- Email: [email protected]
Application of a hydroxylamine nitrate stability model to plutonium purification process equipment
Journal of Loss Prevention in the Process Industries 35 (2015) 12e18
Contents lists available at ScienceDirect
Journal of Loss Prevention in the Process Industries journal homepage: www.elsevier.com/locate/jlp
Application of a hydroxylamine nitrate stability model to plutonium purification process equipment G. Scott Barney a, *, Paul B. Duval b a b
105903 N. Harrington Rd., West Richland, WA, USA Shaw AREVA MOX Services, LLC, Aiken, SC, USA
a r t i c l e i n f o
a b s t r a c t
Article history: Received 17 February 2014 Received in revised form 2 February 2015 Accepted 5 March 2015 Available online 6 March 2015
A mathematical model that predicts hydroxylamine nitrate (HAN) (NH2OH$HNO3) stability is applied to aqueous solutions containing HAN, nitric acid and plutonium that are used in plutonium purification processes. The model estimates the stability of these solutions with respect to the rapid, hazardous, autocatalytic reaction of HAN with nitric acid that generates heat and gas. It also accounts for reaction kinetics, temperature changes, gas generation rates, solution volumes and flow rates, and distribution of plutonium and nitric acid between aqueous and organic phases. The model is applied to three typical process vessels used in solvent extraction purification of plutonium e a countercurrent aqueous/organic plutonium stripping column, an oxidation column used for HAN and hydrazine destruction, and a plutonium rework tank. Both normal and off-normal process scenarios are modeled. Two of the offnormal scenarios lead to the rapid autocatalytic reaction of HAN with nitric acid where heat and gas are generated and that could lead to damage of the process equipment and/or release of hazardous plutonium solution from the vessel. In these two cases, stationary aqueous solutions containing HAN, Pu(III), and nitric acid were allowed to slowly react until conditions for the autocatalytic reaction were reached. © 2015 Elsevier Ltd. All rights reserved.
Keywords: Hydroxylamine stability Plutonium Kinetics Nitric acid Model Solvent extraction Process equipment
1. Introduction A mathematical model for determining the stability of aqueous hydroxylamine nitrate (HAN) (NH2OH$HNO3) solutions used in solvent extraction purification of plutonium was described in an earlier publication (Barney and Duval, 2011). The model calculates rates of the reactions involved in oxidation of HAN under conditions existing in a typical plutonium purification solvent extraction process. Conditions for which the hazardous autocatalytic reaction of HAN with nitric acid occurs were predicted successfully with the model. This paper describes application of the model for predicting HAN stability in solvent extraction vessels where heat generation from chemical reactions, heat loss through vessel walls, gas generation rates, distribution of Pu(IV) and HNO3 between aqueous and organic phases, volume changes, and solution flow into and out of vessels must be considered in the stability calculations. The
* Corresponding author. E-mail address: [email protected] (G.S. Barney). http://dx.doi.org/10.1016/j.jlp.2015.03.004 0950-4230/© 2015 Elsevier Ltd. All rights reserved.
organic phase extractant considered in this paper is a 30% by volume tributyl phosphate (TBP) e branched paraffin hydrocarbon (~C12) solution. Examples of application of the model to three types of process vessels are described, along with normal operations and effects of off-normal conditions.
