The sample preparation and assaying procedures discussed in this chapter apply equally to in-house

facilities as well as commercial laboratories. They are applicable to samples derived from
reconnaissance exploration through development drilling to mine operations.

The focus is on operations in metal mines; indeed, much worldwide mining activity at this writing is
concerned with exploration for and development of gold deposits. The requirements for the
preparation of samples of many gold-bearing materials are more stringent than those for many
other metallic deposits. Thus the practices described here generally fulfill the requirements for most
metallic mineral deposits.

The preparation and testing of coal samples is a specialized field. Most of the pertinent procedures
have been published by the American Society for Testing and Materials (ASTM), and they are
summarized in a later segment (5.4.3) of this chapter.

As is true for coal, the preparation and testing of samples of industrial mineral raw materials is highly
specialized. For many products, the preparation and testing procedures are endproduct sensitive;
that is, the preparation and testing procedures to be used are controlled by the properties desired
in the final product. One must attempt to select sample preparation and testing procedures that
will produce a product comparable to one produced by the actual process procedures to be used.
The book, Industrial Minerals and Rocks (Lefond, 1983), presents perhaps the best overall
background on a wide variety of industrial mineral products, although little specifically on sample
preparation and testing. Few additional general publications exist, and no specific guidelines can be
given here, other than to suggest that sample preparation and testing procedures be established
for individual deposits through frequent and close communication among the exploration and
mining staff, laboratory personnel, mill managers, marketing staff, and end users.

5.4.1 SAMPLE PREPARATION

Sample preparation is the process of converting samples of geologic materials from the large sample
collected in the field or mine into finely divided homogeneous powders suitable for chemical
analysis or other testing. This is accomplished by the mechanical reduction of pieces of rock to a
smaller particle size in a stepwise sequence, alternating with the reduction of sample volume or
mass by an unbiased splitting process.

Error can be introduced in many ways during sample preparation. As a consequence, attention to
detail and thorough cleaning of equipment between samples is necessary. The desired end result of
sample preparation is a rock powder that contains the elements to be analyzed in the same
concentrations and proportions as in the original sample received. The reduction in particle size will
be affected by many factors, including particle shape, hardness, specific gravity, lubricity,
malleability, residual moisture, and the quantity of clay minerals or organic matter present.

5.4.1.1 Sample Preparation Equipment

The equipment required to prepare mineral samples adequately for analysis depends to some
extent on the nature and quantity of the samples, and even on the climatic environment. In hot
desert environments, samples can be air dried adequately under the sun, and a laboratory with only
a handfull of small samples to crush and pulverize per day could do an adequate job with a manual
bucking board and muller.
Dryers:

Electric- or gas-fired ovens are used to remove moisture or free-standing water from samples before
crushing and pulverizing. An airflow is maintained through the oven to remove water vapor released
from the samples. For routine assay purposes, oven temperatures are usually maintained from 220
to 285F (104 to 140 C), the higher temperatures being used on clays, although the temperature
should not exceed 100F (37C) if mercury is to be determined. The submission of larger and
excessively wet drill cutting samples to high-volume minerals laboratories has started a trend
towards drying rooms or buildings equipped to dry large quantities of samples.

Crushers:

Jaw CrushersThe first stage of crushing normally is accomplished in a laboratory-sized jaw crusher.
Many designs still in production date back a half century or more, but the trend in redesigns or new
designs is towards roller bearings, short stroke, and a high reduction ratio. Some of these can be
choke-fed through a hopper with feed up to 4 in. (100 mm) across, and they will produce a 8-
mesh (2.4-mm) product in one pass while the operator is attending to other duties. Jaw crushers
have a relatively high productivity and reduction ratio and are generally easy to clean between
samples.

Cone CrushersNormally used in a second stage of crushing, cone crushers produce a uniform-sized
product with a smaller percentage of fines, considered better for metallurgical testing, and most
can crush to 10 mesh (2 mm). Cone crushers are not effective on clays, but work best on hard
siliceous materials. They have a relatively low productivity and are difficult to clean, especially after
clayey samples.

