XRF and TGA Commissioning outcomes at Cape Lambert Port B R.W.Brunning, C. Andringa-Bate Rio Tinto Iron Ore Introduction The Cape Lambert Port B (CLB) project comprises the construction of a new Port facility adjacent to the Cape Lambert Port A facility (CLA). This involved the construction of a new 100Mt/a train unloading infrastructure, stockyard, ship-loading facilities (including a new ore wharf) and the construction of an automated sampling and analysis laboratory. Figure 1 shows CLB layout with the stockpiles in the background, followed by the rescreening plant and the pale large building in the foreground which houses the CLB sampling facility and analytical cell. Figure 1. The photograph shows the location of Cape Lambert Port B Laboratory in the foreground RTIO has traditionally designed and commissioned laboratories using semi-automated sample weighing systems, fusion machines, Thermo-Gravimetric Analysers (TGA’s) and XRF spectrometers to process the numerous samples produced from mining, processing and ship-loading activities. At CLB the sample station and preparation facilities are fully automated, providing crushing, dividing, drying and sieving Laboratory
17
Embed
XRF and TGA Commissioning outcomes at Cape Lambert Port B CLB Paper 10 Feb 14... · XRF and TGA Commissioning outcomes at Cape Lambert Port B R.W.Brunning, C. Andringa-Bate Rio Tinto
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
XRF and TGA Commissioning outcomes at Cape Lambert Port B
R.W.Brunning, C. Andringa-Bate Rio Tinto Iron Ore
Introduction
The Cape Lambert Port B (CLB) project comprises the construction of a new Port facility adjacent to the
Cape Lambert Port A facility (CLA). This involved the construction of a new 100Mt/a train unloading
infrastructure, stockyard, ship-loading facilities (including a new ore wharf) and the construction of an
automated sampling and analysis laboratory. Figure 1 shows CLB layout with the stockpiles in the
background, followed by the rescreening plant and the pale large building in the foreground which
houses the CLB sampling facility and analytical cell.
Figure 1. The photograph shows the location of Cape Lambert Port B Laboratory in the foreground
RTIO has traditionally designed and commissioned laboratories using semi-automated sample weighing
systems, fusion machines, Thermo-Gravimetric Analysers (TGA’s) and XRF spectrometers to process the
numerous samples produced from mining, processing and ship-loading activities. At CLB the sample
station and preparation facilities are fully automated, providing crushing, dividing, drying and sieving
Laboratory
operations. Although not originally in the scope of works, automated, analytical cell facilities were
added to the project to reduce the need to transport samples around the site.
In 2012 RTIO/SKM/IMP commenced work for the design and development of the analytical facility at
Cape Lambert Port B. The analytical facility would need to have the capacity to process cargos totaling
100Mt/a at ship-loading rates of 13,440tphr. During loading, individual cargos are composited within the
sampling cell for chemical analysis. On completion of loading the cargo; composite material is crushed,
pulverized and sub-sampled before being presented for XRF and TGA analysis. The analytical cell
comprises a vial storage magazine and a HAG-HF (fusion/weigh cell) interfaced to an Axios XRF
spectrometer and automated TGA system.
Figure 2 CLB Analytical Cell Layout
The new automated analytical facilities were required to have accuracy and a precision as good or better
than current port laboratory facilities. This paper examines the results of XRF calibration and TGA
analysis commissioning between August and December 2013. Analytical results obtained from the cell
are evaluated against the iron ore international standards for chemical analysis ISO9516 (2003 - Iron
Ores - Determination of various elements by X-ray Fluorescence Spectrometry) and hygroscopic moisture
ISO2596 (2006 - Iron ores – Determination of Hygroscopic moisture in analytical samples- Gravimetric,
Karl Fischer, and mass-loss methods).
Method
Weighing and HAG-HF Fusion
The CLB HAG-HF unit comprises of a weigh cell (Figure 3), two fusion ovens (Figure 4), and a citric acid
bath, water bath and air drying to allow cleaning of crucibles and molds. Crucibles and molds are moved
between the various areas in the HAG-HF by a single pivot mounted Mitsubishi robot. The CLB HAG-HF
design enables the continuous fusion of samples introduced via the vial storage magazine as cargo
samples are presented. The vial storage magazine also enables the introduction of general QA/QC
standards and other ad-hoc samples.
Figure 3 HAG-HF weigh cell for fusion and TGA Figure 4 HAG-HF Oven layout
The HAG-HF fusion ovens were utilized for the fusion of all XRF calibration samples. However, calibration
samples were manually weighed external to the HAG-HF due to the small weights of pure oxides to be
used and the numerous binary compounds to be fused. Calibration samples were introduced into the
HAG-HF via a manual sample input. All test samples were introduced via the vial storage magazine
utilizing the HAG weigh cell, fusion and cleaning facilities in order to evaluate TGA and XRF performance.
All test standards were allowed to equilibrate to the laboratory atmosphere over a 12hr period prior to
hygroscopic moisture analysis as required by ISO2596. Performance of TGA and XRF system was based
on the using a range of RTIO standards certified using ISO accredited methods.