2. Heat and gas generation Chemical reactions at different stages of the process will produce heat and generate gases. These reactions will occur mainly in vessels where reagents are added to a solution in order to change oxidation states of plutonium or to remove reducing agents (mainly HAN or hydrazine) from the system. These vessels include the plutonium stripping column [where Pu(IV) is reduced to Pu(III) by HAN and removed from the organic phase], the oxidation column (where excess HAN and hydrazine are oxidized) and tanks downstream from the stripping column. The overall HAN reactions involved in heat and gas production are reduction of Pu(IV) to Pu(III),
G.S. Barney, P.B. Duval / Journal of Loss Prevention in the Process Industries 35 (2015) 12e18
2Pu ðIVÞ þ 2NH3 OHþ /2Pu ðIIIÞ þ N2 þ 4H þ þ 2H2 O;
(1)
the autocatalytic reaction,
NH3 OHþ þ 0:5ðx þ 1ÞHNO3 /xHNO2 þ 0:25ð3 xÞN2 O þ H þ þ ð1:75 0:25xÞH2 O; (2) and the nitrous acid scavenging reaction,
NH3 OHþ þ HNO2 /N2 O þ H þ þ 2H2 O:
(3)
Other important heat producing reactions are reoxidation of Pu(III) with nitric acid,
2Pu ðIIIÞ þ 3H þ þ NO 3 /2Pu ðIVÞ þ HNO2 þ H2 O; oxidation of
N2Hþ 5
(4)
and HN3 with nitrous acid,
N2 H5þ þ HNO2 /HN3 þ Hþ þ 2H2 O
(5)
HN3 þ HNO2 /H2 þ N2 O þ H2 O;
(6)
and absorption of NO2 gas in aqueous solutions, þ 2NO2 ðgasÞ þ H2 OðlÞ /HNO2 ðaqÞ þ NO 3 ðlÞ þ HðaqÞ
Kd and Beta terms are the dissociation constant for the PuNO3þ 3 complex and the formation constant for the PuNO2þ complex, 3 respectively. This equation is used in the model to calculate solution temperature as the reactions proceed. The heat capacities of both aqueous and organic phases in the vessels are accounted for in the calculation since they will sometimes both be present and in thermal contact. The heat capacity of aqueous nitric acid solution and organic (30% TBP-hydrogenated propylene tetramer) phases are 0.946 kcal/K-L (for a 1.0 M HNO3 solution at 25 C from Weast, 1980 p. D-119) and 0.387 kcal/K-L [for a 30% (vol.) TBP-hydrocarbon diluent mixture at 25 C, Dean, 1979, pp. 7-206, 7-374, 9-77], respectively. Since the heat capacity of the stainless steel vessel is not accounted for in these calculations, the actual temperature increase is less than the calculated values. Gas generation in process vessels during normal operations occurs by reactions that produce gases such as Reactions (1)e(3) and (6) as HAN solutions or NOx gas (a mixture of NO2, N2O4, and NO gases) are added to the vessel. These reactions will normally occur quite rapidly in the vessels so that gas generation rates will depend only on the feed rate of the aqueous solution or NOx gas to the vessel and the concentration of reactants in the feed and in the vessel. However, the gas generation rate can be very high when the autocatalytic reaction occurs, depending on solution acidity, flow into and out of the vessel, and temperature conditions. This reaction is very fast once started and is strongly exothermic.
(7)
The heats of reaction for each of these reactions are calculated from published standard heats of formation (DH0f ) for the reactants and products (Dean, 1979). These reaction heats are used to calculate a temperature rise for the aqueous phase in processing vessels. The reactions occurring in a solution containing HNO3, NH3OHþ, HNO2, NO2, N2Hþ 5 , HN3, Pu(IV), and Pu(III) are considered. The heat produced by these reactions is 50.78 kcal/mole of Pu(IV) (Reaction (1)), 23.76 kcal/mole of NH3OHþ (Reaction (2)), 57.71 kcal/mole of NH3OHþ (Reaction (3)), 10.23 kcal/mole of Pu(III) (Reaction (4)), 44.1 kcal/mole of N2Hþ 5 (Reaction (5)), 82.49 kcal/ mole of HN3 (Reaction (6)) and 12.8 kcal/mole of NO2 gas absorbed (Reaction (7)). The rate of each of these reactions is calculated by the model so that the contribution of each reaction to the total temperature rise of the solution is:
3. Heat loss from process vessels Some of the heat generated by chemical reactions occurring in the process solutions will be lost to the air in the process cell/glove box atmosphere by three heat transport mechanisms: (1) conduction of heat through the solutions and the walls of the stainless steel vessel, (2) natural convection with heat being transferred from the outer surface of the vessel wall to a flowing layer of air in the cell/glove box, and (3) radiative heat transfer from the outside surface of the vessel to the cell/glove box. The assumption of instantaneous mixing of the entire contents of the vessel is used in the model and results in a uniform temperature throughout the vessel solutions. Heat conduction in the solutions is, therefore, not considered in the heat loss calculation. Heat transfer from the inner
0 1 0 1 2 kcal C kcal C h i þ 2 B B 50:78 23:76 ½PuðIVÞ OH NH dT 3 2B þ moleC moleC ¼k1 h i2 B @ C kcal A þ k2 NH3 OH ½NO2 @ C kcal A dt 2 þ 4 + p p L C L+ C Kd þ NO3 ½PuðIIIÞ H 1 0 0:5 h i0:5 kcal C h i h i ½PuðIIIÞ½HNO2 0:5 Hþ NO3 B 57:71 0 þ moleC i h þ k3 NH3 OHþ ½HNO2 B @ C kcal A þ k4 þ k4 H p L+ C 1 þ Beta NO3 0
1
0
13
1
0
1
(1)
kcal C h i h iB44:18 kcal C h iB82:49 kcal C B10:23 mole þ þ B þ B moleC moleC C B @ C kcal A þ k5 N2 H5 ½HNO2 H @ C kcal A þ k6 ½HN3 ½HNO2 H @ C kcal A p L C p L C p L C
kcal moles NO2 gas absorbed=secondÞ 12:8 mole þ Cp kcal=L C ðsolution volume; LÞ
The Cp term in this equation is defined as the total heat capacity (in kcal/K) of the aqueous and organic solutions in the vessel. The
surface of the stainless steel column to the outer surface will be rapid due to its high heat conductivity and relatively small
14
G.S. Barney, P.B. Duval / Journal of Loss Prevention in the Process Industries 35 (2015) 12e18
thickness and is ignored in the calculation. Only natural convection and thermal radiation are considered in the heat transfer calculation. The rate of heat loss from the vessel due to natural convection depends on the temperature difference between the outer surface of the vessel and the cell/glove box air. The cell/glove box air temperature is conservatively (higher than expected and thus retards heat flow) set to be 40 C except for the oxidation column which was set to 25 C. Another conservative assumption is that there is no forced convection cooling. In fact, cell/glove box ventilation will provide additional cooling capacity not considered in the calculation. The heat loss rate also depends on the surface area of the vessel. The surface area is conservatively calculated from vertical cylindrical portions of the vessels with dimensions assumed for these calculations. This surface area does not include any horizontal surfaces and is thus a conservative estimate. These estimated surface areas are used in all of the heat loss calculations. McAdams (1964) has defined the heat transfer rate from a solid surface to a fluid by Newton's equation
q ¼ hm AðTs TÞ
(2)
where q is the heat transfer rate in Watts, hm is the mean coefficient of heat transfer from surface to fluid in units of W/m2K, excluding any radiation, Ts is the temperature of the surface, T is the bulk temperature of the fluid and A is the surface area of the solid surface. The mean coefficient of heat transfer for vertical cylinders in air at ordinary temperatures has been estimated by McAdams to be
hm ¼ 1:31ðTs TÞ1=3
(3)
so the overall equation for heat transfer by natural convection for the vessels is
q ¼ 1:31ðTs TÞ4=3 A:
(4)
The net heat loss by thermal radiation from a convex surface (a gray surface having an emissivity less than a black body) to a larger enclosure can be estimated by the equation given by Perry and Green (1997).
q ¼ Aεs T14 T24
(5)
where ε is the emissivity of the stainless steel vessel (0.3 for stainless steel having a coarse surface (Weast, 1980), s is the StefaneBoltzmann constant (5.67 108 W/m2K4), T1 is the temperature of the vessel surface, and T2 is the temperature of the black surroundings (cell/glove box temperature). The net rate of temperature change (in K/s) for the solutions in process vessels due to heat loss to the process cell/glove box surroundings by natural convection and by thermal radiation can be estimated by the equation
4=3 dT A 4 4 ¼ 1:31 Tsurf Tcell þ εs Tsurf Tcell dt Cp
(6)
where A ¼ surface area in m2, Cp ¼ the total heat capacity of aqueous and organic solutions contained in the vessel in W-s/K ¼ (kcal/K) (4184 W-s/kcal), ε ¼ 0.3, s ¼ 5.67 108 W/m, Tsurf is the surface temperature of the vessel in K, and Tcell is the temperature of the process cell/glove box air in K.
