DOI: 10.3303/CET25117192 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
Paper Received: 29 December 2024; Revised: 25 February 2025; Accepted: 29 May 2025 
Please cite this article as: Gicquel L., Lefèvre S., Moinot L., Tardivel B., 2025, Innovative Sodium Cleaning Process Using Saline Solutions, 
Chemical Engineering Transactions, 117, 1147-1152  DOI:10.3303/CET25117192 
  

 CHEMICAL ENGINEERING TRANSACTIONS  
 

VOL. 117, 2025 

A publication of 

 

The Italian Association 

of Chemical Engineering 

Online at www.cetjournal.it 

Guest Editors: Fabrizio Bezzo, Flavio Manenti, Gabriele Pannocchia, Almerinda di Benedetto 

Copyright © 2025, AIDIC Servizi S.r.l. 

ISBN 979-12-81206-17-5; ISSN 2283-9216 

Innovative Sodium Cleaning Process using Saline Solutions  

Leïla Gicquel*, Sébastien Lefèvre, Laure Moinot, Benjamin Tardivel 

CEA, DES, IRESNE, DTN, Cadarache, F-13108, Saint-Paul-lez-Durance, France 

 leila.gicquel@cea.fr 

During the development of NOAH process at CEA in the 1980s (Rodriguez, 2021), the addition of sodium 

hydroxide to water was identified as an interesting way to moderate the kinetics of the sodium-water reaction. 

Based on these observations, studies were launched at CEA in 2007 with the aim of developing an innovative 

process for treating remaining sodium in the components of a sodium fast reactor (SFR). This process would 

have the following advantages: control of the sodium-water reaction and limitation of its effects by adding an 

appropriate salt; flexibility and increase in treatment rates by rapid immersion of the components. In order to 

overcome the corrosive nature of sodium hydroxide, new salts were studied. Several salts appeared promising 

in that, for some of them, the reaction no longer generates a shock wave despite an excess of water, the kinetics 

are slowed down and the release of energy occurs gradually in the reaction medium. Following these results, 

new experimental campaigns were launched and enabled to identify the influential parameters and to quantify 

the effects of two salts on the moderation of the sodium-water reaction: tetrasodium EDTA and sodium acetate. 

1. Introduction 

As part of the development of the IVth Generation of nuclear reactors, studies were launched at CEA in 2007 

with the aim of developing an innovative process for treating sodium in the components of a sodium fast reactor 

(SFR). A fuel assembly has a limited lifetime and when the fuel loses reactivity, the assemblies are removed 

from the reactor core and transported to deactivation pools after treating the residual sodium on their surface. 

To do this, the method used consists of reacting sodium and water in cleaning pits or cells. Since the sodium-

water reaction (SWR) can be violent and generate an explosion phenomenon, the cleaning times for the 

assemblies are quite long. New processes would make it possible to control the SWR and limit its effects by 

adding an appropriate salt, with significant benefits such as flexibility and increased processing rates thanks to 

the rapid immersion of components. In order to overcome the corrosive nature of sodium hydroxide, several 

salts were studied at CEA (Lacroix, 2014), some of which appeared promising such as lithium acetate, 

tetrasodium EDTA or even iron(III) chloride. The reaction no longer generated a shock wave despite an excess 

of water, the kinetics were slowed down and the release of energy occurred gradually in the reaction medium. 

These results led to the publication of a patent (Lacroix et al., 2015). A new experimental campaign was 

launched at the end of 2023 to identify influential parameters and quantify the effects of salts on SWR 

moderation (Gicquel et al., 2024). Following these experiments, improvements were made to the device, and a 

new experimental campaign was launched, with the aim of providing information on the chemical kinetics. 

2. Methodology 

2.1 Sodium-water reaction 

The sodium-water reaction (SWR) is an exothermic reaction with a reaction enthalpy of -184 kJ/molNa, taking 

into account the dissolution of sodium hydroxide in water. The reaction equation is the following: 

Na (s) + H2O (l) → NaOH (l) +1/2 H2 (g)     (1) 

In addition to the exothermic nature of the reaction, SWR is almost instantaneous, produces sodium hydroxide, 

which can lead to corrosion problems and hydrogen, which is a flammable and explosive compound depending 

1147



on its concentration (between 4 and 75 % in volume in air) and the presence of oxygen in the environment. 