Automated TGA
The HAG-HF is interfaced to the automated TGA and XRF units via a series of transport conveyors. The
HAG-HF unit extracts a test sample (2g) into a small ceramic crucible for full TGA analysis. A conveyor
then transfers the test sample from the HAG-HF to the TGA. The automated TGA set-up comprises of
four TGA units in which the sample is weighed upon entry into the first TGA (hygroscopic moisture) and
weighed at the completion of 1hr. The test sample is then rotated through the three remaining TGA’s
using the ‘automated transport mechanism. The residence time at each TGA unit is approximately 1hr.
The three remaining TGA’s are individually set at 425°C, 650°C and 1000°C. Each TGA remains at its pre-
defined temperature, and crucibles containing ore samples (as well as crucible blanks) are rotated from
furnace to furnace. Blank crucibles are used to calculate furnace factors. After ignition at 1000°C, the
crucible is removed, allowed to cool prior to being air-cleaned, ready for re-use. Figure 5 shows the four
blue TGA units with the automated transport mechanism suspended above the TGA units.
Figure 5 CLB automated TGA system
Reported shipment grades are based on a calculated iron (Calc Fe %) grade on a dry basis as detailed in
Equation 1 below
Calc Fe % = (100 – oxides % at 1000C] – LOI1000C
) / 1.4297 (1)
Where 1.4297 is the factor used to convert Fe to Fe2O
3.
It is imperative that LOI and all elemental analyses for an iron ore are determined accurately and
precisely. Since these are required to be corrected to a dry weight basis, accurate hygroscopic moisture
determination is of great importance. A simple method developed with the assistance of CSIRO (Division
of Minerals) enables the testing of several ores simultaneously for hygroscopic moisture in a single cycle
of a ‘Parcher’ apparatus (Figure 6). As detailed in ISO2596, hygroscopic moisture is determined by
heating a known quantity of ore at 105°C under a stream of nitrogen for 2hrs. Copper sulphate
pentahydrate has a known loss of 28.5-29.25% when heated under these conditions and can be seen in
the third position from the left in Figure 7. The copper sulphate pentahydrate standard was periodically
analysed with standards to ensure correct operation of the Parcher apparatus.
Figure 6 Parcher apparatus Figure 7 Ore samples place into parcher (prior to heating)
The first TGA oven to which the ore sample is introduced for hygroscopic moisture determination has
been designed to simulate as closely as possible the conditions used in the Parcher apparatus. A hollow
alumina lid allows each crucible at any position to be purged with nitrogen from above (Figure 8). Given
shipping operations are continuous and the TGA system will also operate continuously, a nitrogen
generator has been installed at CLB to supply nitrogen of >99.9% purity to TGA 1. However, during the
commissioning phase industrial grade bottles nitrogen banks were used to purge TGA 1.
Figure 8 TGA 1 lid inverted showing small nitrogen inlet holes
The importance of introducing a nitrogen atmosphere into the first TGA was examined during
commissioning. A series of 12 iron ore replicates was tested with the nitrogen on and off to examine the
effects of an inert atmosphere.
The hygroscopic moisture loss through parching the materials should be equivalent to the weight loss in
TGA 1 prior to the sample proceeding to TGA 2. A series of samples with various mineralogies were
tested over a period of six weeks to examine the accuracy and precision of the TGA 1 compared to the
Parcher method. In each test the recommendations of ISO2596 were adhered to in terms of
equilibration of samples in the laboratory atmosphere prior to analysis.
Using a variety of internal standards (which have been certified using only applicable ISO methods), the
performance of the remaining TGA’s for LOI at the various intermediate and final temperatures was
evaluated.
XRF Calibration
The CLB Port analytical laboratory will be expected to analyse 100-150 samples daily when full ship-
loading rates are achieved in late 2014. The two fusion induction furnace (HAG-HF) arrangement
commercially produced by Herzog shown in Figure 9 below is capable of fusing this volume of samples
Figure 9 CLB HAG-HF Fusion Unit
The HAG-HF fuses two samples simultaneously using two induction ovens, producing 32mm glass discs.
At Cape Lambert A and all other RTIO laboratories, 40mm discs are produced. The fused beads at CLB
were produced using 0.43g of sample to 4.4g of 12-22 lithium metaborate tetraborate flux. These
sample/flux weights are within the parameters specified for the production of 32mm fused discs in
ISO9516. The furnaces inductively heat the samples through a series of mixing and homogenization
steps prior to casting into a flat mold. Increased fusion times approaching 20minutes were utilized for
the fusion of calibration discs. All calibration samples were weighed and mixed with flux manually due to
the low weights of pure chemicals used rather than using the automated dispensing facilities within the
HAG-HF. Some 120 calibration beads were produced over a period of 10 days.