An estimate of the magnitude of the heat loss from a vessel, for example, a vessel having a total volume of 40 L (20 L aqueous and 20 L organic phases) is readily calculated by the above equations by adding the contribution by natural convection and by thermal radiation with a cell/glove box air temperature (Tcell) of 40 C. For surface temperatures (Tsurf) of 45 C and 50 C and a surface area of 1.0 m2 the heat loss by natural convection is 11 W and 28 W, respectively and for thermal radiation, is 11 W and 22 W, respectively. 4. Distribution of Pu(IV) and HNO3 between aqueous and organic phases The model normally assumes instantaneous partitioning of Pu(IV) and HNO3 between the organic solvent and the aqueous phase. This assumption is conservative and reasonable for cases where the aqueous and organic flows are near normal flow rates. When the aqueous flow is stopped (for example in a Pu stripping column) or the rate lowered significantly, and the organic flow rate remains near the nominal rate, a significant amount of Pu(IV) and HNO3 will be removed from the column in the outgoing organic phase. This removal will not allow high concentrations of Pu(IV) and HNO3 to build up in the aqueous phase as the flow continues. To calculate the concentrations of Pu(IV) and HNO3 in the aqueous phase, the organic and aqueous distribution coefficients for these components must be calculated. Equilibrium distribution coefficients are used in the model to calculate these distributions. The distribution coefficients, DPu ([Pu(NO3)4$TBP]o/[Pu(IV)]a), DH1 ([HNO3$TBP]o/[Hþ]a) and DH2 ([HNO3$2TBP]o/[Hþ]a) (where o and a subscripts refer to the organic and aqueous phases, respectively) depend on the composition and temperature of the aqueous phase and on the volume fraction of TBP in the organic solvent. The equations used for calculating these coefficients were obtained from the SEPHIS computer program (Gronier et al., 1971) for solvent extraction calculations. In this program, the pseudo-mass-action equilibrium expressions for the extraction reactions are correlated using a function of total nitrate salting strength (na) where
na ¼ mHþ þ 2mUO2þ þ 4mPuðIVÞ þ mHAN 2
(7)
and mHþ , mUO2þ , mPuðIVÞ , and mHAN are molal concentrations of acid, 2 uranium, plutonium(IV), and HAN in the aqueous phase. The empirical equilibrium expressions are
KU* ¼
UO2 ðNO3 Þ2 ,2TBP h i UO2þ ½TBP2o 2
o
7:3 ¼ 3:7n1:75 þ 1:4n3:9 a a þ 0:011na
a
* KPu
PuðNO3 Þ4 ,2TBP ¼ Pu4þ a ½TBP2o
o
¼ KU* 0:2 þ 0:55F 1:25 þ 0:007n7:23 a
½HNO ,TBPo * ¼ 3 ¼ 0:135n0:82 þ 0:005n3:44 KH1 a a Hþ a ½TBP2o ½HNO ,2TBPo * * KH2 ¼ 3 zKH1 Hþ a ½TBP2o where F ¼ volume fraction of TBP in dry, solute-free solvent. Correction factors in the SEPHIS code allow use of TBP concentrations other than 100% TBP. These equations are
* 4F 0:17 3 e200t KPu þ KPu
G.S. Barney, P.B. Duval / Journal of Loss Prevention in the Process Industries 35 (2015) 12e18
and
* 1 0:54e15F e340t KH1 zKH2 ¼ KH1
where t ¼ 1/t0 -1/298 and t0 ¼ temperature, K. Assuming no uranium is present, free (uncomplexed) TBP concentration (in molar units) is given by
n ½TBP0 ¼
15
Examples of equations used in the HAN safety model for estimating the concentrations, volumes, and temperature changes due to solution flow are derived in the paragraphs below. The chemical components that can be extracted to the organic solvent (a 30% TBP-branched paraffin hydrocarbon solution) were considered in these equations [Pu(IV), Hþ and NO 3 ]. Equations for total Pu in the aqueous phase [Pu(IV) plus Pu(III)] and temperature are also
i1=2 o h ðKH1 mHþ þ 1Þ þ ðKH1 mHþ þ 1Þ2 þ 8 KPu mPuðIVÞ þ KH2 mHþ ð3:651FÞ 4 KPu mPuðIVÞ þ KH2 mHþ
The distribution coefficients are
DPu ¼ KPu ½TBP20 DH1 ¼ KH1 ½TBP0 and
DH2 ¼ KH2 ½TBP20 These equations were used in the model to calculate distribution coefficients for Pu(IV) and HNO3 so that the concentrations in the aqueous and organic phases in the vessels could be calculated.
