British Geological Survey

TECHNICAL REPORT WG/92/30

Mineralogy and Petrology Series

cwd9

Industrial Minerals Laboratory Manual

FLAKE GRAPHITE
C J Mitchell

Mineralogy and Petrology Group

British Geological Survey
Keyworth
Nottingham
United Kingdom NG12 5GG

British Geological Survey

TECHNICAL REPORT WG/92/30
Mineralogy and Petrology Series

Industrial Minerals Laboratory Manual

FLAKE GRAPHITE
C J Mitchell
A Report preparedfor the Overseas Development Administration under the
ODADGS Technology Developmentand Research Programme, Project 91/1

ODA Classification:
Subsector: Others
Subject: Geoscience
Theme: Mineral resources
Project title: Minerals for Development
Reference number: R554
1

Bibliographic reference:
Mitchell, C J
Industrial Minerals Laboratory Manual:
Flake Graphite
BGS TechnicalReport WG/92/30
Subject index:
Industrial minerals, flake graphite,
mineral processing, laboratory techniques
Cover illustration:
A laboratory frothflotation cell
is used to prepare flake graphite
concentrate during beneficiation trials

CONTENTS

Page
1. INTRODUCTION

2. GEOLOGICAL OCCURRENCEAND BACKGROUND

3. USES AND SPECIFICATIONS

4. MINERALOGICAL CHARACTERISATION

5. LABORATORY PROCESSINGTRIALS

11

6. CASE HISTORIES

18

REFERENCES

22

APPENDICES:
1. Specifications of graphite for different applications

23

2. Particle size and carbon content data

for commercial graphites

24

3. Graphitic carbon assay

25

4. Sample preparation andfroth flotation procedure

28

5. Air classification procedure

29

6. Graphite grade, yield and recovery calculation

30

7. Chemistry of the Lac Knifegraphite, Canada

31

Preface
Industrial mineral raw materials are essential for economic development.
Infrastructure improvement and growth of the manufacturing sector
requires a reliable supply of good quality construction minerals and a
wide range of other industrial mineral raw materials.
Although many less developed countries have significant potential
industrial mineral resources, some continue to import these materials to
supply their industries. Indigenous resources may not be exploited (or
are exploitedineffectively) because they do not meet industrial
specifications, and facilities and expertise to carry out the necessary
evaluation and testwork are unavailable. Unlike metallic and energy
minerals, the suitability of industrial minerals generally depends on
physical behaviour, as well as on chemical and mineralogical properties.
Laboratory evaluation ofteninvolves determination of a wide range of
inter-related properties and must be carried out with knowledge of the
requirements of consuming industries. Evaluation may also include
investigation of likely processing required toenable the commodity to
meet industry specifications.
Over the last 10 years, funding from theOverseas Development
Administration has enabled the British Geological Survey to provide
assistance to less developed countries in the evaluation
of their industrial
mineral resources. This seriesof laboratory manuals sets out experience
gained during this period. The manuals are intended tobe practical
bench-top guides for use by organisations such as Geological Surveys
and Mines Departments and are not exhaustive in their coverage
of every
test and specification. The following manuals have been published to
date:
Limestone
Flake Graphite
Diatomite
Kaolin
Bentonite
A complementary series of Exploration Guides is also being produced.
These are intended to provide ideas and advice forgeoscientists
involved in the identification and field evaluationof industrial minerals
in the developing world. The following guide has been published to
date:

Biogenic Sedimentary Rocks

A J Bloodworth
Series Editor
D J Morgan
Project Manager

1
~

~~

~~~

~~

Industrial Minerals Laboratory Manual

Flake graphite
1

INTRODUCTION

This manual concentrates on the laboratory assessment

of flake
graphite, though other formsof graphite are considered briefly.
Overviews of the geological occurrence and the uses and applications of
graphite are given. These are supportedby more detailed informationin
the Appendices. Procedures for mineralogical assessment and
laboratory processing trialsof graphite-bearing rocksare discussed,
along with those for evaluation of graphite concentrates produced in the
laboratory. Detailed step-by-step proceduresfor laboratory processing
trials are givenin the Appendices. Case histories detailing two typical
laboratory investigations with worked examples are also included.
This manual is one
of a series produced as
part of the BGS/ODA R&D
Project Minerals for Development.
Graphite is oneof two naturally occurring allotropesof crystalline
carbon, the other being diamond. In contrast to diamond, graphite
is
soft, greasy to the touch, and soils fingersand paper. Table 1
summarizes the principal properties
of graphite whilst Figure1
illustrates the crystalline structureof graphite. Graphite generally
occurs in one of three forms: microcrystalline or amorphous; crystalline
lump or vein; and crystalline flake. The characteristics
of the different
2. The major consumerof flake
forms of graphite are listed in Table
graphite is the refractories industry. Itis also used in friction materials,
lubricants, foundry core and mould washes, and may undergo property
modification to form expanded graphite and graphite foil.

I
I

Figure 1. Arrangement of carbonatoms in the hexagonal (2H) unit

1987).
cellofgraphite.(AfterKraussetal.,
Mineralogy and Petrology Group, British Geological Survey0 NERC 1992

Flake graphite

Table 1. Properties of graphite

~ _ _ _

Hardness 0.5 - 1 (Mohs

scale).

Colour

Dark steel
grey
iron
toblack;
metallic
lustre;
grey
black
to streak.

Morphology

Hexagonal
crystal
system
(rhombohedral
symmetry);
usually
platy,
occasionally
acicular; perfect basal cleavage, commonly striated, occasionally twinned.

Meltingpoint

Decomposesat 600C inoxidisingconditionsandmeltsat

3550C in non-oxidising
conditions; infusiblein most common fluxes; vaporisation temperature4500C.

Miscellaneous

Good thermalandelectricalconductivity;chemicallyinert;greasytexture(natural
lubricant); low thermal expansion and shrinkage; compressible and malleable; low
absorption coefficientfor X-rays; nontoxic; generally not 'wetted by metals.

Mineralogy and Petrology Group, British Geological Survey0 NERC 1992

Density

- 2.26

Composition

2.09

Flake graphite

2. GEOLOGICAL OCCURRENCE AND BACKGROUND

Graphite generally occurs as a result
of metamorphism (regionalor
contact) of organic matterin sediments. Flake graphiteis assumed to be
in organic matter. As
derived from fine-grained sediments rich
metamorphic grade increases, carbonaceous material converts to
'amorphous' graphite. Flake graphite formsfrom its amorphous
precursor in rocks at or beyond amphibolite grade metamorphism
(Landis, 1971). Vein graphite is assumed to form by partial
volatilization of graphite and subsequent recrystallization during regional
granulite and/ or charnockite facies metamorphism. Amorphous
graphite is generally considered to have originated
by thermal or
regional metamorphismof coal or carbonaceous sediments.
Positive vanadium and nickel anomalies and negative boron anomalies
if geochemical survey dataare
are possible signatures for graphite
available. The presenceof sulphides and trace amountsof uranium may
also be an indicator.

Table 2. Summary ofcharacteristics of the differentformsofgraphite

Vein

Flake

> 150 pm; fine <150 pm

coarse crystals
Mostly >4 cm

Microcrystalline
<70 pm

metamorphism

Epigenetic; regional
metamorphism

Syngenetic; contact and

/ or
regional metamorphism

ore

5 - 30% graphite;
stratabound, tabularor lenses

98%+ graphite; veins Seams, often folded and

faulted
and fractures

Product grade

75 - 97% graphite

98 - 99.9% graphite

60 - 90% graphite

Main uses

Refractories, brake
linings, lubricants and
batteries

Carbon brushes,
brake linings and
lubricants

Refractories, steel
industry, paint and
batteries

sri Lanka.

China, S Korea,
Czechoslovakia, Austria
and N Korea

Description
coarse
Crystalline
flakes:
regionalSyngenetic: Origin

. ...

,.,

India,Brazil,China,Major
Madagascar,
pdUCl3I.S

Germany,

Norway, Canada and

Zimbabwe
Adapted from Fogg and Boyle (1987)

Mineralogy and Petrology Group, British Geological Survey 0 NERC 7992

Flake graphite

Flake graphite occurs asflat, plate-like crystals, with angular, rounded

or irregular edges, disseminated throughout originally carbonaceous
metasediments. The host rocks include quartz-mica schist, feldspathic
in
or micaceous quartzite and gneiss. Flake graphite may also occur
metamorphosed carbonate rocks, though these occurrences are currently
of little economic significance. Flake graphite deposits are usually strata
bound, with individual bedsor lenses from30 cm to more than30 m
thick and extend for lengths of2 kilometres or more. Ore bodies are
normally tabular, occasionally lenticular, and occur locally as irregular
bodies in the hinge zones of folds. Most economic deposits of flake
graphite are of Archean to late Proterozoic age. These rocks may
contain up to 90% graphite, although 10 - 15% graphite is a more
typical ore body grade.
Graphite flakes rangein size from 1 to 25 mm, with an average sizeof
2.5 mm. Commercially flake graphite is divided intocoarse (150 - 850
pm in diameter) andfine (45 - 150 pm in diameter)flake. Fine flake
may be further subdivided into medium flake(100 - 150 pm), fine flake
(75 - 100 pm) and powder (-75 pm). Impurities include minerals
commonly found in metasediments, usually quartz,
feldspar, mica,
amphibole, garnet and calcite, with occasional amphiboles, pyrrhotite,
pyrite and magnetite
Vein-type ('high-crystalline' or'lump') graphite is found as interlocking
aggregates of coarse and/or microcrystalline platy
or (less commonly)
acicular graphite. The veins are hosted
by igneous and metamorphic
rocks (gneiss, schist, quartzite and marble).
'Amorphous' (microcrystalline) graphite is formed
of aggregates of fine
graphite crystals which give the material
a soft, black, earthy
appearance. This material is usually hosted by quartzites, phyllites,
metagreywackes and conglomerates. 'Amorphous'graphite is defined
as being finer than40 pm in diameter, but some trade statistics define
the upper limit at70 pm (40 - 70 pm is the limit of resolution
of the
human eye). Deposits with grades over80% carbon are considered to
be economically viable.

Mineralogy and Petrology Group, British Geological Survey0 NERC 1992

Flake graphite

3. USES AND SPECIFICATIONS

Specifications for graphite used in a wide variety

of applications are
given in Appendix 1. Particle-size data and carbon contentsof a large
number of commercial graphite products from major graphite producing
2. The requirements of industries
countries are given in Appendix
which use graphite as a raw material are discussed below.A brief
review of current economic trends within
the graphite marketis also
given.
3.1 Refractories

The major consumerof flake graphite is the refractories industry. Flake

graphite improves the thermal shock and erosion resistance
of refractory
products. Flake orientation enhances the strength
of the refractory.
Graphite mustbe employed in reducing conditions because of its low
oxidation temperature. Accordingly, graphite-bearing refractories such
as magnesia-carbon bricks are extensively employed as linings
in basic
oxygen converters, electric arc furnaces and steel ladles. Magnesia
graphite bricks usedin electric arc steel makingin the UK contain 15 20% graphite. Japanese magnesia graphite bricks may contain
20 -25%
graphite which doubles the lifespan
of the refractory. Alumina graphite
bricks are generally used in continuous casting processes and torpedo
ladles. Crucibles, for handling molten metal in foundries,
are made
from either clay graphite (large flake/85% carbon)
or Sic graphite (fme
flake/80% carbon) mixtures. Flake and amorphous graphite are used in
washes on foundry moulds and cores. Sulphides are undesirable in
graphite used as a foundry sand additive.
Flake graphiteis also used to make expanded graphitefor insulation in
steel production andis added to polyurethane foam
as a flame retardant.
Expanded graphite is produced
by using acid (chromic and sulphuric) to
weaken the bonds between graphite plates and this allows expansion on
heating to about 700C. 'Amorphous' graphiteis used as a foundry
additive, in refractory bricks andin plastic and castable refractories.
Graphite powder can
be used as a thermal blanket for the continuous
casting of steel.
3.2 Non-refractory uses

Flake graphiteis used in friction materials (brake

/ clutch linings),
lubricants and in the manufacture
of graphite foil and long-life alkaline
batteries. Brake linings use mixturesof natural (flake and vein) and
synthetic graphite(60:40 blend) in order tolower the wear rate, and
form up to 15% of the lining. A wide rangeof lubricants employ
graphite (flake and vein), including high-temperature bearings and oils
Mineralogy and Petrology Group,British Geological Survey 0 NERC 1992

ow - OZZ$

ooT - oooT$
O O O T - OLL$
009 - OZS$
0 9 - SZ$
009-OoV$

OOOT
002 T

- osV$
- OS9$

OoSI - OSL$

3 %SS/OS
3 %66/L6
3 %L6/S6
3 %26/06
3 %SSlOS
3 %S6/08
3 %06/S8
3 %06/S8
3%S6/26

Flake graphite

4. MINERALOGICAL
CHARACTERISATION

Mineralogical characterisationof graphite-bearing rocks with economic

potential shouldaim principally to determine the graphitic carbon content
andgraphiteflakesize.These two properties will determinethe
.
economic valueof the graphite. Mineralogical characterisation also
provides the basis for the planning
of laboratory beneficiation trials.
An initial thin-section examinationof graphite-bearing rocks is useful
for determining the size and shape
of the graphite flakes, their
relationship to other minerals present, and likely liberation size. A full
petrographic descriptionof a typical graphite schistis given as part of
the graphite case history in Section
6 of this manual. A sirmlar
examination of poorly-indurated weathered graphitic rock and graphite
rich soil can be carried out by binocular microscope. Identification of
non-graphite minerals maybe confirmed by X-ray diffraction(XRD).
Landis (1971) proposed a graphitisation series between flake graphite
and amorphous graphite based on XRD peak width and shift.
Comparison by BGS laboratories of XRD data from selected
commercial crystalline and amorphous graphite samples (Figure
2)
shows a limited amount
of peak broadening in thelatter material, butno
evidence for peak shift. It is likely that particle size differences between
the two materials will account for this contrast
in peak width. Care
should thereforebe taken not to confuse trade descriptions
of graphite
such as amorphous and crystalline with strict mineralogical terms.
Determination of the carbon contentof graphite-bearing rock (graphite
assay) is generally carried outby a simple loss-on-ignition (LOI) test.
The laboratory procedure is given in Appendix
3. Graphite content is
determined as the percentage weight loss
of a sample heated inair to
1000C. LO1 should be performed on duplicate samples to monitor the
precision of the analysis.
Graphite-bearing soils contain clays and iron oxyhydroxides which will
also lose weight on heating. Other minerals present in graphite-bearing
be accounted for.
rocks may also show this behaviour and must
Microscope and/or XRD examinationwill indicate the presenceof any
minerals which might interfere with the determination of graphitic
carbon by LOI. A modified LO1 test has been devised to take this into
consideration. A list of some possible minerals which may cause
problems is given in Appendix 3.
Thermogravimetry (TG) affordsan alternative methodfor determination
of graphitic carbon content. The technique
is of particular value if other
minerals which evolve volatiles on heatingare present. The sample is
under nitrogen,by which
heated at a controlled rate 900C
to
temperature any clays or carbonates present will have decomposed. The
instrument records weight losses in the sample resulting from these
reactions. The temperatureis maintained at 900C and oxygenis
Mineralogy and PetrologyGroup, British Geological Survey 0 NERC 1392

Flake graphite

introduced to the sample. The graphitic carbon

is oxidised leaving a
residue of ash. Graphitic carbon content canbe calculated from the
resulting weight loss curve. Provided the sampleis well-characterised
optically orby XRD, it may be possible to quantify other volatileevolving minerals from the weight loss curve. Figure
3 illustrates a
typical weight loss curve.
LO1 presents avery simple and rapid technique for assay
of graphite. It
is a technique whichis useful in the preliminary assessmentof large
number of raw samples from the field, as well
as in the monitoringof
graphite products fromany subsequent beneficiation trials. Relatively
large sample weights used in this test ensure that sampling errors
are
kept to a minimum. Though it potentially provides more detailed
temperature-dependent data,TG requires much more sophisticated
equipment and tends to
be more prone to sampling error because
of the
small sample weight requiredby the thermobalance. Itis very difficult
to obtain a small sub-sample of flake graphite-bearing material which is
truly representative.

Mineralogy and Petrology Group, Btitish Geological Survey 0 NERC 1992

Flake graphite

U
C
0
J

J
4

L
J

25.0

27.0

29.0

31.0

33.0

Peak position ("28)

Figure 2. XRD traces of flake graphite fromZambiaand

Chinashowingcontrasting
peakwidth: height ratio.

Mineralogy and Petrology Group, British Geological Survey 0 NERC 7992

.
.
..

'amorphous' graphite from

35.0

10

Flake graphite

Volatiles
(3%)

Carbon
(71%)

Ash
(26%) ::
~~

