Determination of total fluoride and boron in preservative-treated wood fluoride by ion selective electrode without steam distillation, and boron colorimetrically using azornethine-h
P A Collins, M J Kennedy
An analytical method was developed to determine total fluoride and boron in wood treated with aqueous or solid diffusing fluoride or borate-fluoride formulations. Ground wood was mixed with 30% sodium hydroxide solution and oven-dried for 2 hours before furnacing at 600° C for 1 hour. The cool fused product was dissolved in hot water, transferred to a plastic beaker and neutralised. Fluoride was measured by ion-selective electrode (ISE), after adding Ionic Strength Adjustment Buffers (ISAB), against appropriately pH and matrix-matched calibration standards containing 1, 3, 5, 10 and 50 ppm F. Boron was determined colorimetrically on the same neutralised solution using azomethine-H chemistry.
Citric acid, sulphuric acid and nitric acid were evaluated as neutralising agents. While the choice of neutralising acid had no significant influence on the determination of boron by azomethine-H, speed and stability of the ISE for fluoride was best with nitric acid at 3.75M.
The ISABs Sulphosalicylic acid and EDTA (SSA-EDTA); tri-sodium citrate and potassium nitrate; tiron and sodium nitrate; and TISAB were all evaluated for fluoride determination. Only SSA-EDTA was found to complex aluminium adequately, also accommodating borate, silicate and iron. Although SSA-EDTA has been traditionally used at pH 9.5 for the determination of fluoride in large amounts of aluminium, we found that use at this pH led to significant problems from hydroxide ion interference, slow electrode response, and pH changes due to loss of ammonia, all resulting in poor detection limits and reproducibility. To improve method performance, the pH of standards and samples was adjusted to 8.7, and all samples were spiked with an additional 1 mg/L fluoride. Interference from aluminium increased marginally, but the SSA-EDTA buffer was still capable of masking aluminium at a ratio of 20:1 AI:F, when measuring solutions containing 5 mg(F)/L.
The method compared favourably with two alternatives; one employed the time-consuming distillation step included in the AWPA A2-1994 method, with a titration finish, and the second employed a double fusion and ISE finish, using TISAB buffer. Mean recovery from spiked Tasmanian oak (Eucalyptus regnans) at 0.05% - 0.20% m/m B/F was 95-105%. Precision at 0.2% m/m B/F in wood was about 3% for F and 1% for B. Detection limits of 30 mg/Kg for F and 20 mg/Kg for B were adequate for monitoring movement of toxicants within remedially-treated hardwood poles.