Electronic Supplementary Information

1. Graphite oxide (GO) decomposition kinetics

All the numerical calculations in this work have been done with either Python or R
project for statistical computing software.
The reaction rate of the type solid ---> solid + gas reaction could be expressed as:

Ea
dx
= k(1 − x)n = Ae−RT (1 − x)n (1)
dt

where k is reaction rate constant, x is the mass fraction reacted (from 0 to 1), dx/dt
is reaction rate, n is the empirical reaction order which is assumed to stay constant
through most part of the reaction, Ea is activation energy and A frequency factor.
Decomposition reaction activation energy and frequency factor were estimated
applying an equation from Kissinger [1]:

𝜙 𝐴𝑅 𝐸
ln (𝑇 2 ) = ln ( 𝐸 ) − 𝑅𝑇𝑎 (2)
𝑚 𝑎 𝑚

where 𝜙 is heating rate, Tm is the temperatures at which the heat flow reaches
maximum during various heating rate cycles and R universal gas constant.
𝜙 1
Plotting ln (𝑇 2 ) vs. offers linear regression, from which activation energy Ea
𝑚 𝑇𝑚

and frequency factor A are calculated.

The empirical reaction order n can be estimated by applying the following
equation[1]:

𝑛 = 1.26 𝑆 0.5 (3)

where S is curve shape index defined as the absolute value of the ratio of the slopes
of tangents to the DSC curve at the two inflection points.
In order to verify the validity of our kinetic parameters determined, the actual
heat flow was calculated applying these parameters and using the following
equation:

Ea
𝑑𝑞
= −∆𝐻Ae−RT (1 − 𝑥)𝑛 (4)
𝑑𝑡

𝑑𝑞
where ΔH is the enthalpy of decomposition, , is the change of heat in time.
𝑑𝑡

2. Self-heating analysis
The governing equation for non-adiabatic energy balance of self-heating can be
written as:

𝑑𝑇 𝑑𝑥
𝑚𝐶𝑝 𝑑𝑡 = −𝑚∆𝐻 𝑑𝑡 − ℎ𝑆(𝑇 − 𝑇0 ) (5)

where m is mass of GO, Cp is specific heat capacity of GO (assumed to be 0.5

Jg-1K-1), x is mass fraction reacted, ΔH is the enthalpy of decomposition, h is
convective heat transfer coefficient, T0 is temperature of the surrounding
environment (also the initial temperature of GO) and S, lateral surface area of bulk
GO exposed to gas atmosphere.
The term on the left hand side in equation 5 is the energy accumulation term
(which at steady state is zero), and the terms in right hand side of the equation 5
are the decomposition reaction heat generated or lost (depending if the reaction is
exo- or endothermic) and convectional heat loss. We ignore radiation heat losses or
gains in this work.
To study the complex self-heating phenomenon, we begin with the simplest case,
i.e. adiabatic self-heating, for which the heat transfer term is ignored in equation 5
and the governing equation transforms to:

𝑑𝑇 𝑑𝑥
𝑚𝐶𝑝 𝑑𝑡 = −𝑚∆𝐻 𝑑𝑡 (6)

After integrating equation 6 and approximating n = 1 one can show the following
relation for the time t for self-heating,

𝑐 𝑐𝑏
𝐸𝑖( ) 𝑒 𝑏𝑇0 +1 1 𝑏
𝑡= 𝑇
− ∗ 𝐸𝑖(𝑇 − 𝑏𝑇 +1)| 𝑇𝑇 (7)
𝑎 𝑎 0 0

∆𝐻𝐴 𝐶𝑝
where Ei(x) is exponential integral, constants a, b ,c are defined as － , － ∆𝐻,
𝐶𝑝

𝐸𝑎
, respectively.
𝑅

In this calculation the following numbers are applied to determine constants a, b

and c: heat of decomposition, ΔH = -1600 Jg-1, specific heat capacity Cp=0.5
kJkg-1K-1 (which is a value in the common range of most oxygen-containing
carbon materials), A= 3.5*1013 s-1, Ea= 142 kJmol-1 (both values estimated from
kinetic analysis). Time t is calculated for given T in Equation (7), and from which
self-heating T-t plots are created.
For a more realistic scenario, heat loss due to convection should be taken into
account as shown in equation 5. In the analysis below, we estimate convective heat
transfer coefficient h and the equation 5 can be rewritten as:

𝑑𝑇 ∆𝐻𝐴 𝐶𝑝 (𝑇 − 𝑇0 ) 𝑛 − 𝐸𝑎 ℎ𝑆
=− (1 + ) 𝑒 𝑅𝑇 − (𝑇 − 𝑇0 )
𝑑𝑡 𝐶𝑝 ∆𝐻 𝑚𝐶𝑝
from which we can show the following Equation 8:

𝑇 𝑑𝑇
𝑡 = ∫𝑇 −
𝑐 (8)
0 𝑎(1−𝑏(𝑇−𝑇0 ))𝑛 𝑒 𝑇 −𝑑(𝑇−𝑇0 )
∆𝐻𝐴 𝐶𝑝 𝐸𝑎 ℎ𝑆
where constants a, b, c and d are defined as − , − ∆𝐻 , and ,
𝐶𝑝 𝑅 𝑚𝐶𝑝

respectively.
There is no analytic form of solution for the integral on the right hand side in
equation 8, but numerical results is calculated.