~~~

1000

Temperature ("C)
Figure 3. TG weight loss curve of a Zambian flakegraphite.

Mineralogy and Petrology Group, British Geological Survey 0 NERC 1992

Flake graphite

5. LABORATORY PROCESSING TRIALS

Laboratory processing trialsof flake graphite-bearing samples should
aim to establish the optimum method
by which flake of a particular size
range mightbe separated with the maximum possible grade and
recovery. Recovery is the proportionof the total graphite present in the
head sample or starting material which reports
to the concentrate and is
calculated using grade and yield data gathered during processing trials.
Graphite grade is generally determinedby the graphitic carbon content
of a concentrate. Yield is the proportionof the total weightof the
starting material containedin the concentrate. An example of a grade
and recovery calculation for a flake graphite
is given in Appendix6. An
assessment of the economic potentialof samples from a deposit will
require an indication of the likely processing scheme, including grade
and recovery data over a range
of flake sizes. The resultsof various userelated tests may alsobe required.
Graphite is relatively easyto separate fromits host rock, though it is
more difficultto produce high grades(90% and above), and many
producers process their graphite several times to eliminate impurities.
Graphite with very high carbon contents(96 - 99%) must be chemically
and thermally treated to reduce the level
of impurities. A flow chart
illustrating a suggested laboratory processing scheme
is given in Figure
4. A typical commercial processing schemeis given in Figure5.