included. Only examples of the equations for calculating changes in the Pu(IV) concentrations in the aqueous phase are given here. The terms used in these equations are defined as follows: t ¼ time (s), T ¼ temperature ( C) V ¼ volume (L), F ¼ flow rate (L/s), In and Out ¼ flows into or out of the vessel Tk ¼ tank, Ex ¼ extractor org ¼ organic phase, aq ¼ aqueous phase Dx ¼ equilibrium distribution coefficient for x between organic and aqueous phases Cp ¼ heat capacity of solutions (kcal/L- C) Rz ¼ rate of Reaction z (moles/L-s) DHz ¼ heat of reaction for Reaction z (kcal/mole) For each time step Dt,
5. Solution flow into and out of vessels The rate of change for concentrations of chemical species, solution volumes, and temperatures in process solutions over time will depend on solution flow to and from the process vessels. This is in addition to the change in concentrations and temperatures due to chemical reactions in the solutions. The calculation becomes more complex when two liquid phases are present (aqueous and organic phases in some of the cases considered here) and some of the chemical species will be distributed between these phases. The flow can be a simple addition of solutions to a tank or counter-
Daq ½PuðIVÞTk Faq ¼ Dt
aq ½PuðIVÞIn
aq
5.1. Aqueous þ organic flow into a tank (with no outflow) The changes in concentrations of Pu(IV) over time in the aqueous phase of a tank receiving flows of aqueous and organic solution are given by the equation,
þ Forg org ½PuðIVÞIn DPu aq ½PuðIVÞTk 1 aq org V2 þ DPu V2
current flows of aqueous and organic solutions through an extraction vessel (pulsed column or mixer-settlers). These flow situations require different equations for calculating the effects of flow on the concentrations of important solution components and temperatures.
Daq ½PuðIVÞEx Faq ¼ Dt
Dt ¼ t2 e t1 DV ¼ V2 V1 ¼ DtF.
aq ½PuðIVÞTk 1
aq
V ðR20 R19 Þ þ 2aq org V2 þ DPu V2
5.2. Aqueous þ organic flow into an extraction vessel The changes in concentrations of Pu(IV) over time in the aqueous phase of an extraction vessel receiving flows of aqueous and organic solution are given by the equation,
½PuðIVÞIn aq ½PuðIVÞOut þ Forg org ½PuðIVÞIn DPu ðV aq þ DPu V org Þ
aq ½PuðIVÞOut
þ
V aq ðR20 R19 Þ ðV aq þ DPu V org Þ
16
G.S. Barney, P.B. Duval / Journal of Loss Prevention in the Process Industries 35 (2015) 12e18
It can be seen that the equations derived for calculating the rate of change of concentrations of Pu(IV), PuTotal, Hþ, and NO 3 and temperature have the same form when used for adding aqueous and organic solutions to a tank or to an extractor. The only difference is that the concentration or temperature in the tank or extractor calculated for the previous time step is labeled [X]Tk or [X]Ex or [X]Out. They are all calculated in the same manner, so that it is not necessary to modify the computer code when applying the model to an extractor or a tank. 6. Examples of model results from application to process scenarios Results from model calculations for three different process scenarios are presented in this section. A number of assumptions are required in these calculations because of a lack of data and a need for simplification of the computations. The assumptions made are conservative and lead to results that give higher temperatures and greater quantities of gases generated than would be the case for actual conditions in the process vessels. These assumptions are listed below along with justifications for making them: 1) Extrapolation of rate constants to a temperature higher or lower than the temperature at which the rate constants are measured using the Arrhenius equation requires making the assumption that the reaction mechanism remains constant over the relevant temperature range. This is justified since the temperature range is generally small (a few tens of degrees centigrade) and changes in reaction mechanism are rare for reactions in aqueous solutions over this temperature range. Many of the rate constants used in the model have been shown to follow the Arrhenius law over the entire required temperature range. 2) The heat capacity of the stainless steel vessels is ignored in the calculations due to the complexity of the vessel configuration and heat flow. This results in calculated temperature increases that are higher than actual values, which is a conservative assumption for chemical reaction rates. Heat conduction through the walls of the vessel is assumed to have no temperature gradient and no forced air convection cooling is assumed outside the vessel walls. 3) Instantaneous partitioning of Pu(IV) and HNO3 from the organic phase to the aqueous phase in the extraction column and tank is assumed in order to simplify the calculations. Partitioning of these components in PUREX-type solvent extraction systems is known to be rapid and equilibrium is assumed in most models of these systems. This assumption is conservative since it allows more time for reaction in the aqueous phase. 4) Instantaneous mixing of the entire volume of aqueous solution in the vessels with input solution is assumed (uniform composition throughout the solution volume). This assumption does not account for the multistage countercurrent operation of the pulsed columns (stripping column), which leads to more complete reactions of the solution components. In all cases run with the model, however, the Pu(IV) reduction reactions with HAN are 80e95 percent complete after steady-state operation is reached. 5) For process vessels that contain both aqueous and organic phases, it is assumed that both phases are at the same temperature because of good mixing of the phases, 6) Constant volumes for the aqueous and organic phases are assumed in the pulsed column (stripping column) calculations and input/output solutions to all the vessels are assumed to be additive (no volume changes due to mixing). The model output for major reaction product gases (N2 and N2O)
