APPENDIX 2

THE THERMAL OXIDATION OF GRAPHITE

A. Blanchard

A2.1. General

Graphite is used in a variety of nuclear reactor types; principally for moderator, reflector, fuel
sleeve and fuel tube material. Since it is a form of carbon, like coal and charcoal, its oxidation
behaviour might be expected to be of concern to the graphite chemist. The oxidising
environments of particular interest are air (oxygen), carbon dioxide and steam (water).

A2.2. Reactions and Thermodynamics

A2.2.1. Reaction with Oxygen

½ O2 + C = CO ∆H = - 110.5 kJ.mol-1 (A2.1)

O2 + C = CO2 ∆H = - 393.5 kJ.mol-1

(A2.2)

where ∆H is the standard enthalpy of formation at 298 °C.

Reaction (A2.1) maximises the amount of carbon which may be removed by a given
mass of oxygen (as carbon monoxide). Reaction (A2.2) maximises the amount of heat
produced by oxidising a given mass of carbon (to carbon dioxide). Reaction (A2.2) may also
be regarded as proceeding in stages, with reaction (A2.1) followed by

½ O2 + CO = CO2 ∆H = - 283.0 kJ.mol-1

(A2.3)
(obtained by difference - Hess’s law)

Reaction (A2.3) can take place wholly in the gas phase.

The above reactions are exothermic and favoured thermodynamically. Despite this fact,
pure dense nuclear graphites do not readily react with air, so kinetic factors are obviously of
importance.

A2.2.2. Reaction with Carbon Dioxide

This can be readily inferred by multiplying reaction (A2.1) by 2, including the ∆H term.
The result is then added to reaction (A2.2) reversed. Overall, the (Boudouard) reaction is

C + CO2 = 2CO ∆H = +172.5 kJ.mol-1 (A2.4)

A2.2.3. Reaction with Water (water gas reaction)

C + H2O = CO + H2 ∆H = +131.3 kJ.mol-1 (A2.5)

also
207
 C + 2H2O = CO2 + 2H2 ∆H = -82.4 kJ.mol-1 (A2.6)

The hydrogen produced can then react with carbon:

C + 2H2 = CH4 ∆H = -74.81 kJ.mol-1 (A2.7)

Subtracting (A2.5) from (A2.6) gives

CO + H2O = CO2 + H2 ∆H = - 213.7 kJ.mol-1

(A2.8)

This is the water gas shift reaction which takes place in the gas phase.

In many practical cases, products from the above reactions are free to escape, such that
thermodynamic equilibrium is not reached. The enthalpy changes are of importance, however,
since they give a measure of the heat produced in exothermic reactions.

A2.3. Mechanisms, Regimes and Kinetics

A2.3.1. Mechanisms

The oxidation mechanisms consist of a series of physical and chemical steps. Non-catalysed
oxidation typically follows the route:

(i) Transport of oxidant to the graphite surface.

(ii) Adsorption of oxidant onto the graphite surface (physisorption).

(iii) Formation of carbon-oxygen bonds (chemisorption).

(iv) Formation of carbon-hydrogen bonds in reaction (A2.7) (reduction).

(v) Breaking of the carbon-carbon bonds.

(vi) Desorption of carbon monoxide, or other product.

(vii) Transport of reaction product from the graphite surface.

Any of the above steps may be rate controlling, i.e. develop the major reactant
concentration gradient.

A2.3.2. Controlling Factors

Factors controlling the rate of oxidation may include the following:

(viii) The rate at which the oxidant is supplied to the surface.

(ix) The partial pressure of the oxidant.

(x) The reactive surface area available to the oxidant at the surface.

208
(xi) The amount and distribution of catalytic impurities in the graphite.

(xii) The temperature.

(xiii) The rate at which reaction products are removed.

(xiv) The fast neutron damage to the graphite.

(xv) The amount of pre-oxidation (radiolytic or thermal burn-off).

(xvi) The quantity of in-pore deposits.

(xvii)The effective diffusion coefficient.