..

Determination of flake sizewill indicate the size at which the graphite

may become liberated from the host rock.This is important as it will
minimise themount of crushing andgrinding required prior to mineral
separation. Flakes in the size range250 pm to 1 mm will command the
highest prices, with fine graphite flakes (down 125
to pm)also in some
demand. An excess of graphitic fines will reduce the
flake size and
therefore the valueof the final product.Also fine graphite will coat
other minerals, which may then act as graphite during froth flotation and
be recovered with the graphite concentrate
- thereby reducing the grade
of the product. Mica will often occur interlayered with graphite which
may be difficult to remove during preparation. Fine material (such
as
clay and lateriticsoil) may coat the graphite. These minerals will
complicate graphite processing.
5.1 Preparation of graphite-bearingrock
beneficiation

for

The concentrationof graphite from graphite-bearing material

is
dependant on the graphite flake morphology and liberation size.
As
graphite is resistant to weathering it may often
be pre-concentrated in
graphite-bearing soil, but with graphite-bearing rock the graphite needs
to be liberated. Roller mills and cone grinders (or high-speed impact
Mineralogy and PetrologyGroup, British Geological Survey 0 NERC 1992

12

Flake graphite

mills in larger-scale operations) produce a differential

grinding effect
in the coarser size fractions.
that tends to concentrate the graphite
Granular materials, suchas quartz and feldspar, are ground into the
fines whereas graphite flakes slip through relatively unscathed.
However this maybe countered by graphite being ground by contact
with the granular material during grinding to produce
an even
distribution of graphite. Care should be taken not to over grind the
graphite as it will tend
to smear out over the other minerals present.

5.2 Graphiteconcentration
Examination of the size fractions by binocular microscope indicates
if
the correct liberation size was chosen, as essentially all the graphite
flakes willbe free of the host rock. If a large proportionof graphite is
'locked into the host rock then attempts at separation will produce poor
results andthe larger size fractions may need to
be ground further. If
this is required, the next largest sieve diameter is used asnew
the
liberation size. For example,if the original liberation size chosen was2
mm, then 1 mm is selected as the new liberation size (andso on down
the sieve series 2 mm, 1 mm, 500 pm, 250 pm and 125 pm). The
process of grinding and microscopic examination is repeated until the
correct liberation size is achieved.A procedure for grinding and
4.
screening priorto graphite concentration is given in Appendix
5.2.1 Froth flotation

Graphite is generally concentrated from crushed by

rock
froth flotation.
Historically, itwas the first ore tobe concentrated using flotation.
Graphite is a naturally hydrophobic mineral and will float
with little
encouragement. Despite its natural floatability, the separation
of
feldspar, quartz,mica and carbonate gangue is normally improved by
the additionof reagents such as kerosene (which coats the graphite) and
pine oil (which stabilises the froth). Selectivity may
be improved by
adjusting the pH with caustic soda to the
7.5 - 8.5 range (Crozier,
1990).
Sized material is essential for efficient froth flotation. Material less than
125 pm will not concentrate effectively. Laboratory flotation
is
normally carried out as a batch process in glass cells using a mixer
capable of injecting a streamof bubbles into the sample. Appendix 4
outlines the methodof graphite flotation used in BGS laboratories. This
procedure produces a concentrate (material that floats) and a tailing
(material that sinks). A second flotationis carried outto 'clean' the first
graphite concentrate, as well
as to removeany mica and material
smeared with graphite during preparation. The two-part flotation
process resultsin three products: a concentrate (from second flotation),
Mineralogy and Petrology Group, British Geological Survey 0 NERC 1992

13

Flake graphite

middling (material that sank during the second flotation) and tailing
(material that sank during the first flotation). Only small amountsof
reagents are required;3-4 drops of each are usually sufficient after
which no more froth will appear, and it is important not to add too much
initially. Froth flotation generally works beston a 'starvation diet' of
reagents. Graphite concentrates gradingup to 95% can be produced by
froth flotation in the laboratory; grades higher than this can only be
achieved by chemically leaching out the impurities from the concentrate.
5.2.2 Air classification

Industrial-scale froth flotation concentration of graphite requires large

amounts of water and may not be anappropriate processfor mines
situated in arid areas. Air classification can separate graphite flakes
from their host materialby exploiting their low density and platy nature.
Using either 'zigzag' air classifiers or air tables, closely-sized material
can be upgraded by effectively 'winnowing out' the graphite. A
schematic diagram of a zigzag classifieris shown in Figure 6. Air
classification requires different operating conditions for different size
fractions, as larger flakes require
a higher air flow to separate them and
vice versa. An outline procedurefor air classificationof flake-graphitebearing material is given in Appendix5. Pre-concentrates gradingup to
70% graphite and recoveries of70% graphite can be achieved, though
the presenceof mica or other mineralswith a platy habit may prejudice
the effective pre-concentrationof graphite using this method. Where
other platy minerals are absent, air classification may halve the total
weight of material required to be processedby froth flotation,with the
attendant savings in water and reagents.