is given in terms of aqueous solution molarity for ease in calculating the rate of gas generation. Obviously, their solubilities will be exceeded at some point and they will evolve from solution as gases. The solubilities will depend on a number of factors (solution composition, pressure, temperature, etc.). Since these relationships are unknown, the gas generation rate calculation conservatively assumes that the gases are initially saturated in the aqueous solution and any additional N2 or N2O formed will be evolved as gases. The model requires input of initial conditions in the process vessel of interest and input for any flows that are entering the vessel. For the three process scenarios considered here (a plutonium stripping column, an oxidation column, and a plutonium rework tank) values for the initial parameters and the input flow parameters are given in Table 1. 6.1. Stripping column The stripping column is used to separate plutonium from the organic extraction phase (a TBP-branched paraffin hydrocarbon mixture) by reducing the TBP-extractable Pu(IV) nitrate to lessextractable Pu(III) nitrate with aqueous HAN. The purified Pu(III) nitrate moves to the aqueous phase, leaving uranium and other impurities behind in the organic phase. The column operates in a countercurrent mode with the lighter organic phase loaded with Pu(IV) nitrate fed near the bottom of the column and the HANHNO3 stripping solution fed near the top of the column. The column is pulsed so that small droplets of the organic solvent can rise through the continuous aqueous phase and the Pu(IV) can be rapidly stripped and reduced to Pu(III). For the stripping column run simulated by the model, the concentration and temperature curves given in Fig. 1 show that calculated steady state concentrations of Pu(IV), Pu(III), and HAN in the aqueous phase are reached before about 300 min and are about 0.018 M, 0.18 M, and 0.32 M, respectively. This is for the normal operation of the column with feed concentrations, feed rates, and temperatures of the feeds given in the table. The temperature of the solutions in the column are somewhat elevated (about 0.1 C) compared to the feed, due mainly to the exothermic reduction of Pu(IV) with HAN. The effects of the off-normal condition of stopping the aqueous flow and continuing the organic flow are also shown in Fig. 1. When the aqueous flow is stopped and the organic flow continues, the acid concentration increases because of the Pu(IV) reduction with HAN and extraction of acid from the organic phase. The Pu(III) concentration rises as the HAN continues to reduce more Pu(IV) from the organic phase until the HAN is spent (along with the hydrazine and hydrazoic acid) and will be rapidly oxidized by the nitrous acid generated by the autocatalytic reaction. The concentration of Pu(IV) in the aqueous phase never reaches that of the Pu(III) maximum since the organic phase will continue to extract it. Stopping the aqueous flow results in a rapid drop in temperature because the main heat-producing HAN-Pu(IV) reaction slows as the HAN is used up. The autocatalytic reaction that occurs about 760 min after stopping the aqueous flow generates enough heat to raise the temperature of the column by about 0.12 C. The maximum gas generation rate during this reaction was 14 L/min of N2O. 6.2. Oxidation column The purpose of the oxidation column is to remove reducing agents (HAN, hydrazine, and hydrazoic acid) from the aqueous solution coming from the stripping column by oxidation so that they will not cause safety problems downstream. The solution in the column is contacted with a mixture of NOx gases (mainly NO2) and air that is fed to the bottom of the column as the solution flows
G.S. Barney, P.B. Duval / Journal of Loss Prevention in the Process Industries 35 (2015) 12e18
17
Table 1 Compositions of solutions in the vessels initially and solutions flowing into the vessels, volumes, flow rates, temperatures, cell temperatures, and wetted surface area of the vessels. Vessel Variable
Stripping column
[HAN]aq (M) [NO2]aq (M) NO2 gas flow rate (mole/s) [HNO2]aq (M) [HNO2]org (M) [NO]aq (M) [HNO]aq (M) [Hþ]aq (M) (M) [Hþ]org (M) [NO 3 ]aq (M) [NO 3 ]org (M) [N2Hþ 5 ]aq (M) [HN3]aq (M) [N2O]aq (M) [N2]aq (M) [Pu(III)]aq (M) [Pu(IV)]aq (M) [Pu(IV)]org (M) [Fe(III)]aq (M) [FeTotal]aq (M) Vaq (L) Vorg (L) Taq ( C) Torg ( C) Faq (L/s) Forg (L/s) Cell temperature ( C) Wetted surface area (m2)
0.5 4E-12 N/A 4E-10 0 2E-13 0 0.5 0.126 1.1 1.21 0.1 0 1E-5 1E-5 1E-6 1E-5 3.62E-5 1E-6 1.0001E-6 15.6 20.4 40 40 N/A N/A 40 2.40
Oxidation column
Pu rework tank
Fig. 1. Aqueous solution concentrations of chemical species and temperatures for normal and off-normal (aqueous solution flow stopped) process scenarios for the stripping column over time.