3. Thermochemistry of the decomposition reaction, estimating GO enthalpy of formation

applying group contribution method

The GO thermal reduction is essentially a disproportionation reaction and can be

shown as follows:

GO−→ rGO + x CO2 + y CO + z H2 O (9)

As determined by Jung et al, 2009 the x:y ratio during GO decomposition is

around 2 [2]. The coefficient z is estimated by assuming all the hydrogen present
in GO is converted into H2O.
The enthalpy of decomposition ΔH at standard conditions can thus be estimated by
subtracting the enthalpy of formation of reactants from those of the products if the
enthalpy of formation of bulk GO and reduced GO at standard conditions, i.e.
ΔHf,GO and ΔHf,rGO are known.
The standard enthalpy of formation for bulk GO ΔHf,GO (or ΔHf,rGO for rGO)
consists of two parts. Namely, the enthalpy of formation for monolayer GO
ΔHf,monolayer GO (or ΔHf, monolayer rGO for rGO) and stacking enthalpy for multilayer
structure ΔHstack. The value of the latter one is still in controversy and a
well-accepted value is not available [3-8]. However, if we assume that stacking
energy of GO and rGO is similar and relatively small compared with the GO
enthalpy of formation, the influence of stacking enthalpy can be offset when
calculating ΔHd by using ΔHf, monolayer GO and ΔHf, monolayer rGO instead of ΔHf, GO and
ΔHf, rGO.
The ΔHf, monolayer GO and ΔHf, monolayer rGO were estimated with a group contribution
method developed and reported by J. Marrero and R. Gani [9]. This method
estimates compounds thermodynamic properties based on their molecular
structures at three levels. The first level contracts with the proximity effects. The
second level contracts with the polyfunctional, aromatic, polar/non-polar effects,
and the third level with the more complex structural fragment effects whose
description is insufficient by the first and second levels prediction. The model
molecular structure of GO is a monolayer sheet consisting of 192 carbon atoms
with a 12 by 7 array of six-membered rings with varying amount of epoxide or
hydroxyl groups on its surface and edges. To minimize the influence of edge sites,
we assumed a periodic boundary condition, i.e. the sheet is an infinite 2D structure.
The molecular structure of rGO is similar but with no heteroatoms other than
carbon on the sheet. If the enthalpies of formation of GO and rGO are determined
the enthalpy of decomposition at standard conditions can be calculated, since the
rest of the enthalpies of formation in equation 9 are known.

4. The influence of reduction agents during thermal decomposition

GO samples before and after incubation with hydrazine and sodium borohydride
(NaBH4) were also investigated. In each case, as shown in Figure 1S, partial
chemical reduction lowers the enthalpy of thermal decomposition, but raises the
reactivity of the product sample (chemically reduced GO), seen as a lowering of
the thermal decomposition onset temperature. Washing with DI water after
reduction, but before DSC experiment cannot eliminate this effect, implying the
chemical change of GO structure rather than the reducing reagent is the cause of
this thermal behavior change. A complete chemical reduction by 100 mg mL-1 of
hydrazine totally eliminates the enthalpy of decomposition (see Figure 1S A).

Fig. 1S - Chemical reduction effects on GO thermal decomposition. The GO samples

used for these characterizations are recovered and dried from suspension incubated
with hydrazine or NaBH4. DSC thermograms of hydrazine (A) or NaBH4 (B) treated
GO. In both cases, partial reduction lowers the enthalpy of decomposition, but at
the same time, chemical reduction lowers decomposition reaction onset temperature.
The washed sample shows similar behavior. Complete reduction with high
concentration reducing agents however removes most of the exothermicity, probably
due to the complete removal of active oxygen function groups.

References

[1] Kissinger HE. Reaction Kinetics in Differential Thermal Analysis. Anal Chem.
1957;29(11):1702-6.
[2] Jung I, Field DA, Clark NJ, Zhu YW, Yang DX, Piner RD, et al. Reduction Kinetics
of Graphene Oxide Determined by Electrical Transport Measurements and Temperature
Programmed Desorption. J Phys Chem C. 2009;113(43):18480-6.
[3] Allinger NL, Yuh YH, Lii JH. Molecular Mechanics - the Mm3 Force-Field for
Hydrocarbons .1. J Am Chem Soc. 1989;111(23):8551-66.
[4] Charlier JC, Gonze X, Michenaud JP. Graphite Interplanar Bonding - Electronic
Delocalization and Van-Der-Waals Interaction. Europhys Lett. 1994;28(6):403-8.
[5] Jeong HK, Lee YP, Lahaye RJWE, Park MH, An KH, Kim IJ, et al. Evidence of
graphitic AB stacking order of graphite oxides. J Am Chem Soc. 2008;130(4):1362-6.
[6] Marchand D, Fretigny C, Lagues M, Batallan F, Simon C, Rosenman I, et al.
3-Dimensional Band-Structure of Graphite Studied by Angle-Resolved Photoemission
Using Ultraviolet Synchrotron Radiation. Phys Rev B. 1984;30(8):4788-95.
[7] Schabel MC, Martins JL. Energetics of Interplanar Binding in Graphite. Phys Rev B.
1992;46(11):7185-8.
[8] Zacharia R, Ulbricht H, Hertel T. Interlayer cohesive energy of graphite from
thermal desorption of polyaromatic hydrocarbons. Phys Rev B. 2004;69(15).
[9] Marrero J, Gani R. Group-contribution based estimation of pure component
properties. Fluid Phase Equilibr. 2001;183:183-208.