己醇作为改性剂的HYDRA - POINT Coulometric Gen与二甲苯的60/40试剂混合物中添加9%（v/v）的癸醇，可将误差从3.6%降至1.6%。对于更高浓度的癸醇，脉冲电流必须降至100mA，这可能解释了为何未观察到进一步的改善。通过降低试剂中二氧化硫浓度来最小化干扰的额外尝试未产生显著效果。然而，借助基于连续而非脉冲电流（包括大阴极电流密度）的自制计算机控制库仑仪器，对于所研究的最佳试剂，有可能实现接近100%的回收率。讨论了这种改进的原因。