RTIO has traditionally used a series of pure standards with a commercially available synthetic calibration
standard (Syncal, XRF Scientific Batch Number 070708DB) to calibrate its XRF’s. The Syncal is mixed at
various ratios with silicon dioxide (min 99.99% SiO2) or iron oxide (min 99.998% Fe2O3) to increase the
number of calibration points for various minor and trace elements. Progressing from the mid 1990’s,
RTIO has increased the number of elements analysed from 14 to 24. The latest elements to be added
routinely for analysis have been Na and Cl. XRF Scientific introduced Na and Cl into the Syncal matrix in
2008 as NaCl on request from iron ore producers. As ISO9516 requires treatment of the pure
compounds to eliminate moisture and contamination, the Syncal standard was prepared by ignition in a
muffle furnace manually at 900°C for 20mins. This treatment of the Syncal had detrimental effects for
the As, Pb, K2O and Cl calibrations as detailed in the results section.
The inductive heating of the HAG-HF combined with the transport of molds and crucibles to and from
weigh cell–fusion-cleaning cell requires the platinum-ware to be resistant to high temperature and
robot gripping forces. Crucibles were manufactured using a 95%Pt/5%Au/0.2%ZrO2 alloy whilst
95%Pt/gold were used for the molds. The Pt/Au/Zr alloy is considerably harder than the normal Pt/Au
alloy. Temperature calibration of the furnaces is configured using an optical pyrometer which monitors
the exterior wall of the crucibles as they are heated.
Results
HAG-HF Fusion
The HAG-HF fusion equipment at CLB produces 32mm glass beads with the flux and sample weights of
4.4g and 0.43g respectively (ratio of 10.2:1). The HAG-HF fusion machine is programmed to pre-weigh a
small amount of flux followed by dosing 0.43g of sample, and finally topping with flux to achieve the
ideal sample ratio prior to mixing. Table 2 shows the sample and flux weights and ratio’s achieved over a
period of a single month running 60 unknown standards provided by Geostats. As can be seen in Table
2, all samples are well within the range of the weights specified in ISO9516, and the flux to sample ratio
is extremely small in variability.
Table 1 Flux and sample masses dispensed by the HAG-HF over a period of a month
The HAG-HF produces beads through high temperature induction, whereby the temperature of the melt
is monitored by an optical pyrometer measuring the outside crucible wall temperature between the
induction coils. Initial sulphur losses as per ISO9516 (duplicate CaSO4/Fe2O3 fusions) indicated a small
0.1% difference between duplicates on S count rates. (<2% is considered acceptable). After several
weeks of operation however, sulphur content of shipment sample duplicates showed concentration
differences of 0.01% as is indicated in Table 2.
Flux / Sample masses achieved from the HAGISO9516 / g Mean Mass / g Min Mass /g Max Mass / g
Flux 4.1 - 4.61 4.2653 4.2044 4.4049
Sample 0.41 - 0.44 0.4171 0.4134 0.4308
Target Flux/Sample Ratio 10.23
Min Flux/Sample Ratio 10.16
Max Flux/Sample Ratio 10.29
Table 2 Failure of sulphur on duplicate samples at the CLB laboratory
Sample ID S / %
Average Result 0.015
A11 0.013
A12 0.017
Average Result Fail S
B11 0.016
B12 0.006
Average Result 0.016
A21 0.016
A22 0.015
Average Result Fail S
B21 0.016
B22 0.007
The large difference between duplicate analyses was traced back to the fusion process, which was
observed to be exceeding the set temperature of 1050°C both visually and experimentally (verified using
a thermocouple). While the over temperature problem was resolved by adjusting the feed to the
induction coil, introduction of new platinum-ware still causes fusion temperature to significantly
increase. Sulphur losses were evaluated by preparing duplicate 10% CaSO4 90% Fe2O3 beads using old
and new crucibles (Figure 10). Fusion of the CaSO4/Fe2O3 mixture in a new and old crucible showed a
17% difference in sulphur count rates, while less than 2% is considered acceptable according to ISO9516.
The newer crucible showed the lower sulphur count rates, indicating overheating and sulphur loss
during the fusion process. We are currently investigating whether the temperature of the melt could be
monitored rather than the temperature of the platinum crucible wall to eliminate sensitivity of the
temperature control mechanism to the platinum surface. Alternatively temperature calibrations may
need to be performed when old crucibles are replaced.
Figure 10 - Old and new crucible exterior wall
Automated TGA
The automated TGA instrumentation determines the hygroscopic moisture, intermediate LOI’s at 425°C
and 650°C and final LOI at 1000°C of iron ore samples introduced. During commissioning the hygroscopic
moisture was evaluated using the automated TGA concurrently with determinations on the Parcher
apparatus at 105°C. Copper sulphate pentahydate sample losses during the initial commissioning of the
Parcher averaged 28.6% (expected 28.5-29.25%), confirming the Parcher to be accurate in hygroscopic
moisture determination. In October 2013, TGA 1 was changed from a routine ceramic based furnace to
an alumina block furnace in which there are multiple N2 purging points compared to 2 points in the
ceramic based TGA. The temperature we tested and adopted for hygroscopic moisture is 140°C. There
are 47 bench-top TGA’s in our Pilbara operations and previous test-work has indicated that most TGA’s
need to operate at a temperature of 140°C for the hygroscopic moisture step. Figure 11 shows the