A2.3.3. Oxidation Regimes and Kinetics

Regime 1

At low oxidation rates (generally at relatively low temperatures for the particular oxidation
reaction) the oxidant may be at essentially the same concentration throughout the transport
pores of the graphite. This ‘chemical’ regime is characterised by the fact that the reaction rate
is largely determined by the intrinsic reactivity of the graphite (steps (ii) to (vi), above).
Different parts of the structure may react at different rates; the binder being more reactive
than the grist particles and edge atoms being more reactive than basal plane atoms, for
example. It may also be the case that the gas composition varies in non-transport pores and
that the effect of a given gas composition varies with pore shape (because of gas phase
reactions).

Reaction between air and pure nuclear graphites is generally not measurable below
about 350 °C and only becomes significant in the region of 400 °C. The rate of reaction is
typically of the order of 3-8 x 10-10 kg.kg-1.s-1 at this temperature (historically expressed in
µg/gh; 1µg/gh being 2.78 x 10-10 kg.kg-1.s-1). The chemical regime then extends typically up
to 550-600 °C.

The units for oxidation rate imply a rate law of the form

dm
= km (A2.9)
dt

where
m = graphite mass (kg)
t = time (s)
k = a (rate) constant (s-1)

However, for a solid reacting body, rate laws of the form shown below would be
expected for reaction at the superficial surface (with some simple assumptions):

dm
Slabs =k (A2.10)
dt

209
 dm 1
Cylinders = km 2 (A2.11)
dt

dm 2
Spheres = km 3 (A2.12)
dt

The reaction is, of course, taking place within the porous body of the graphite. It cannot,
however, be related to the initial open pore volume (which might be expected to be
proportional to graphite mass) since the reaction must take place at the surface of the pores.
The rate law is thus only superficially similar to a homogeneous first order gas phase
expression and care must be taken in its use, particularly in describing the time dependence of
graphite burn off.

The variation of reaction rate with temperature is of importance:

E
−
k = Ae RT
(A2.13)

where
A is a pre-exponential factor (s-1)
E is the apparent temperature coefficient of reaction (‘activation energy’) (J.mol-1)
R is the gas constant (J.mol-1.K-1)

E has a typical value of 170 kJ.mol-1.

Similar considerations apply to reaction with carbon dioxide and water vapour (H2O).
The reaction with CO2 is of less importance, however, since it is negligible at 625 °C
(Thurlbeck, 1962) and does not pose a problem even at the highest AGR inner sleeve
temperatures of 675 °C (Prince, 1976). The reaction is also endothermic (equation (A2.4))
and so does not have the same safety implications as the reaction with air.

The reaction with H2O is of particular importance in HTR reactors, because of the
(generally small) inleakage from the steam side into the gas circuit, where it can react with
hot graphite fuel tubes. Since the partial pressure of the water vapour is a variable in this
system, rate equations of the form

r = A Pn (A2.14)

are applied, where

r = the specific reaction rate (kg.kg-1.s-1)
A = a (rate) constant (s-1.(N.m-2)-1)
P = partial pressure of water vapour (N.m-2)

The reaction with water vapour is generally insignificant below 800 °C and
approximately obeys equation (A2.14) with n=0.5 over the temperature range 1000-1200 °C.
The kinetics can also be described by a Langmuir-Hinshelwood scheme (Walker et al, 1959;
Atkins, 1987; Stairmand, 1990).

Regime 2

In this regime, the reaction rate becomes high enough for access of the gas to the in-pore
210
structure to be significantly limited by diffusion control (steps (i) and (vii), above). This can
also arise from a particularly restrictive structure. For the air-graphite system, this occurs
approximately in the range 600-900 °C. The ‘activation energy’ is halved in this regime and
the kinetic expressions involve the effective diffusion coefficient for the graphite (Walker et
al, 1959; Giberson and Walker, 1965).