Roll CrushersAs an alternative to cone crushers for second-stage crushing, roll crushers have a
higher productivity and can produce a 10-mesh (2-mm) product. However, they produce a poorly
sorted product with a higher percentage of fines and are noisy, dusty, and difficult to clean. The
feed to roll crushers must be sized to 3/8 or 1/2 in. (10 or 12 mm), and the feed rate must be
controlled to prevent choking.

Hammer MillsHammer mills have a high productivity and the potential to produce a product
suitable for splitting and pulverizing in one pass from feed as large as 4 in. (100 mm). However, they
are extremely noisy, dusty, hard to clean, and subject to excessive wear when processing tough
siliceous materials. Their product is not well sorted, typically consisting of a large percentage of fines
with a small percentage of very coarse fragments. Hammer mills are more often used to crush clays,
limestone, coal, and similar softer materials.

Splitters: At some point in the sample size-reduction process, it becomes impractical and
unnecessary to further reduce the particle size of the entire sample. The sample volume is then
reduced by half or more, depending on the procedure selected as appropriate for the material at
hand, by using a sample splitter.

Riffle SplittersThe riffle splitter, or Jones splitter, is most commonly used for sample size
reduction. The technique of splitting must be carefully monitored to insure that statistically valid
splits are taken. Splitter chutes should be at least three times as wide as the diameter of the largest
particles in the sample, and the delivery pan should be no longer than the distance across all of the
chutes. The sample should be evenly distributed along the length of the delivery pan and should be
poured along the center line of the splitternot against one side or the other. The rate of pouring
must be slow enough to avoid choking the chutes. The splitter must be cleaned between samples.

Rotating Sectorial SplittersThe most effective splitter, in terms of sampling error, is the rotating
sectorial splitter (Allen and Khan, 1970). Various designs exist, most being adapted to splitting the
product from rotary drill rigs and thus larger than appropriate for use in the laboratory. Shop-made
laboratory-sized versions of the rotating sectorial splitter are in use in some sample preparation
facilities. One design consists of a rotating circular table driven by a variable-speed motor at about
10 rpm. Four plastic cartons of 1-pt or 1-qt (1/2 or 1-L) capacity rest on the rotating table. The gaps
between the cartons are covered by pieces of angle iron. The sample is fed from a feed hopper by a
vibratory feeder.

Pulverizers: Once the sample is reduced to an appropriate weight (typically 1/4 to 1 lb, or 100 to
500 g), it is then pulverized to a nominal 100 to 200 mesh (150 to 75 m). Two basic types of
pulverizers are in common use: the plate or disk pulverizer and the vibratory ring mill.

Plate PulverizersPlate pulverizers reduce the particle size of samples through a shearing action as
the sample passes between a fixed and a rotating plate. Two basic designs exist: horizontal shafts
and vertical shafts. Some are designed only for intermittent use, whereas the more recent versions
are extremely rugged and built for high volume work. Controlling the spacing between the plates
(and thus the product size) has been a problem, and has required almost constant attention by the
operator, but recent models of plate pulverizers have better mechanical or even pneumatic controls
that maintain a constant plate separation. Plate pulverizers have a high productivity, they can
pulverize a large sample, and are easily cleaned. However, they are dusty and their product is
inhomogeneous, requiring thorough blending before assaying. Mikli (1986) does not recommend
plate pulverizers for the final pulverizing of nuggety gold ores, as the plate pulverizer does not
significantly reduce the particle size of gold nuggets. Plate wear is high, requiring frequent changing
and refinishing or replacing of plates.

Vibratory Ring MillsTwo basic versions of vibratory ring mills are made. The first consists of a steel
bowl with lid, the bowl containing a cylindrical steel puck plus one or two steel rings that surround
the puck. The crushed sample is placed in the open space between the walls of the bowl, the puck,
and the rings. The bowl is clamped in a housing which is made to oscillate around a vertical axis by
an electric motor carrying an eccentric weight on its shaft. The oscillatory motion causes the puck
and rings to revolve in a planetary pattern inside the bowl, thus grinding the sample.

The second version of this mill also consists of a bowl with lid but only a single grinding element, a
discus- or flying saucer-shaped oblate spheroid of steel with a flat rim. The center of gravity of this
steel element is off center so that when the bowl oscillates, the grinding element revolves in a
planetary pattern within the bowl. Bowls of this style have a capacity of 1.8 to 11 lb (800 g to 5 kg)
of sample, whereas the puck and ring bowls range from 0.1 to 0.9 lb (50 to 400 g) in capacity.