down from the top. The column is filled with Lessing rings to allow efficient contact between the gas and liquid phases. The oxidized solution is collected in a 10-L pot below the column. During normal operation, the NO2 feed rate to the column is about 0.01 moles/s, as shown in the table. This oxidation by NO2 results in very low concentrations of HAN and hydrazine in the pot, as shown in Fig. 2. Pu(III) is also oxidized to Pu(IV), nitrous acid concentrations reach a high level, and the temperature increases (about 3.8 C) due to the oxidation reactions and absorption of NO2 by the solution. The off-normal condition of stopping the NOx flow after about 123 min is shown in Fig. 2. The temperature rapidly drops and the concentrations of HAN, hydrazine, HNO2, Pu(III), and Pu(IV) approach those of the aqueous feed solution after about 300 min. The HAN-HNO3 autocatalytic reaction conditions are not
Oxidation column
Pu rework tank
0.25 4E-12 0.01 4E-12 N/A 2E-13 0 0.75 N/A 1.6 N/A 0.06 0.03 1E-5 1E-5 0.185 0.00239 N/A 1E-6 1.0001E-6 N/A N/A 25 N/A 0.005 N/A 25 2.0
0.4 1E-10 N/A 1E-9 0.0001 5E-12 9E-7 0.7 0.2 1.43 0.3 0 0 1E-5 1E-5 0.15 0.00384 0.03 1E-6 1.0001E-6 N/A N/A 40 40 0.003 0.004 40 10.0
Input flow Variables
Initial conditions in vessel 0 4E-12 0.01 0.001 N/A 2E-13 0 0.78 N/A 1.65 N/A 0.06 0 1E-5 1E-5 0.0001 0.185 N/A 1E-6 1.0001E-6 10.0 N/A 25 N/A N/A N/A 25 2.0
Stripping column
0 6E-9 N/A 1E-5 0 1E-10 0 0.1 0.01 0.1 0.000552 0 0 1E-5 1E-5 1E-7 1E-6 1E-9 1E-6 1.0001E-6 30 30 40 40 N/A N/A 40 10.0
0.5 4E-12 N/A 4E-12 0.001 2E-13 0 0.14 0.22 0.71 0.80 0.1 0 1E-5 1E-5 1E-5 0 0.145 1E-6 1.0001E-6 N/A N/A 40 40 0.0055 0.073 40 2.40
Fig. 2. Aqueous solution concentrations of chemical species and temperatures for normal and off-normal (NOx flow stopped) process scenarios for the oxidation column over time.
reached during this process scenario. 6.3. Plutonium rework tank The plutonium rework tank is a 300 L slab tank that receives solutions from the stripping column and can recycle the organic phase to the stripping column. The aqueous solution from the stripping column, which normally contains HAN, hydrazine, and hydrazoic acid, is sent to a second tank where these reducing agents are oxidized with NOx gas. In this computer simulation, aqueous and organic solutions from the stripping column are fed to the tank that initially contains 30 L each of aqueous and organic solutions. The compositions, temperatures and feed rates of the initial and input solutions are given in Table 1. Fig. 3 shows a
+
G.S. Barney, P.B. Duval / Journal of Loss Prevention in the Process Industries 35 (2015) 12e18
10
1
Flows To Tank Stopped
40.020
-1
10
40.015
-3
10
[NH OH ] [Pu(IV)] [Pu(III)] [Pu(IV)] [HNO ] Temperature
-5
10
-7
10
40.010
Temperature, °C
[NH3OH ], [Pu(IV)], [PuIII)], [HNO2], M
18
40.005
-9
10
10
2
3
4 5 6
100
2
3
4 5 6
1000
2
3
4 5 6
Time, min Fig. 3. Aqueous and organic solution concentrations of chemical species and temperatures for normal and off-normal (aqueous and organic flows stopped) process scenarios for the plutonium rework tank over time.
