Maggi Loubser. Group Chief Chemist, PPC Cement, South Africa Maggi.Loubser@ppc.co.za
Her career in X-Ray fluorescence Spectroscopy
started as a laboratory technician at the Atomic Energy Corporation of South Africa
in 1988. There she learned XRF in a process control laboratory.
During this time she studied part-time towards a BSc in Chemistry, which she
obtained in 1996 from the University of South Africa (UNISA), and a BSc(hons)
in 2003 from University of Pretoria. She has just completed a MSc in
Chemistry (University of Pretoria) with a dissertation entitled: Chemical and physical
aspects of Lithium borate fusion.
In 1994 she joined the Geology department of the
University of Pretoria (UP) where for the next fourteen years the X-Ray
laboratory was build up to a state of the art facility presenting annual short
courses in XRF and Representative Sampling and Sample preparation. Since 2005
Maggi has been presenter at the annual XRF Short Course at University of
Western Ontario in Canada.
She joined PPC Cement in November, 2008, where her
duties involve the development
of training material, training, mentoring and capacity building of the
chemists in the group. She has the
responsibility to ensure
world-class analytical laboratory practices at PPC’s
laboratories manage compliance to laboratory minimum standards and identify and
implement best practices in the analytical laboratories. She has presented numerous papers at
international conferences and national symposia and served on the executive of
the South African Spectroscopic Society and organising committee for CSIXXXIII and Geoanalysis 2009.
Representative sampling and the errors introduced in
sampling and sample preparation
Maggi Loubser.
PPC Cement, Group Quality Services,
Most of the analytical techniques currently used in process control environments are mature, stable techniques. In the past many manufacturing concerns typically focussed only on major elements. With the improved sensitivities and precision of analytical equipment along with the analytical demands related to the use of alternative raw materials and to environmental monitoring, the range of elements analysed has expanded from major elements to include minor and trace elements.
As analytical chemists we are prone to being arrogant regarding our analytical abilities being supported by standard quality control practises routinely implemented in the laboratory. In addition accreditation of a laboratory is often used as first line of defence when results are challenged by clients.
In reality it is often forgotten is that unless the sample is truly representative of the bulk it represents, the data are of little value. Typically the total sampling error is often hundred times larger than the analytical error.
Sampling theory originated with Pierre Gy in the nineteen fifties, the principles of which are still applicable even though analytical techniques have improved greatly. Most natural materials are compositionally and texturally heterogeneous and for a sample to be representative of the bulk, the correct size sample has to be taken, taking particle size and the concentration of minor and trace elements of interest into consideration. The problems do not end once the primary sample was taken, as only unbiased statistically reproducible mass reduction will ensure representative analytical specimens.
In this paper an overview of the theory of
sampling will be presented elucidating the different approaches used. The objectives of sampling in a process
control environment will be discussed and will be related to the theory of
sampling. Some of the practical problems
associated with implementing theoretical sampling principles in an industrial
environment will be described and some examples given on how appropriate compromises
are made.
Different methods of sampling will be discussed with real life examples from different manufacturing industries including cement, metallurgical and mining environments.
These concepts will be carried through to
basic sample preparation and the errors introduced at each subsequent step in
the process.