SWR is explosive, leading to overpressure due to the pressure wave released (Figure 1). 

 

Figure 1: Sodium-water reaction with a 4-gram pellet of sodium 

2.2 Experimental device 

The experiments have been carried out in LAVINO device (Innovative Cleaning), presented in the Figure 2. This 

device comprises several parts including: 

• A 3.5 L reactor. For each experiment, 700 mL of solution were introduced into the reactor, leaving a 

total of 2.8 L of gaseous phase volume; 

• A sodium pellet introduction device. This device allows the pellet to be transported from the glove box 

where it has been manufactured to the reactor, while avoiding contact between the sodium and the 

surrounding air. Once placed on top of the reactor, it is connected to a pneumatic cylinder that allows 

the basket containing the sodium pellet to be immersed into the solution; 

• Three thermocouples (in the solution, in the gas and on the reactor wall), a dynamic pressure sensor, 

and a gas analysis board (O2, H2). These acquisition tools enable data collection on each experiment; 

• A high-speed Phantom camera (1,000 – 28,000 fps, resolution 1,024 x 768) as well as a GoPro camera, 

allowing the reaction to be filmed with varying degrees of precision through the reactor's portholes. 

 

Figure 2: Picture and sketch of LAVINO facility 

2.3 Theoretical calculations 

The experimental data is used primarily for energy balance between, before and after the reaction. The LAVINO 

reactor is an irreversible closed adiabatic system whose gas phase is assimilated to a mixture of perfect gases. 

Thus, the Antoine’s law is used to calculate the saturated vapour pressure of water: 

ln(𝑃𝑠𝑎𝑡) = 𝐴 −
𝐵

𝐶+𝑇
  (2) 

And the mole numbers of each constituent are calculated from the equation: 𝑃𝑖 =
𝑛𝑖

𝑛𝑡𝑜𝑡
𝑃     (3) 

The enthalpy of such a mixture depends only on the temperature according to the first law of Joule for 

monoatomic or diatomic gas: ∆𝐻 =
5

2
𝑛𝑅∆𝑇, ∆𝐻 =

7

2
𝑛𝑅∆𝑇        (4) 

1148



The molar latent heat of vaporization of water will be added to complete the calculation of the variation of 

enthalpy in the gas phase. For the solution, the variation of enthalpy shall be calculated as:  

∆𝐻 = 𝑚𝐶𝑝∆𝑇          (5) 

The thermal capacity used will be approximated by that of water, for which no data have been found for high 

concentration solutions. The mass is calculated using the density of the solution while considering that 700 mL 

of sodium acetate solution was introduced. For EDTA, due to lack of data, water density will be used. 

Theoretically, it is expected that: 𝑄𝑡ℎ𝑒𝑜 = ∆𝐻𝑠𝑜𝑙 + ∆𝐻𝑔𝑎𝑧                       (6) 

assuming that the change in enthalpy of the gas phase is negligible compared to that of the solution, the 

theoretical temperature difference of the solution should be such that: ∆𝑇 =
𝑛.∆𝐻𝑟

𝑚.𝐶𝑝
    (7) 

2.4 Control tests with water 

Control tests (CT) were conducted with 2-gram balls and 4-gram balls. SWR trials “explode” almost 

systematically according to the literature. However, of the five tests with 2-gram balls, only one was observed 

to explode. The peak pressure phenomenon associated with the explosion occurred for both 4-gram tests. 

