10
Analytical Instrumentation
A Fast Method for Determining PCB/Chlorine Contamination in Transformer Oils by Monochromatic X-Ray Fluorescence (MCXRF)
David Malone, Rudy Haas, HORIBA Scientific 17671 Armstrong Avenue, Irvine, CA 92614 USA, Tel: HQ (949)250-4811,
www.horibalab.com,
rudy.haas@horiba.com
Photo courtesy of : Power Substation Services, Wheeling, WV.
On an annual basis, millions of electrical transformers are removed from service in the global power grid. This activity generates many tons of solid waste and hundreds of millions of gallons of dielectric fluid / coolant (transformer oil) that can be recycled or must be incinerated as a hazardous waste. Although many chemical analysis criteria are utilised to determine the fate of both the solid and liquid reclaimed materials, the need to demonstrate regulatory compliance and control reclamation processes make testing for the presence of polychlorinated biphenyls (PCBs) contamination in used transformer oil absolutely essential.
Introduction / Abstract
Polychlorinated biphenyls (PCBs) are a liquid mixture of chemically stable chlorine containing compounds. They were widely used in the twentieth century for many electrical applications because of their excellent heat resistance and electrical insulation (dielectric) charact- eristics. Beginning in the late 1970’s, production of high purity PCBs-based transformer oil material ceased in the United States and was eventually banned in most countries because of evidence they build up in the environment and can cause harmful health effects.
Today PCBs contamination in transformer oil is measured and controlled as millions of gallons of various dielectric/transformer oil materials are produced, recycled and traded on a global basis. Most analysis techniques take advantage of the high chlorine content in the individual compounds that comprise PCBs mixtures and measurement of specific PCBs compounds can readily be accomplished by a variety of analytical techniques, such as gas chromatography and mass spectrometry. However, analysis for the specific compounds that make up a PCBs mixture is not always practical or necessary, particularly when large number of individual samples must be characterised.
As indicated above, measurement of chlorine can be an excellent predictor for actual PCB contamination in transformer oil and other fluids of interest. X-Ray fluorescence (XRF) technology provides a reliable, fast and easy-to-use chlorine measurement methodology. The proven technique has been used for decades for primary measurement and most importantly for screening applications that reduce the need to analyse large numbers of materials with more expensive and time consuming analytical techniques.
The following describes the findings of a thorough examination of a new bench-top Monochromatic Energy Dispersive X-Ray Fluorescence (MCXRF) spect- rometer. The compact (13”Wx 20”D x 16”H ) and lightweight (30 lbs) analyser was found to be able to readily measure chlorine content in transformer oils to part per million levels, in duplicate, in six minutes. The analyser’s performance characteristics easily meet transformer oil recycling industry requirements whenever rapid reliable chlorine measurement is essential to routine screening and process feed control activities.
Project Background and Data Origin
Analytical Services Inc. (ASI), supplies certified reference materials used for analytical instrument calibration and provides various project management and instrument validation services. They do not endorse, nor serve as an advocate for any instrument manufacturer. ASI was contracted to conduct an independent testing of the Horiba MESA 6000 chlorine in transformer oil capability and all data presented was generated by ASI personnel at their The Woodlands, TX facility. Two instruments were utilised in the study and a side-by-side testing regime
was used to demonstrate instrument agreement, accuracy, repeatability and reproducibility. The sample materials tested contained certified chlorine (Cl) content that was gravimetrically prepared in fresh transformer oil (TO) using chlorobenzene as the Cl source. Each instrument was provided separate sample sets, calibration materials, and a performance check sample. All testing was completed in less than 8 days.
Calibration – Linearity
One fundamental assessment of analytical instrument performance is verification of calibration curve linearity. Both instruments tested by ASI during the trial demonstrated excellent linearity throughout the 0-500 mg/kg Cl in transformer oil calibration range. The following graph depicts the typical near-perfect linearity exhibited by both apparatus tested and provides an indication of the quality of the calibration materials utilized (see figure 1 and note 1).
Additionally, each of the sample materials analysed had certified assigned values (see note 1) that were unknown to the analyst and were held in confidence until completion of testing. An analysis of ASI’s testing results can be found in table 1 and allow the following observations.
• Final results agreement between Instrument 1 and 2 was very good as standard deviations ranged between 0.5 and 10 mg/kg Cl.
• Sample set quality and instrument accuracy were affirmed, as percent recovery values derived from the average of Instrument 1 and 2 final results were found to be very near 100% of the certified value of the samples
• No significant bias was identified as results varied above and below the certified chlorine concentrations.
Figure 1: Chlorine Calibration Curve
Following confirmation of calibration curve linearity, instrument accuracy was demonstrated when separate performance check materials, each having a certified reference value of 40 mg/kg, were measured to contain 39 mg/kg chlorine by both apparatus (see note 1).
Note 1: The calibration, performance check and sample materials were certified to be within their assigned value at an 0.8% expanded degree of uncertainty, based on errors from assay analysis and the weighing of raw materials at a 95% confidence level (k=2).
Instrument Accuracy and Results Agreement
The ASI designed testing plan, required that both instruments measure the chlorine in each sample in their assigned sample set on 2 consecutive days. Each days result was derived from the average of two consecutive 180 second analyses of 5 mls of transformer oil in the same sample analysis cup. The day one and day two results were then averaged to yield a final result for each sample from both instruments. The test was designed to allow a close evaluation of between instrument result agreement and short-term calibration stability.
Table 1: Two Instrument Result Comparison and Percent Recovery Data
Equivalency
In order to determine how the MCXRF methodology performance might compare to another estab- lished chlorine measurement technology, a separate but identically prepared calibration standard and sample set was analysed with Wavelength Dispersive X-ray Fluorescence (WDXRF) Instrumentation. Following instrument calibration with chlorine in transformer oil standards, the six samples were analysed in duplicate to yield the WDXRF result. An analysis demonstrates that the MCXRF and WDXRF results readily compare to each other (see table 2).
Instrument Precision and Estimation of Level of Detection
As previously discussed, during testing each instrument was supplied with a certified 40 mg/kg Cl performance check sample that was used to confirm calibration curve validity . To determine precision, ASI personnel used instrument 1 to analyse the performance check material 21 consecutive times. As with previous testing,
February/March 2010
Page 1 |
Page 2 |
Page 3 |
Page 4 |
Page 5 |
Page 6 |
Page 7 |
Page 8 |
Page 9 |
Page 10 |
Page 11 |
Page 12 |
Page 13 |
Page 14 |
Page 15 |
Page 16 |
Page 17 |
Page 18 |
Page 19 |
Page 20 |
Page 21 |
Page 22 |
Page 23 |
Page 24 |
Page 25 |
Page 26 |
Page 27 |
Page 28 |
Page 29 |
Page 30 |
Page 31 |
Page 32 |
Page 33 |
Page 34 |
Page 35 |
Page 36 |
Page 37 |
Page 38 |
Page 39 |
Page 40 |
Page 41 |
Page 42 |
Page 43 |
Page 44 |
Page 45 |
Page 46 |
Page 47 |
Page 48 |
Page 49 |
Page 50 |
Page 51 |
Page 52