5.3 Assessment of products

The effectivenessof the laboratory processing trialsis measured by
calculation of the grade and recoveryof the graphite products. Grades
are generally determined using theLO1 technique described above and
recoveries calculated using the method set
out in Appendix 6. Recovery
values will serve to indicate the efficiency
of the beneficiation scheme
adopted and the 'amenability'of the graphite-bearing rock to
processing. The relatively high economic value of graphite products
may mean that low rates of recovery can sometimes be tolerated. Data
on graphitic carbon content(or grade) of laboratory concentrates froma
specific size range may
be compared directly to commercial
specifications.
Chemical analysis of the ash produced
by LO1 tests on concentrates will
indicate if undesirable oxides are present. These include major elements
such as SiO2,Al2O3, MgO, CaO and Fe2O3. Trace metals such as
Mineralogy and Petrology Group, British Geological Survey 0 NERC 1992

14

Flake graphite

copper, cobalt,tin and arsenic are undesirable in graphite used for

alkaline batteries and carbon brushes. Appendix
7 gives the chemical
analysis of the Lac Knife graphite from Canada.
5.3.1 Use-related testing

A number of tests related to specific applications

of flake graphite canbe
carried out on concentrates from laboratory processing
trials. Userelated tests for graphite used in the manufacture
of crucibles include
apparent density, flakiness and toughness. Apparent density is
largely controlledby the thickness of the flakes, assuming a close size
distribution. A value of45 g per 100 cm3 is considered good and
anything over58 is poor. This simple bulk density valueis determined
by the weightof graphite required tofill a 100 cm3 container.
Flakiness is determined
by placing 100 g of graphite (greaterthan 250
pm) into a 250 ml measuring cylinder. The cylinder is then tapped until
the material is completely packed down.The volume which the material
occupies is known asthe flakiness index, which for crucible grade
flake
should notbe less than 129 ml. Toughness is determined by grinding
a sampleof graphite of a fixed size distribution under standard
conditions and then determining the amount of less than
75 pm material
produced.

Mineralogy and Petrology Group, British Geological Survey0 NERC 1992

15

Flake graphite

I (Assay,SUB-SAMPLE
I
mineralogy and liberation

HEAD MATERIAL

size of graphite)

JAW CRUSHING OR
ROLLER MILLING

(To liberation
size)

I
-

SCREENING

>

(Using sieveseries 2 mm, 1 mm, 500 pm, 250 pm and 125 pm)

SUB-SAMPLE
(Assay)

u/

(Bypass air classification if initial

characterisation indicates that it would
not produce adequate preconcentration)

AIR CLASSIFICATION
(To preconcentrate graphite flakes)
I

FROTH FLOTATION
(To produce graphite concentrate of high grade and recovery)

\1/

CHARACTERISATION OF GRAPHITE CONCENTRATE

(Assay, mineralogy and chemistry)

Figure 4. Suggestedflowsheet

for laboratory processing of flake graphite.

Mineralogy and Petrology Group, British Geological Survey0 NERC 1992

16

Flake graphite

Run of mill ore

1-1

crusher
Cone

ore bin

Jaw crusher
Rougher flotation cells

Tailings to dam
I

""mill

Cleanerflotationcells

r--l

Centrifuges

Screens

r
Bagging

Figure 5. Processflowsheet

at Lynx graphite mine, Zimbabwe (Muchemwa, 1987).

Mineralogy and Petrology Group, British Geological Survey0 NERC 1992

17

Flake graphite

Air flow

Air flow

Cyclone

Zig-zag air
classifier
0 column

'

Overflow
(Light and flaky
particles)

Direction
of air flow

_ .
.. .

Underflow

Figure 6. Schematic diagram of a zig-zag airclassifier.

Mineralogy and Petrology Group, British Geological Survey 0 NERC 7992

18

6. CASE HISTORIES
Laboratory evaluationof contrasting flake graphite-bearing rocks from
Mozambique and Zambia are described briefly below.
6.1 Mozambique

Three samples of graphite-bearing schistfrom Ancuabe, Mozambique

were evaluated. These consistedof fresh core (1 1%graphite);
weathered rock(13% graphite) and eluvialsoil (9% graphite). They all
contained flake graphite(0.5-2 mm in size), with varying amountsof
quartz and feldspar; the weathered sample was iron stained and the
of iron-oxide-rich soil. The
eluvial samples contained a large proportion
samples were processed down
to 1 mm, which was taken as the
liberation size. Graphite-rich >1 mm fractions were produced during
stage cone grinding, as a direct result
of preferential concentration,The
>125 pm size fractions were pre-concentrated using 'zigzag'
air
classification and the air classified concentrates further upgraded using
froth flotation. The results are given
in Table 4.
Air classification produced concentrates with
up to 73% graphite andup
to 7 1% recoveries. Froth flotation produced concentrates with
up to
90% graphite and recoveriesof up to 72%. Approximately 85% of the
head material was rejected
by using air classification priorto froth
flotation. This highly effective pre-concentrationof graphite would
final
result in considerable savingsof water and reagents during the
froth flotation process.
Table 4. Concentrate gradeandrecovery