 

Figure 3: Comparison of two conventional SWR; left: control test 2; right: control test 3 

These graphs (Figure 3) show the difference in reaction behaviour. The reaction starts with a fairly similar 

behaviour with a fairly regular pressure rise and then a slight stabilisation (the reaction slows down), the 

difference being mainly on the absence of the explosion phenomenon compared to the control test 2 whose 

reaction ends at that moment. For the other tests, a more marked reaction is observed and the ball is completely 

consumed. Table 1 compares the difference in gas phase pressure between the initial and final states of the 

hydrogen production, as well as the difference in maximum pressure reached. The differences are mainly due 

to the reactor pressure drop of about 30 mbar per minute, but also to unexpected sodium combustion 

phenomena which will be discussed in section 3. 

Table 1: Energy balances of several control tests 

Test 
mNa/S 

(g/cm2) 
QTheo (kJ) QExp (kJ) Rel. Dev.*  ΔPTheo (bar) ΔPExp (bar)**  Rel. Dev. 

CT  1 0.26 15.93 12.74 20.0% 0.382 0.303 / 0.465 2.6% 

CT  2 0.26 15.85 17.55 10.7% 0.375 0.294 / 1.13 2.7% 

CT  3 0.26 16.01 11.95 25.3% 0.378 0.287 / 0.494 3.2% 

CT  4 0.26 16.25 10.61 34.7% 0.387 0.328 / 0.627 1.8% 

*relative deviation, **first value is ending pressure, second value is maximum pressure during the test 

The maximum pressure (ΔPExp max) indicates whether the reaction is explosive (CT 2) or not (CT 1, 3 and 4). 

In the case of control tests 1, 3 and 4, the pressure variation exceeds the expected one, despite reactor pressure 

drop, due to small explosion of a piece of sodium on the surface as well as an increase in the gas-phase 

temperature from the combustion of the sodium. 

3. Experimental results with saline solutions 

This new experimental campaign focused on the study of sodium acetate and tetrasodium EDTA. These salts 

were chosen based on their performances in moderation as well as their costs or environmental issues. Each 

experiment was conducted twice, in order to determine the repeatability between two experiments under the 

same operating conditions. 

1149



3.1 Results with sodium acetate (NaOAc) 

Of the 20 tests carried out, the explosion phenomenon with pressure peak only occurred for two tests (8.1 and 

8.2). At a concentration of 1 mol/L, the explosion phenomenon disappears when a 2-gram pellet reacts, but this 

phenomenon is still present during the same reaction with a pellet twice as heavy and therefore thicker. In terms 

of concentration, it is clear that a 3 mol/L solution is more effective than a 1 mol/L solution. The comparison is 

based first on the pressure rise, which appears smoother and more regular at higher concentrations, and also 

in terms of reaction time, which is much longer when using a 3 mol/L sodium acetate solution. With a 4-gram 

pellet and 1 mol/L solution, the pressure peak appears after 6 seconds, whereas it appears after 4 seconds for 

SWR. This delay can be attributed to the difference of oxide’s layers or due to an ionic layer in solution. These 

hypotheses will be studied during a PhD to come in the laboratory. 

 

Figure 4: Pressure and temperature curves for test 2.1; mNa = 2 g; C°NaOAc = 2.5 mol/L 

The graphs presented on Figure 4 and Figure 5, representing pressure and temperatures as a function of time, 

demonstrate whether the reaction is well moderated or not. At a concentration of 1 mol/L, the explosion 

phenomenon disappears when a 2-gram pellet reacts, but this phenomenon is still present during the same 

reaction with a pellet twice as heavy (and therefore thicker). On Figure 5 the pressure peak is correlated to the 

increase of temperature in liquid and gas phases, linked to the exothermicity of reaction released in the medium. 

 

Figure 5: Pressure and temperature curves for test 8.1; mNa = 4 g; C°NaOAc = 1 mol/L 

The Table 2 compares the pressure difference in the gas phase between the initial state and the final state 

reflecting the production of dihydrogen, as well as the maximum pressure difference reached. The theoretical 

value is never reached due to several factors, but the experimental values still demonstrate the production of 

dihydrogen.  

Table 2: Energy and pressure balances for several tests with sodium acetate 

Test 
mNa/S 

(g/cm2) 

[Salt] 

(mol/L) 

QTheo 

(kJ) 

QExp 

(kJ) 

Rel. 