Regime 3

This is the mass transfer regime (Burnette et al, 1979; Raeder and Gulden, 1989), where
reaction at the superficial surface of the graphite is so high that most of the oxidant is
consumed there, the oxidant concentration gradient generally developing across the laminar
sub-layer. The reaction rate is now expressed in terms of the superficial surface area of the
graphite (kg.m-2.s-1) together with any oxidant partial pressure dependence. The change from
one regime to another may be progressive and mode 2 may appear to be missing in some
cases.

Two other ‘regimes’ appear trivial, but can usefully be distinguished:

Regime 4

If there is a fixed rate of ingress of oxidant to the system, for example as a known quantity of
impurity in the make-up gas, the rate of oxidation cannot exceed the rate of supply of oxidant
(rate balance). The preferred site of any resulting oxidation may not be known, however.

Regime 5

If the system contains a fixed amount of oxidant, for example that remaining after blowing
down and recharging the coolant gas, the extent of oxidation is limited by the amount of
oxidant available (mass balance). Neither the location, nor the rate of reaction, may be
known in this case.

A2.3.4. Catalysis

The kinetics in the chemical regime may be further complicated by catalysis (McKee, 1981).
The catalyst (impurity) particles act to increase the reaction rate by offering an alternative
reaction pathway. This lowers the activation energy. Reaction rates are particularly increased
at lower temperatures (this leads to a ‘compensation’ effect).

A simple catalytic model involves the oxidation of metal atoms (M) by oxygen,
followed by reduction of the oxide by carbon:

½ O2 + M = MO (A2.15)

MO + C = CO + M (A2.16)

The reaction may proceed by a ‘tunnelling’ mechanism.

Reaction inhibitors are also known (McKee, 1991) and some substances are able to act
to either promote or inhibit reaction, perhaps by competing for active sites (e.g.
water/oxygen). Boron (with phosphorus) is known as an inhibitor for thermal oxidation, but
when intercalated into the graphite structure will promote oxidation (Karra et al, 1995)

211
(perhaps by mimicking fast neutron damage).

A2.4. Requirements for the Information

Early designs of graphite moderated reactor operated with air as the coolant and there was a
requirement to understand both the likely ongoing oxidation behaviour of the graphite and its
behaviour during a temperature excursion. As these reactors were superseded by carbon
dioxide cooled designs, an understanding of the reaction of graphite with air remained
important for the following reasons:
(i) Safety case information relevant to both major and minor ingresses of air to the system
under fault or other conditions (Dodson, 1960; Nairn and Wilkinson, 1960; Blanchard and
Fitzgerald, 1978).
(ii) The possible requirements to carry out deliberate oxidations to
(a) Remove deposits from fuel.
(b) Open up the structure of low diffusivity graphite to improve inhibitor (e.g.
methane) access.
(c) Remove deposits from fuel pin and heat exchanger surfaces to improve heat
transfer.
(d) Estimate the amount of deposit in graphite moderator and fuel sleeves by
differential thermal oxidation (Welch, 1972; Oxley and Dymond, 1972; Baguley and Livesey,
1972) (so as to be able to correct the weight loss).
(e) Alter the structure of experimental graphites in controlled ways to improve
theoretical knowledge of the interaction between reactivity and structure.
A2.5. Essential Measurements and Knowledge
The complexity of graphite oxidation behaviour is such that the following are generally
required:
(i) Good statistical data on the relevant oxidation rates for a range of blocks and heats.
(ii) An awareness of the relevant oxidation regime and the rate laws which are likely to
apply.
(iii) Information on the temperature coefficient of reaction.
(iv) Information or measurement on the heat change on reaction.
(v) Theoretical or experimental information on factors affecting the oxidation rate, such as
fast neutron damage, burn-up, deposition of potential catalysts, etc.
(vi) Modelling knowledge to extrapolate from small scale samples, or full scale tests.
As some graphite reactors are coming to the end of their lives, there is an increasing
requirement to carry out assessments for long term storage or disposal (Wickham et al, 1996).
The potential oxidation behaviour is also of concern in this context.
A2.6. Acknowledgements
In compiling this account, the author has been particularly grateful for reviews by John
Stairmand (Stairmand, 1990) and Tony Wickham et al (Wickham et al, 1996) who in turn
cite the pioneering work of P.L. Walker (Walker et al, 1959) and Dubinin (Dubinin, 1966).
Any important omissions are the author’s own, however.