Vibratory ring mills exhibit the following advantages: they require no operator adjustment, produce
a relatively homogeneous product that requires no further blending, create Little dust, have a low
noise level (because of a supplied noisesuppressant cabinet), and have a moderate productivity of
some 12 to 20 samples per hour. The productivity depends on the character of the samples, with
the lower productivity being achieved on high-clay samples, which require less than 1 min to
pulverize, but several minutes cleanup time between samples. Vibratory ring mills have been
described by Mikli (1986) as the only type of pulverizer that actually can reduce the particle size of
gold nuggets. However, to reduce the particle size of coarse nuggety gold or to pulverize a 4.5-lb (2-
kg) or larger sample thoroughly to 80 mesh (177 m) or finer can require from 10 to 30 min and
result in excessive bowl wear.

Blenders and Pulp Splitters: Pulps prepared on plate pulverizers require blending. The most common
approach (although not the most effective) is to roll the pulp on a rubberized cloth. Taggart (1945,
Sec. 19, p. 71) gives instructions for proper rolling: Rolling is accomplished by drawing the corners
of the cloth horizontally toward diagonally opposite corners, causing the sample to roll over and
over on itself. If the corner is lifted instead of drawn horizontally, the sample merely slides along the
surface of the cloth and no mixing occurs. Plastic sheeting should not be used for blending because
of the static charges that build up, causing retention of some particles and difficulty in cleaning.
Some question exists as to whether or not rolling actually is very effective in homogenizing a pulp.

A far superior alternative to rolling is the use of a mechanical blender. Large V-shaped blenders have
been used to blend bulk samples for use as reference standards, but small sizes suitable for blending
assay pulps are not common. Individual sample blenders are slow. A multisample mechanical wheel
blender that meets the productivity requirements of a high-volume minerals laboratory has been
described by Gilbert (1987).

Mechanical splitters for use on pulps exist, but they are expensive, slow, and require extensive
cleaning between samples. A more effective approach is to roll the pulp on a rolling cloth into a
sausage, flatten the sausage to the height of the scoop being used, and then cut several
increments from the sausage using a flat-bottomed vertical-sided scoop until the required weight
has been withdrawn.

5.4.1.2 Typical Preparation Methods

The procedure given in Table 5.4.1 is typical of many minerals laboratories. It is inexpensive and
assumes the sample is relatively homogeneous and the ore particles small.

People experienced with sampling for nuggety gold (or any ore with a potentially large nugget
effect) generally collect larger samples and prefer a sample preparation procedure similar to that
outlined in Table 5.4.2.

5.4.1.3 Selecting a Sample Preparation Procedure

Virtually every mineral deposit has its own eccentricities, and an individual sample preparation
procedure must be developed for each, unless the deposit is known to be fine-grained and relatively
homogeneous. A safe sample preparation procedure is given by Royle (1989, p. 40) based on a
method originally developed by Gy (1977). Similarly, Pitard (1987) presents sampling nomographs
that enable one graphically to analyze an existing sample preparation process and to develop an
optimum protocol. Pitard markets programs for personal computers that describe tests to perform
on gold ores of unknown characteristics and enable one to plot sampling nomographs from which
an appropriate sampling protocol can be developed. In their Chapter 1, Sampling, Ingamells and
Pitard (1986, pp. 1-84) present a good review of the necessity for intelligent sampling, both before
and during the sample preparation and analytical stages.

5.4.2 ASSAY METHODS

Two basic classes of assay methods are available: geochemical and quantitative. Geochemical
procedures typically are used in prospecting and the early stages of exploration when results of high
accuracy and precision are not as necessary, but low levels of detection are required. Quantitative
procedures are used during exploration drilling, sampling and analysis for ore reserve estimation,
and subsequent stages of mine development and operation. Geochemical methods of analysis in
the past have been considered semi-quantitative, but many of these now approach quantitative
methods in accuracy and precision, and they offer the advantage of considerably lower levels of
detection.