of graphitefromMozambique.
~

Air

~~~~

Sample
Fraction

Core
Weathaed
Eluvial

Fraction

Assay

Recovery

Wt %

Wt %

Assay

Recovery
Wt %

Wt %

Wt %

Wt %

14.4
13.7
6.5

53.7
73.2
55.3

63.7
71.1
36.7

9.5
11.5
3.2

88.4
90.2
90.5

65.1
72.2
29.3

Note: Due to lack of material the air classification concentrate and middlingof the -250+125 pm fractions were
combined prior to froth flotation, Consequently some of thefroth flotation recovery figures are higher than
those for air classification.

Mineralogy and Petrology Group, British GeologicalSurvey 0 NERC 1992

19

Flake graphite

6.2

Zambia

Three samplesof graphite-bearing schist from Njoka and Petauke,

Eastern Province, Zambia, were examined. The Njoka graphite samples
comprised a rubbly graphite-schist
(A) and a consolidated graphiteschist (B). Both consisted mainly of feldspar with a small amountof
mica, graphite (16%in A and 18% in B)and quartz. The petrography
of Njoka (B) is briefly described below and illustrated
by the
photomicrographs in Figures7 and 8. The graphite in both Njoka
samples is the fully ordered variety. The samples were crushed e1
to
mm (which was taken as the liberation size) by a combination
of jaw
crushing, disc millingand cone grinding, and then screened down to
125 pm. Graphite-rich >1 mm fractions were produced during stage
cone grinding. The >250pm fractions were pre-concentratedby air
classification and the products
of air classification, together with
the 250125 pm fraction, were concentratedby froth flotation. The results are
given in Table5. Air classification produced relatively low-grade flakegraphite concentrates with very low recoveries, due
to mica
concentrating with the graphite. Froth flotation produced higher-grade
concentrates with good recoveries. The main impurity in the
concentrates was found to be mica.
The Petauke graphite schist contains mainly feldspar and mica, with
smaller amountsof graphite (6%)and quartz. Material was processedin
a similar fashion to that from the Njoka deposit, although the 250-125
pm fraction was pre-concentratedby air classification andthe products
upgraded by froth flotation. Results are also given in Table 5. Air
classification produced a very low-grade flake graphite concentrate with
of mica. Froth flotation produced a
a low recovery due to the presence
high-grade flake graphite concentrate
with a very low recovery due to
the mica, and further processing would be required to retrieve the
tadings.
graphite lost in the flotation
Table 5. Concentrate gradeandrecovery

flotation
Sample

FrothAir classification

Wt %

Recovery

Fraction

Wt %

Wt %

Wt %

Wt %

Wt %

12.9
11.0
11.2

53. I
70.7
24.9

40.8
4 1.7
40.2

16.0
23.3
2.3

78.2
73.2
87.1

70.1
73.5
39.0

Fraction

Njoka A
Njoka B
Petauke

of graphitefromZambia.

Assay

Mineralogy and Petrology Group, British GeologicalSurvey 0 NERC 1992

Assay

Recovery

20

Flake graphite

6.2.1 Petrographic description of Njoka (B) graphite

Njoka (B) graphite is a quartz-albite-schist containing graphite and mica

form a crudely foliated
(possibly phlogopite). The quartz and feldspar
mosaic with curved, slightly interlocking interfaces. The foliation is
typically createdby elongation of the feldspar grains. Thefeldspar has
patchy turbid areas of (sericitic) alteration. Parallel to the quartzfeldspar lithons (lenses) is a slightly undulose schistose fabric formed
of
graphite flakes, intergrownwith mica. The mica shows some alteration.
Graphite flakes areup to 3-4 mm long, generally0.4-2 mm, and from
0.05 mm thick.
Figures 7 and 8 show photomicrographsof Njoka (B) graphite in
reflected and transmitted light respectively. Figure7 shows the graphite
as light grey to white elongated grains with the quartz, feldspar and mica
represented by the darker grey material surrounding the graphite. Most
of the graphite flakes are oriented normal
to the plane of the
photomicrograph, with some inclined (thelight grey areas). Figure 8
shows the same area as in Figure
7 (at a slightly lower magnification)
with black graphite surrounded by quartz and feldspar. Pale yellowishbrown mica is aligned parallel to the graphite.
Based on study of the thin section, a liberation size of1 mm was
selected as the most appropriate for processing
of the Njoka graphite.
The close associationof mica with graphiteis an indicator of potential
problems during processing. Mica tendsto concentrate with graphite
during air classification and may affect recovery of graphite during froth
flotation.

Mineralogy and PetrologyGroup, British Geological Survey 0 NERC 1992

22

Flake graphite

REFERENCES
Bonneau, J and Raby, R (1990) The Lac
Knife graphite deposit.Mining
Magazine, July. 12- 8.
1
Crozier, RD (1990) Flotation - Theory, Reagents and Ore Testing.
Course handbook, Camborne Schoolof Mines, UK.

Fogg, CT and Boyle,EH (1987) Flake and high-crystalline graphite

availability - Market economy countries.US Bureau of Mines
Information Circular 9122.
Graffin, GD (1983) Graphite. In:
Industrial Minerals andRocks,
Lefond, SJ (Editor), American Instituteof Mining, Metallurgicaland
Petroleum Engineers, 2.
Harben, PW (1 992)Industrial MineralsHandybook. Metal Bulletin
PLC, London.
Gauss, UH, Schimdt, HW, Taylor, AT and Sutphin, DM (1989)
International Strategic Minerals Inventory
Summary Report - Natural
Graphite. US Geological Survey Circular 93043.
Landis, CA (197 1) Graphitization of dispersed carbonaceous material
in
metamorphic rocks.Contr. Mineral. and Petrol. 30, 34-45.
Muchemwa, E (1987) Graphitein Zimbabwe. Mineral Resources Series
20, Zimbabwe Geological Survey.

Suggested further reading

Russell, AR (1988) Graphite- Current shortfallsin flake supply.
Industrial Minerals, 255,23-43.
Sutphin, DM and Bliss,JD (1990) Disseminated flake graphite and
amorphous graphite deposit types:
an analysis using grade and tonnage
models. CIM Bulletin, August, 85-89.