Dev.  
ΔPTheo (bar) ΔPExp (bar)  

Rel. 

Dev. 

1.1 0.26 1.0 16.25 11.84 27.1% 0.391 0.350 10.4% 

4.1 0.26 1.5 16.17 15.45 4.5% 0.381 0.352 7.6% 

6.1 0.20 1.0 8.01 3.46 56.8% 0.190 0.160 15.7% 

8.1 0.32 1.0 31.78 31.81 0.1% 0.748 0.701 6.2% 

The deviations are mainly due to the reactor pressure loss, of the order of 30 mbar per minute, but also to a 

sodium combustion phenomenon, which is unexpected and will be discussed in section 3. For energy balance, 

supposing that the enthalpy variation in the gas phase is negligible compared to that of the liquid phase, we 

supposed that all the sodium is consumed and the enthalpy of reaction fully transferred to the liquid phase. The 

higher the temperature in the gas phase, the less the temperature variation corresponds to that expected, with 

the exception of the tests with 1-gram balls, found here. It should also be noted that the salt concentration 

influences this temperature because the reaction is less violent at high concentration. The relative difference is 

1150



the lowest for the only test where the pressure wave phenomenon occurred and test 6.1 with a  

1-gram ball has a significant difference compared to the other tests. For the sodium acetate solutions, at the 

same concentration, the results are showing a greater reaction rate when the sodium mass increases. This can 

be supported by the fact that at 2 mol/L, 4 grams of sodium burn at a rate of 0.24 g/s while 2 grams burn at a 

rate of about 0.09 g/s. 

3.2 Results with tetrasodium EDTA 

All these trends can also be observed in tests with tetrasodium EDTA solutions (Figure 6 and Figure 7). For a 

concentration of 0.1 mol/L, the reaction is moderated whether it is for a pellet of 1, 2, or 4 grams; however, it is 

not for a concentration of 0.075 mol/L with a 2-gram pellet or a concentration of 0.05 mol/L with a 4-gram pellet. 

 

Figure 6: Pressure and temperature curves for test 11.1; mNa = 4 g; C°EDTA = 0.1 mol/L 

 

Figure 7: Pressure and temperature curves for test 12.2; mNa = 4 g; C°EDTA = 0.05 mol/L 

On the temperature graph of Figure 6 (right side), it can be seen that the temperature measured in the gas 

phase is increasing significantly during the test until almost 40°C. The higher the gas phase temperature, the 

less the temperature variation corresponds to that expected, with the exception of the experiments with 1-gram 

pellets, for which the temperature variations are lower than expected. In addition, it seems that when the 

explosion phenomenon occurs, the temperature variation is closer to the theory. However, the temperature in 

the gas phase does not seem to depend on the mass of sodium involved, nor on the salt concentration. This 

means that the high temperature rise in the gas phase is not only related to the quantity of dihydrogen produced 

and to sodium combustion. When the reaction is moderated, the pressure rise only slightly exceeds the expected 

final variation, showing a successful moderation of the reaction. The differences between theoretical and 

experimental overpressures are rather satisfactory. It is also possible to say that the explosion phenomenon 

does not affect the production of dihydrogen. It was shown in a previous campaign that the uncertainty of the 

pressure sensor can be taken equal to 10% of the measurement made for low pressure measurements (< 2 bar).  

Table 3: Energy and pressure balances for tests with tetrasodium EDTA 

Test 
mNa/S 

(g/cm2) 

[Salt] 

(mol/L) 

QTheo 

(kJ) 
QExp (kJ) 

Rel. 

Dev.  
ΔPTheo (bar) ΔPExp (bar)  

Rel. 

Dev. 