Whether establishing an in-house laboratory or selecting a commercial laboratory, there is no

substitute for a chemist with experience in the analysis of earth materials. The wide variety of
materials to be analyzed, constituents to be determined, and matrix compositions to be
encountered pose a myriad of challenges to the minerals analyst. Many constructive comments on
the art of geochemical analysis are given by Ingamells and Pitard (1986).

5.4.2.1 Wet Chemistry

Wet chemical methods of analysis have been and probably will continue to be the mainstay of the
minerals laboratory, in contrast to nondestructive instrumental methods. Wet chemical methods
have undergone significant changes with time, perhaps the most dramatic following the evolution
of microelectronics and computers that allowed the development of instrumental methods of
measurement.

Classical Volumetric and Gravimetric Methods: Volumetric and gravimetric methods of analysis now
are used principally for the determination of the higher concentrations of various elements in ores
and concentrates. Typical examples are the determination of calcium in limestone; iron in iron ores;
ferrous iron in most mineral materials; copper, lead, zinc, etc., in ores and concentrates; and
uranium in ores and yellow cake. Most classical methods of analysis are well-established. Many are
given in references such as Furman (1962), Young (1971), and Dolezal et al. (1968).

Colorimetric Analysis: Colorimetric methods of analysis were used extensively in the mineral
industry in the mid-1900s, especially in trace analysis for geochemical exploration purposes. More
recently, colorimetric methods of analysis have been supplanted by instrumental methods such as
atomic absorption and inductively coupled plasma spectrometry.

Colorimetric methods of analysis are based on the development of a colored compound in solution
by the reaction of the element in question with a specific chemical reagent. One of the principal
colorimetric methods still in common use is the thiocyanate assay procedure for molybdenum in
ores and concentrates (Ingamells and Pitard, 1986, p. 523). Simple colorimetric methods of analysis
useful in geochemical exploration in remote areas are ammonium citrate-soluble heavy metals
(Bloom, 1955) and cold acid-extractable copper (Canney and Hawkins, 1958). Colorimetric analytical
methods for tungsten and molybdenum produce results comparable with atomic absorption
procedures, but they are not as rapid. Instrumental Analysis: Instrumental methods of analysis have
dominated minerals laboratories since the development of the atomic absorption
spectrophotometer in Australia in the mid-1950s and the inductively coupled plasma emission
spectrometer in the 1970s.

The atomic absorption spectrophotometer (AA) enables the analyst to measure the concentration
of some cations down to a fraction of a part per million and with specialized attachments down to
a part per billion. The AA instrument generally is specific for the element selected, although there
are various interferences and operational nuances that the analyst must be aware of and either
avoid or compensate for.

Principles of AA operation are given by Slavin (1968). Many methods of analysis are given in
publications of the US Geological Survey and the Geological Survey of Canada as well as the Atomic
Absorption Newsletter (now titled Atomic Spectroscopy) published by the Perkin Elmer Corp.

Inductively coupled plasma spectrometers (ICPs) were developed as modifications of existing direct-
reading emission spectrometers or atomic absorption spectrophotometers. The direct readers
became ICPs capable of measuring 30 or more elements simultaneously, whereas the AAs became
ICPs that measure a number of elements sequentially, commonly utilizing more than one
wavelength per element. Thus the simultaneous instrument is faster for routine work on similar
materials, whereas the sequential instrument is more versatile. Most ICP instruments utilize the
high energy of an argon plasma to excite atoms of various elements in a sample solution that is
aspirated into the plasma. The excited atoms emit light that is passed through a spectrometer,
wherein the energy of light emitted at various wavelengths is measured electronically and
converted into the concentration of each element in the sample solution.

ICPs are supplanting AAs in many high-volume mineral laboratories because they allow the
determination of a large number of elements in a short period of time, whereas the AA determines
one or at best a few elements simultaneously. The accuracy and precision of ICP instruments and
AAs are comparable, and the capital costs, considering the productivity, probably favor the ICP. The
ICP can achieve detection levels as low or perhaps slightly lower than the AA, and in addition the
ICP instruments have a linear response over some five to seven orders of magnitude of
concentration of the element in question, a much broader operating range than the AA.

ICPs are used to determine both trace elements and major elements in most earth materials. A wide
variety of digestion and separation procedures are used to prepare the sample for ICP
measurement, many of them similar to AA digestion procedures. A large literature exists.