Taylor, HA, Jr. (1985) Graphite. In:Mineral Facts and Problems, US
Bureau of Mines, Bull. 675.
Tron, AR (1964) The Production and Uses ofNatural Graphite.
Department of Scientific and Industrial Research, Warren Spring
Laboratory. HMSO, London.

Mineralogy and Petrology Group, British Geological Survey 0 NERC 1992

Appendix 1: Specifications of graphite for differentapplications.

Type
Application

graphite
graphite
additive

Comments
size

Rehctories
Magnesia
Alumina
Crucibles
Expanded graphite
Foundry
Foundry core
and
mould
washes

Brake
linings
/ clutch
Bearings
Lubricants
batteriesDry cell
batteries
Alkaline
steel
Recarburizing
brushes
Carbon
Electrical
Pencils
coatings Conductive
aints
Packing
Polishes
(lubricating)
mud
Drilling
Explosives
(control of burning rate)
Nuclear reactors (moderators & reflectors)
prevention)
(scale
Boilers

F
F

AF
F
AFV

AF
AFV

FV
AFV
A
AF
A
AFV

FV

AF
A

FV
AFV

Flake
C content

85 - 90%
Min. 85%
80 - 90%
Min. 90%
40 - 70%
70 - 90%

150 - 710 pm
150 - 500 pm
+150 pm
200 - 1700 ym
53 - 75 yrn
-75 ym

Min. 98%
90 - 93%
98 - 99%
Min. 88%
Min. 98%
98 - 99%
95 - 99%
93 - 95%
95 - 97%
50 - 55%
85 - 90%
85 - 90%

<75 pm

AF

80%+
65%

F
F

93 - 95%
50%+

Ash <2%,often up to 10%.

Bulk density 48 - 54 g per 100 cm3

Sulphides and other readily fusible minerals undesirable

+150 pm

53 - 106pm
85% <75 pm
5-75pm
-5 pm
4 3 ym
+150 ym
+150 pm
<75 ym
4 5 0 pm
4 5 0 pm
N/A
4 5 0 ym
N/A
N/A

Free from sulphides, abrasive material and metallic contaminants.

No metallic impurities andS less than 0.5%
No impurities such as Cu, Co, Sb and As.

< 1%ash / silica. No abrasive or metallic contaminants.

Free of gritty impurities
May contain up to 25% silica, butfree from gritty impurities.

About 41b per b a d of mud

Free from sulphides and acids.Low moisture content
Free from high neutron absorbing elements, e.g. boron

Key :A - amorphous; F - flake; V - vein; N/A - Not available

Table adapted from Russell (1988), Harben (1992), Goossens (1982), Muchemwa (1987) and Tron (1964).
t3
W

24

Flake graphite

Appendix 2: Particle size and carbon content data for

commercialgraphite

Source
China
Madagascar

Norway
Canada (Lac Knife)
India
Zimbabwe (Lynx mine)
S r i Lanka (vein graphite)

Mexico
Italy
Austria

Large flake
Medium flake
Small flake
Large flake
Medium flake
Fine flake
Extra fine flake
Large flake
Medium flake
Small flake
Large flake
Medium flake
Small flake
Large flake
Small flake
Flake A
Flake powder
Large lump/ lump
Chippy dust
Powders
Amorphous
Amorphous
Amorphous

Particle size

% carbon

80% > 300 pm

80% > 180 pm
50% > 180 pm
97% >250 pm
97% A 8 0 pm
95% > 180 pm (min)
N/A
A 5 0 pm
>lo0 pm
<75 pm
>150 pm
>75 pm
<75 pm
180 - 600 pm
4 5 0 pm
A 6 0 pm
4 6 0 pm
>10 mm
4 mm
<75 pm

85 - 90%
85 - 90%
80 - 90%
85 - 94%
80 - 92.5%
75 - 90%
70 - 90%
85 - 95%
85 - 95%
80 - 95%
92 - 95%
90 - 94%
85 - 89%
70 - 88%
70 - 88%
86 - 95%
80 - 90%
92 - 99%
80 - 99%
70 - 99%
80 - 90%
60 - 80%
66 - 99%

Table adapted fromHarben (1992), Muchemwa (1987),Graffm (1983) and Bonneau (19%).

Mineralogy and Petrology Group, British Geological Survey 0 NERC 1992

25

Flake graphite

Appendix 3: Graphitic carbon assay

Assuming a representative sample (about
5- 10 kg) of the graphite has
been taken, the laboratory evaluation should follow two roughly similar
paths. This depends on the degree of induration of the sample.

Preparation of graphite-bearing rock

1.

Retainareferencesampleandjaw-crushtheremainingmaterial
to less than 4 mm.

2. Extract a 50g sub-sample

than 500 lm.

3.

by cone and quartering. Grind to less

Determineparticle-sizedistribution of crushedmaterial by
screening remainder on the sieve series mm,
2 1 mm, 500 pm,
250 pm and 125pm.

Preparation of graphite-bearing soils

1.

Retainareferencesampleandsub-sample
quartering.

1 kg by cone and

2.Determineparticle-sizedistribution
of sub-sample by dry
screening from4 mm down to 125pm. Extract small
sub-samples from each size fraction
and grind to less than 500
pm.

3.

Screen remainder of sample at 4 mm and reduce >4 mm by jaw

crushing. Extract a50 g sub-sample by cone and quarteringand
grind to less than500 p.m.

Mineralogy and Petrology Group, British Geological Survey 0 NERC 1992

26

Flake graphite

Appendix 3 (continued)
Assay of graphite-bearing rock

2.

Dry overnightat105Candweighaccurately(to
places).

4 decimal

3. Heat in furnace to 1000C for 2 hours to completely oxidise the

graphite. Coolin a desiccator and weigh.

4.

Calculategraphitecontentfromfollowingequation;
Percentage graphitic carbon
= 105C weight -

1000C weight X 100

105C weight

Assay of graphite-bearing soil

1.

Followpoints1-2 in theprocedure for assay of graphite-bearing

rock.

2. Heat to 375C for 16 hours to oxidise and remove all the organic
matter and dehydrate iron hydroxides. Cool in a desiccator and
weigh.

3.

Heatto600Cfor1hourtodehydroxylateclaysandany
remaining iron hydroxides.Cool in desiccator and weigh.

4.

Heat in furnace to 1000C for 2 hours to completely oxidise the

graphite. Cool in desiccator and weigh.

5. Calculate graphite contentfrom following equation;

Percentage graphitic carbon
= 600C weight - 1000C weight X 100

105C weight

Mineralogy and Petrology Group, British Geological Survey 0 NERC 1992

27

Flake graphite

Appendix 3 (continued)

Minerals likely to interfere withgraphiteassay

Mineral

Calcite
Dolomite
Kaolinite
Limonite
Goethite
Gibbsite
Brucite
Talc*
Tremolite*

Composition

by loss-on-ignition

Decomposition

Weight
loss

Decarbonation at 920- 975C

Decarbonation at 720-950C
Dehydroxylation at 450- 700C
Dehydroxylation at 350- 400C
Dehydroxylation at 250- 360C
Dehydroxylation at 240- 560C
Dehydroxylation at 400- 450C
Dehydroxylation at850 - 1000C
Dehydroxylation at 990- 1070C

43.97%
47.7%
-13%
>10.1%
10.1%
34.6%
30.89%
4.75%
2.8%

Temperature ranges for carbonates apply only under dynamic heating conditions