212
 REFERENCES

[1] Atkins P.W. Physical Chemistry, 3rd edition, Oxford University Press, 783 (1987).
[2] Baguley G. and Livesey J.N. DTO Applied to Graphite/Carbon Deposit Systems - Paper
11, The Attack Rate of the underlying Graphite during Oxidation of In-pore Deposits,
Symposium on Differential Thermal Oxidation held at Berkeley Nuclear Laboratories
(1972).
[3] Blanchard A. and Fitzgerald B. The Air Reactivity of WAGR Moderator and Sleeve
Graphites, UKAEA Report WAGR/TC/P(78)21 (1978).
[4] Burnette R.D., Velasquez C. and Harrington R.O. Studies of the Rate of Steam Oxidation
of Graphite at High Gas Velocity, 14th Biennial Conference on Carbon, Pennsylvania
State University (1979).
[5] Dodson M.H. The Depth of Oxidation of Graphite: A Theoretical Approach, UKAEA
Report DEG 148 (CA) (1960).
[6] Dubinin M. Chemistry and Physics of Carbon, 2 (Ed. P.L. Walker, Jr), Marcel Dekker,
NewYork, 51 (1966).
[7] Giberson R.C. and Walker J.P. Reaction of Nuclear Graphite with Water Vapour - Part
1: Effect of Hydrogen and Water Vapour Partial Pressures, Battelle Northwest Report,
BNSA-181 (1965).
[8] Karra M., Zaldivar R.J, Rellick G.S., Thrower P.A. and Radovic L.R. Substitutional
Boron in Carbon Oxidation - Inhibitor or Catalyst?, Proc. 22nd Biennial Conference on
Carbon, San Diego (1995).
[9] McKee D.W. Chemistry and Physics of Carbon, 16 (Ed. P.L Walker, Jr), Marcel Dekker,
New York, 1 (1981).
[10] McKee D.W. Chemistry and Physics of Carbon, 23 (Ed. P.L Walker, Jr), Marcel Dekker,
New York, 173 (1991).
[11] Nairn J.S. and Wilkinson V.J. The Prediction of Conditions for Self-sustaining Graphite
Combustion in Air, Paper GCM/UK/11, US/UK Compatibility Conference (1960).
[12] Oxley E. and Dymond B. DTO Applied to Graphite/Carbon Deposit Systems - Paper 8,
DTO Studies of Irradiated Graphite using Gas Chromatography, Symposium on
Differential Thermal Oxidation held at Berkeley Nuclear Laboratories (1972).
[13] Prince N. Experience in the Design of Graphite Moderator Structures, Paper presented to
the Nuclear Engineering Society (1976).
[14] Raeder J. and Gulden W. Safety Analyses performed by NET, IAEA Technical
Committee Meeting on Fusion Reactor Safety, Jackson, Wyoming (1989).
[15] Stairmand J.W. Graphite Oxidation - A Literature Survey, AEA Technology Report
AEA-FUS-83 (1990).
[16] Thurlbeck A. Summarised Compatibility Review of Reactor Materials for CO2 -cooled
Graphite Moderated Reactors, TRG Report 267(W) (1962).
[17] Walker P.L., Rusinko F. and Austin L.G. Advances in Catalysis, 11, 123 (1959).
[18] Welch E.E. DTO Applied to Graphite/Carbon Deposit Systems - Paper 1, Principles and
Problems, Symposium on Differential Thermal Oxidation held at Berkeley Nuclear
Laboratories (1972).
[19] Wickham A.J., Marsden B.J., Pilkington N.J. and Heath T.G. The Possibility and
Consequences of Graphite Core Degradation during ‘Care and Maintenance’ and
‘Safestore’, AEA Technology Report AEA/16423595/R/001 (1996).

213
214