Several other instruments are used for specific determinations in minerals laboratories. Specific ion
electrodes are used to measure the concentration of cyanide in dilute solutions as well as the
content of fluorine, chlorine, and several other constituents in earth materials, and the fluorometer
is the standard instrument for determining trace amounts of uranium.

5.4.2.2 Fire Assay

A fire assay is a chemical fusion method for separating, concentrating, and measuring the content
of gold and silver in exploration samples, ores, and concentrates. The pulverized sample is weighed,
mixed with a lead oxide-alkali carbonate flux and a small amount of reducing agent such as flour,
and fused in a fire-clay crucible. The reduced lead collects the precious metals as it settles down
through the melt. The molten charge is then poured into a mold to cool. The lead sinks to the bottom
of the mold and is broken from the glassy slag when cool. The precious metals are separated from
the lead by cupellation. The lead button containing the precious metals is placed in a cupel of bone
ash or magnesia which is heated in the furnace under oxidizing conditions. The cupel acts as a semi-
permeable membrane allowing the lead oxide formed from the button to be absorbed into the
cupel, leaving the precious metals in a tiny bead on the surface of the cupel. The gold and silver are
separated chemically and quantified gravimetrically or by instrumental analytical techniques.

Fire assay is the standard method of the industry. Details, theory, and variations of the method are
described in references such as Bugbee (1940), Haffty et al. (1977), and Heady and Broadhead
(1976).

5.4.2.3 Nondestructive Analysis

X-Ray: The X-ray diffractometer is used in some wellequipped minerals laboratories to determine
the mineral species present in a sample by recording their characteristic crystallographic patterns.
The instruments are not extensively used in mine and project assay work.

X-ray fluorescence spectrometry (XRF) is used for the rapid analysis of silicate rocks and the routine
determination of the concentration of selected elements in exploration samples, ores, concentrates,
and mill products. XRF analysis is most satisfactory when applied to a continuing series of samples
of similar matrix. The method is subject to matrix effects and interelement interferences, most of
which can be corrected for in the comprehensive computer software that accompanies all modern
instruments. XRF instruments require for calibration previously analyzed standards similar in bulk
composition to the unknowns. The instruments are capable of excellent precision, but without
proper calibration and intelligent operation, they can be rather inaccurate.

Radiometric Analysis: Radiometric analysis is used to measure the concentration of uranium in its
ores by simultaneously counting the beta and gamma radiation emitted by the sample.
Multispectral radiometric analysis provides a determination of the concentrations of the radioactive
elements uranium, potassium, and thorium, all of which will provide a positive response to a simple
gamma ray detector such as a Geiger-Mueller counter or a scintillometer.

Neutron Activation Analysis: Neutron activation analysis (NAA) is available principally through
commercial laboratories with access to nuclear reactors. Most of the laboratories offering
commercial NAA analyses of earth materials at reasonable cost are Canadian-based. NAA is
indicated when (1) a conventional technique does not have acceptable limits of detection for the
element of interest, (2) the sample is unique and cannot be consumed in analysis, (3) only a small
quantity of sample material is available, or (4) conventional methods of analysis are unacceptable
because of interferences or inherent instrumental errors. The sample typically is pulverized, loaded
into a plastic capsule (rabbit), and introduced to a reactor in which it is bombarded by neutrons.
After recovery of the capsule from the reactor, the radioactivity induced in the sample is measured
and analyzed, thus giving a measure of the concentration of each element present in the sample.
NAA currently has much application in the analysis of vegetation or mull for trace amounts of gold,
and in the analysis of the platinum-group elements and the rare earth elements.

5.4.3 COAL PREPARATION AND ANALYSIS

The procedures used in the preparation of coal and coke samples for analysis are similar to those
for rock samples. The ASTM standard method of preparing coal samples is No. D 2013-86, published
in the Annual Book of ASTM Standards, Pt. 05.5 (Anon., 1989). The principal difference in coal
sample preparation is the use of lower temperatures and, indeed, even air drying to preclude
oxidation as much as possible. Crushing and pulverizing is accomplished with much the same
equipment as used on rocks. However, samples are pulverized only to 60 mesh (250 m), and only
0.1 lb (50 g) is retained for analysis.

