Summary: all these reactions are possible under the right conditions (for example, temperature, pressure, phase, and catalyst); however, for thermal decomposition at atmospheric pressure, the first reaction to become spontaneous (at about $\pu {400 k}$, called the thermal decomposition temperature of $\ce{NaHCO3(s)}$), is the decomposition of $\ce{NaHCO3(s)}$ into $\ce{Na2CO3(s)}$, $\ce{H2O(g)}$, and $\ce{CO2(g)}$.
There is no (accurate and entirely) theoretical way to predict a reaction (as of now). However, we can check if a reaction is possible based on thermodynamic-, to what extent it occurs based on equilibirum-, and weather it happens in real-time based on kinetic-analysis.
Note: This are all semi-empirical analysis methods since the values of thermodynamic potentials, equilibrium constants and kinetic rate coefficients are often obtained experimentally.
I will do the thermodynamic analysis here.
Thermodynamic Analysis
First, let us convert all your reaction-components into standard states in the simplest stoichiometries.
Reactions
At $\pu{298.15 k}$, the reactions are.
$$
\ce{2 NaHCO3(s) -> Na2CO3(s) + H2O(l) + CO2(g)} \tag{1}
$$
$$
\ce{4 NaHCO3(s) -> 4 Na + 2 H2O(l) + 4 CO(g) + 3 O2(g)} \tag{2}
$$
$$
\ce{NaHCO3(s) -> NaOH(s) + CO2(g)} \tag{3}
$$
At $\pu{400 k}$, the reactions are:
$$
\ce{2 NaHCO3(s) -> Na2CO3(s) + H2O(g) + CO2(g)} \tag{1}
$$
$$
\ce{4 NaHCO3(s) -> 4 Na + 2 H2O(g) + 4 CO(g) + 3 O2(g)} \tag{2}
$$
$$
\ce{NaHCO3(s) -> NaOH(s) + CO2(g)} \tag{3}$$
This is because of the phase transition $\ce{H2O(l) \xrightarrow{T>\pu{373.15 K}} H2O(g)}$
Data
The following table shows the standard enthalpies and free energies of formation for each component at $\pu {298.15 K}$ and $\pu {400 K}$.$^1$ All values are in $\pu{kJ mol^{-1}}$
Chemical |
$\Delta_\text{f} H^\circ$($\pu {298.15 K}$) |
$\Delta_\text{f} H^\circ$($\pu {400 K}$) |
$\Delta_\text{f} G^\circ$($\pu {298.15 K}$) |
$\Delta_\text{f} G^\circ$ ($\pu {400 K}$) |
$\ce{O2(g)}$ |
$\pu{0}$ |
$\pu{0}$ |
$\pu{0}$ |
$\pu{0}$ |
$\ce{H2O(g)}$ |
— |
$\pu{-242.847}$ |
— |
$\pu{-223.951}$ |
$\ce{H2O(l)}$ |
$\pu{-285.830}$ |
— |
$\pu{-237.141}$ |
— |
$\ce{Na(s)}$ |
$\pu{0}$ |
$\pu{0}$ |
$\pu{0}$ |
$\pu{0}$ |
$\ce{NaHCO3(s)}$ |
$\pu{-950.810}$ |
$\pu{-953.885}$ |
$\pu{-852.851}$ |
$\pu{-819.095}$ |
$\ce{NaOH(s)}$ |
$\pu{-425.931}$ |
$\pu{-428.299}$ |
$\pu{-379.737}$ |
$-363.788$ |
$\ce{Na2CO3(s)}$ |
$\pu{-1030.768}$ |
$\pu{-1135.717}$ |
$\pu{-1048.005}$ |
$\pu{-1019.363}$ |
$\ce{CO(g)}$ |
$\pu{-110.541}$ |
$\pu{-110.129}$ |
$\pu{-137.180}$ |
$\pu{-146.354}$ |
$\ce{CO2(g)}$ |
$\pu{-393.505}$ |
$\pu{-393.580}$ |
$\pu{-394.364}$ |
$\pu{-394.646}$ |
Calculations
Now, let us calculate heats and free energies of reactions (1), (2), and (3) using $\Delta_r H^\circ = \sum_\text{products} \Delta_f H^\circ - \sum_\text{reactants} \Delta_f H^\circ$ and $\Delta_r G^\circ = \sum_\text{products} \Delta_f G^\circ - \sum_\text{reactants} \Delta_f G^\circ$ (considering the stoichiometry). All values are in $\pu{kJ mol^{-1}}$
Reaction |
$\Delta_r H^\circ$ ($\pu{298.15 k}$) |
$\Delta_r H^\circ$ ($\pu{400 k}$) |
$\Delta_r G^\circ$ ($\pu{298.15 k}$) |
$\Delta_r G^\circ$ ($\pu{400 k}$) |
1 |
$\pu{119.517}$ |
$\pu{135.626}$ |
$\pu{26.192}$ |
$\approx \pu{0}$ |
2 |
$\pu{1657.56}$ |
$\pu{1755.526}$ |
$\pu{1359.666}$ |
$\pu{1249.894}$ |
3 |
$\pu{131.374}$ |
$\pu{132.006}$ |
$\pu{78.750}$ |
$\pu{60.661}$ |
Results and Discussion
This tells us that all the reactions are endothermic and non-spontaneous except for reaction (1) at $\pu{400K}$ for which $\Delta_r G = 0.230 \approx 0$. This is (nearly) called the thermal decomposition temperature of $\ce{NaHCO3}$ to convert to $\ce{Na2CO3}$.
This also tells us is that the reactions are endothermic and non-spontaneous at ambient temperature. This why $\ce{NaHCO3}$ is required to be heated for the decomposition. According to reference$^2$ the thermal decomposition temperature of $\ce{NaHCO3}$ is $\pu{373.70 k}$, but our value is closer to $\pu{400 k}$. You can obtain even better (negative values of $\Delta_r G$ for reaction (1) at higher temperatures.
References
- Thermodynamic Data of Pure Substances. Ihsan Barin, Gregor Platzki. Third edition. VCH Verlagsgesellschaft mbH, Weinheim (Federal Republic of Germany), VCH Publishers, Inc., New York, NY (USA)
- Thermal Decomposition of Sodium Hydrogen Carbonate and Textural Features of Its Calcines. Miloslav Hartman, Karel Svoboda, Michael Pohořelý, and Michal Šyc. 10.1021/ie400896c
P.S. Could someone check the data and calculations once more (because the decomposition temperature didn't match with that of reference [2]). I just want to be sure. What is causing the $\Delta_r G$ to be slightly positive even above the thermal decomposition temperature$^2$ of $\ce{NaHCO3}$?