(e.g. TG analysis). Under static
heating, carbonates will decompose at least100C lower.

* It is not possible to distinguish the weight losses from these minerals from that of graphite.

Mineralogy and Petrology Group, British Geological Survey 0 NERC 1992

28

Flake graphite

Appendix 4: Sample preparation and frothflotation

procedure
Sample
preparation:
1.

Weighthesamplepriortostagegrindingusingeither
a roller
mill or cone grinder until
all material passes the liberation size.

2.

Dry screenthegroundsamplesthroughthesieveseries,
beginning with the sieve immediately below the liberation size.
Weigh the material retained on each
sieve and calculate the size
distribution.

3.

Extractarepresentative 50 gsub-samplefromeachsizefraction
for determination of carbon content. Grind to less than 500 pm.

Frothflotation:
1.

Add 200 - 500 g of sized material to the cell and fill with
deionised water.

2.

Add a dropof kerosene to the suspension andmix. The

kerosene collector coats the graphite.

3.

Add a dropof methyl isobutyl carbinol (MIBC) frother and

continue mixing.

4.

Inject air into the sample. This generates bubbles from the base
of the cell. The graphite adheres to the MIBC-stabilised bubbles
a froth.
and floats to the surface forming

5.

Scoop out the graphite froth.

6.

Repeat steps 2-5 until no more graphite remains.

7.

Following steps 2-6, refloat the initial graphite concentrate.

No
further additionof reagents is required.

8.

Oven dry concentrates, middlings and tailings overnight at 60C.

Weigh and calculate% yield.

9.

for
Extract a representative50 g sub-sample from each product
determination of carbon content. Grind to less than500 pm.

10.

Determine grade using procedure set

out in Appendix 3.

11.

Calculate recovery using procedure set out

in Appendix 6.

Mineralogy and Petrology Group,British Geological Survey 0 NERC 1992

29

Flake graphite

Appendix 5: Air classification procedure

Air classificationis a method of concentrating graphite flakes by using a
stream of air to separate granular minerals such as quartz and feldspar.
The equipment used at BGS is
an Alpine zigzagair classifier, which
consists of a zigzag columnup through whichair flows. Material treated
by air classification mustbe closely sized and fall within the range 125
pm and 6 mm (see Appendix 4). Graphite flakes are entrained in the
upward air flow, reporting to the overflow, and granular material
falls
through the airflow, reportingto the underflow. Theaim of air
classification shouldbe to produce a graphite preconcentrate with a high
grade and recovery and low yield.
1.

The conditions for the separation of flake graphite from granular

material depend on the grain size, with higher
air flow rates
required for coarser size fractionsand vice versa. Air
classification of graphite-bearing material is a process
of trial and
error.

2.

Passthegraphite-bearingsizefractionthroughtheair-classifier
at afiied air flow rate, aimingto remove graphite flakes only
into the overflow (graphite concentrate). If granular material
reports to the ovefflow then the fraction should
be recombined
and re-passed at a lower
air flow rate.If there appears tobe an
excess of graphite flakes reporting to the underflow then re-pass
at a higherair flow rate.

3.

Onceagraphiteconcentrate has beenproduced,theunderflow

material shouldbe re-passed ata higher air flow rate in order to
remove all the remaining graphite.This time granular material
will report to the overflow (middlingfraction) andan attempt
should be made to producean underflow as free from graphite
as possible (tailing fraction).

Mineralogy and Petrology Group, British Geological Survey 0 NERC 1992

30

Flake graphite

Appendix 6: Graphitegrade,
calculation.

.. ..

yield andrecovery

The following is aworked example of grade, yieldand recovery

calculations for a series
of products from laboratory beneficiation
trials
on a flake graphite.

. .

Weathered graphite schist:

Head sample weight = 20 kg

Head sample grade = 13.2% graphite equivalent to2.64 kg graphite
Sample is screened:
lmm-500pm fraction yield = 7.68 kg or 38.4 wt %
lmm-500pm fraction grade= 17.3% graphite equivalent to 1.32 kg graphite
% graphite recovered from the head sample to the lmm-500pm fraction
= 1.32D.64 X 100
= 50% recovery

lmm-500pm fraction is air classified:

Product
yield
19.7
62.7
100

1.51
4.81
7.68

17.6
conc
mid
tail
total

wt (kg)

wt % size
offraction
graphite
grade

(%)

82.5
11.5
2.4

1.35

In order to calculate the

% graphite recovered from the head
to each product, first
calculate the weightof graphite in each product from the yield and grade values:
wt graphite in conc = 1.35/100 X 82.5 = 1.11 kg
wt graphite in mid = 1.51/100 X 11.5 = 0.17 kg
wt graphite in tail = 4.81/100 X 2.4 = 0.11 kg
therefore the % graphite recovered from the head to each
of the products isas follows:

lmm-500pm fractionconcentrate
lmm-500pmfractionmid
lmm-500pmfraction tail

Mineralogy and Petrology Group, British Geological Survey0 NERC 1992

= 1.1 1/2.64 X 100 = 42.2% recoverv

= 0.17/2.64 X 100 = 6.4% recovery
= 0.11/2.64 X 100 = 4.2% recovery

31

Flake graphite

Appendix 7: Chemistry of Lac Knifegraphite,

Canada

500- 150 ym
wt %

<150 pm
wt %

4.32

95.5
0.62
0.38
0.79
3.57

90.2
0.23
0.78
1.37
6.73

47.0
14.1
22.8
5.05
4.39
0.55
3.26
0.53
0.15
0.24

44.5
15.0
24.2
5.32
3.92
0.51
3.58
0.54
0.17
0.15

46.0
15.3
23.O
4.15
4.94
0.84
3.20
0.43
0.15
0.18

Component
>500

pm
wt %

94.1
Graphitic
carbon

co2
S
Volatiles
Ash

75

Ash composition
(i) Major elements
S io2
A1203
Fe203
MgO
CaO

Na20
K20
Ti02
Mn02
P205

. .

(ii) Trace elements(PPM)

2,800 V
3,800Mo
688
cu
195
co
220
U
Ni

0.72
0.35

3,100
1,900
603
133
230

Data from Bonneau & Raby (1990).

Mineralogy and Petrology Group, British Geological Survey 0 NERC 1992

2,200
5,400
883
33
120
449

Quebec,