A variety of tests are available for coal samples. Some of these include sieve analyses, washability,
Hardgrove grindability, moisture, sulfur, ash content, carbon and hydrogen content, and calorific
value. Two of the most common determinations are proximate and ultimate analyses. A proximate
analysis, as described in ASTM Standard Method D 3172-73, covers the determination of moisture,
volatile matter, and ash, as well as the calculation of fixed carbon. According to ASTM (Anon., 1989,
Method D 3172-73, p. 1), the results of a proximate analysis are used to establish the rank of coals,
to show the ratio of combustible to incombustible constituents, to evaluate the coal for
beneficiation and other purposes, and to provide a basis for buying and selling.

An ultimate analysis, according to ASTM Standard Method D 3176-84, p. 1 (Anon., 1989), when
tabulated along with a proximate analysis, provides the data for a cursory valuation of coal for use
as a fuel and coke for metallurgical purposes. An ultimate analysis includes the determination of
carbon and hydrogen as well as sulfur, nitrogen, ash, and the calculation of oxygen by difference.
Typically, moisture is reported as well. The procedures for an ultimate analysis are also specified by
ASTM (Anon., 1989).

In addition to the analytical procedures described previously, data on the major, minor, and trace
elements in coal and coke ash often are of use in the evaluation of coal quality. Accordingly, methods
for these determinations are presented in ASTM Standard Test Method D 3682-87, Major and
Minor Elements in Coal and Coke Ash by Atomic Absorption, ASTM Standard Test Method D 4326-
84, Major and Minor Elements in Coal and Coke Ash by X-Ray Fluorescence, and ASTM Standard
Test Method D 3683-78, Trace Elements in Coal and Coke Ash by Atomic Absorption. The atomic
absorption measurement of the major and minor elements is accomplished after a lithium
tetraborate fusion of the sample followed by hydrochloric acid dissolution of the fusion product,
whereas the trace elements are determined following a mineral acid dissolution process.

5.4.4 QUALITY CONTROL

Most laboratories have an adequate quality control program covering their analytical work. Some
commercial laboratories actually provide results of check assays with their reports of analysis. It
would appear, however, that few laboratories include sample preparation in their quality control
system; yet, as Royle (1989) has commented, Wait until you see the numbers that show all the
horrible things that happen in sample preparation of gold samples!

During sample preparation, maximum particle size readily can be monitored by screening at various
stages as the material is crushed and pulverized. Other parameters are more easily checked by
actually analyzing the material.

Analysis of several pulps prepared from different splits of the crushed bulk sample will provide data
on reproducibility of splitting. Analogously, the homogeneity of a pulp can be checked by taking a
number of replicate portions for analysis from the same pulp and calculating the precision of
replication. A simple procedure to evaluate the effectiveness of sampling drill cuttings is presented
by Schwarz (1989).

Analytical precision and accuracy are best established and maintained through the use of reference
standard samples and analytical control samples (Hill, 1975). Certified standard samples of mineral
materials are available from some governmental agencies and professional or trade groups in
several countries (Abbey, 1983; Flanagan, 1986). Such standard materials are expensive and limited
in quantity. The elements present and concentration ranges in certified standards may not
adequately cover the character of materials that the laboratory is engaged with; nevertheless, such
standard materials may help to provide initial calibration. Matrix-matched control samples to
provide batch-to-batch and day-to-day calibration and quality control in the laboratory can be
prepared and blended from the reject samples from the project. The analytical development of
these in-house control samples can be calibrated to the certified standards. Ideally, every set of
analyses produced by the laboratory should have one or more control samples and duplicates
included for quality control and assurance. Some governmental agencies that contract out a large
quantity of sample preparation and analytical work arrange for 3 out of each 20 samples to be
controls or replicates. Those more mathematically inclined can find much of value regarding
analytical quality control in the laboratory in the Quality Assurance Handbook compiled by the
Center for Analytical Chemistry of the National Bureau of Standards (Anon., 1987). ASTM gives
guidelines for evaluating laboratories performing analysis of coal and coke (Anon., 1989, p. 357).

When the results of analysis of the controls do not agree with the established values within
acceptable limits, the entire group of assays should be rejected, the problem identified and
resolved, and the group of samples rerun. The timely realization and resolving of sample preparation
and analytical problems is the essence of quality control.