9.2 0.20 0.10 8.164 5.116 37.3% 0.195 0.167 14.3% 

10.2 0.26 0.10 16.008 12.356 22.8% 0.380 0.318 16.3% 

11.2 0.32 0.10 32.015 28.191 11.9% 0.760 0.717 5.6% 

12.2 0.03 0.05 31.775 30.647 3.5% 0.742 0.675 9.0% 

As with the sodium acetate tests, it is possible to determine experimental consumption rates. The conclusion is 

that a 2-gram pellet and a 4-gram pellet are consumed at the same rate at a concentration of 0.1 mol/L. Indeed, 

these two experiments led to a consumption rate of around 0.27 g/s. Conversely, a 1-gram pellet has a slower 

consumption rate, around 0.06 g/s. 

3.3 Discussion 

The reaction of sodium with salt solutions can be described according to the following 3 steps (Figure 8): (1) 

dissolution of soda/oxide layer: start of reaction, first hydrogen bubbles appear; (2) runaway of the reaction: 

1151



formation of an unstable gas film with large bubbles that coalesce and move to the surface; (3) slowdown of the 

reaction: the gas film around the bubbles is stabilizing, few bubbles are released until the end of the reaction. 

 

Figure 8: Classic pathway of a mitigate reaction 

The characteristic white smoke of sodium oxide production and the red/orange colour of the particle show that 

a combustion phenomenon is almost systematic during each test, with abnormal variations of the gas phase 

temperature. This implies that all sodium does not react with water, according to the SWR reaction balance 

equation. As explained above, sodium samples are prepared in an argon-inert glove box and stored in jars. 

Unfortunately, the glove box used still contains enough traces of oxygen and water (< 200 ppm) causing 

oxidation of the sodium ball on its surface (Figure 9), but it is impossible to estimate the oxidation layer and 

therefore the mass of pure sodium lost. SWR being a surface reaction, it cannot take place as long as there is 

an oxide layer around the ball. The enthalpy of this reaction is about -330 kJ/mol (with dissolution of soda) and 

does not produce dihydrogen. This could therefore be a cause of deviations in the expected overpressure and 

the energy balance, with an enthalpy of reaction almost twice as high as a SWR. 

 

Figure 9: Comparison of a pure sodium ball (on left) with an oxidised ball 

4. Conclusion 

These experimental results provide first trends in the moderation of the sodium-water reaction. In order to 

improve the results, in the aim of establishing the reaction kinetics, future experiments will be performed with 

some modifications. Specific weight moulds will be used to manufacture the sodium pellets in a pure argon 

environment. This modification will allow to confirm the concentration thresholds (perfectly spherical pellets, no 

oxide crust on the surface). In addition, some evolutions of LAVINO device will be carried out in order to be able 

to collect samples of the liquid solution and the gas during the reaction. These modifications will allow to 

determine the reaction kinetics in a near future for an industrial scale-up. This new process with saline solutions, 

seems to be a good alternative to traditional water cleaning for big sodium fast reactors components and fuel 

assemblies treatment in order to optimise handling and dismantling operations. 

Acknowledgments 

The authors acknowledge former internship Sacha Bruno, the former research team who worked on this subject 

in CEA Cadarache and the project managers and team managers for their financial and material support. 

References 

Gicquel L., Lefèvre S., Bruno S., Moinot L., Tardivel B., 2024, GLOBAL2024, Tokyo, Japan. 

Gicquel L., Lefèvre S., Bruno S., Moinot L., Tardivel B., 2024, 19ème Congrès de la SFGP, Deauville, France. 

Lacroix M., 2014, Étude de l’effet modérateur de sels sur la réaction sodium-eau pour le lavage d’assemblages 

de combustible irradiés issus de réacteurs à caloporteur sodium, PhD, Pierre and Marie Curie University. 

Lacroix M., Lorcet H., Perrais C., 2015, Procédé de traitement du sodium par une solution de sel aqueux et 

procédé de lavage d’assemblages combustibles de réacteur nucléaire utilisant ledit procédé de traitement, 

WO2015/110480. 

Rodriguez G., 2021, Le caloporteur sodium, Tech Ingénieur, BN 3680v2. 

1152


	CET-vol117-b.pdf
	194gicquel.pdf
	Innovative Sodium Cleaning Process using Saline Solutions