Fig. 4. Gas generation rates for N2O and N2 gases and solution temperatures during the autocatalytic reaction in the plutonium rework tank.
gradual increase in the concentrations of HAN, Pu(III), and Pu(IV) in the aqueous phase as the feed solutions are added to the tank. This is expected since the original concentration of each of these components in the tank is very low or zero. The gradual decrease in the HNO2 concentration is due to dilution of the initially higher concentration in the tank (1E-5 M) with the lower feed concentration (1E-9). The small temperature increase is mainly the result of HAN reduction of Pu(IV) from the organic phase of the feed solution. After about 400 min the volumes of aqueous and organic phases in the tank are about 102 L and 126 L, respectively, and the flows to the tank are stopped. Concentration of Pu(III) increases slightly and the concentration of Pu(IV) decreases initially in the aqueous phase as the reduction of Pu(IV) by HAN continues. HAN concentration decreases slowly at first and then more rapidly as the autocatalytic
reaction is approached at about 6300 min (4.38 days). Since HNO2 is a product of the autocatalytic reaction (Reaction (2)), its concentration slowly increases initially until the rapid autocatalytic reaction is approached. The temperature of the solutions rapidly increase by about 0.022 C due to the reaction. Pressurization of the tank is a safety concern when the autocatalytic reaction occurs because the reaction produces N2O gas as a major product. The rate of gas generation can be calculated using the model and the results for this tank are given in Fig. 4. The figure shows that the maximum gas generation rate for N2O is about 10 L/ min for the scenario used in the model run. The maximum pressure in the tank resulting from this gas generation rate is not calculated here, but will depend on the gaseous headspace in the tank and on the vent size and configuration and on the temperature. 7. Conclusions A mathematical model for predicting the stability of HAN in plutonium-containing process solutions is applied to three typical types of process vessels in solvent extraction plutonium processes and for normal and off-normal process scenarios. The model allows calculation of the effects of important process parameters such as temperature, volume, flow rates, gas generation rates, and concentrations of chemical components over time. Two of the offnormal scenarios led to the rapid autocatalytic reaction of HAN with nitric acid where heat and gas are generated and that could lead to damage of the process equipment and/or release of hazardous plutonium solution from the vessel. Both of these autocatalytic reactions resulted from the aqueous phase of the process solutions being allowed to remain stationary in the vessel while the Pu(III) in these solutions slowly generated nitrous acid by Reaction (4) and which used up the HAN by Reactions (3) and (1). Eventually, at some combination of temperature, concentration of nitric acid, nitrous acid and HAN, the solution becomes unstable and the autocatalytic reaction occurs. It can be concluded that all solutions containing nitric acid, plutonium and HAN are inherently unstable and will eventually lead to the autocatalytic reaction occurring. References Barney, G.S., Duval, P.B., 2011. Model for predicting hydroxylamine stability in plutonium process solutions. J. Loss Prev. Process Ind. 24, 76e84. Dean, J.A., 1979. Lange's Handbook of Chemistry, 12th ed. McGraw-Hill Book Company, Inc., New York. Gronier, W.S., Rainey, R.H., Watson, S.B., 1971. An analysis of the transient and steady-state operation of countercurrent liquid-liquid solvent extraction process. Ind. Eng. Chem. Process Des. Dev. 15, 385e390. McAdams, W.H., 1964. Heat Transmission, third ed. McGraw-Hill Book Company, Inc., New York. Perry, R.H., Green, D., 1997. Perry's Chemical Engineering Handbook, seventh ed. McGraw-Hill Book Company, Inc., New York. Weast, R.C. (Ed.), 1980. CRC Handbook of Chemistry and Physics, 61st ed. CRC Press, Inc., Boca Raton, Florida.
Contents lists available at ScienceDirect
Journal of Loss Prevention in the Process Industries journal homepage: www.elsevier.com/locate/jlp
Application of a hydroxylamine nitrate stability model to plutonium purification process equipment G. Scott Barney a, *, Paul B. Duval b a b
105903 N. Harrington Rd., West Richland, WA, USA Shaw AREVA MOX Services, LLC, Aiken, SC, USA
a r t i c l e i n f o
a b s t r a c t
Article history: Received 17 February 2014 Received in revised form 2 February 2015 Accepted 5 March 2015 Available online 6 March 2015
A mathematical model that predicts hydroxylamine nitrate (HAN) (NH2OH$HNO3) stability is applied to aqueous solutions containing HAN, nitric acid and plutonium that are used in plutonium purification processes. The model estimates the stability of these solutions with respect to the rapid, hazardous, autocatalytic reaction of HAN with nitric acid that generates heat and gas. It also accounts for reaction kinetics, temperature changes, gas generation rates, solution volumes and flow rates, and distribution of plutonium and nitric acid between aqueous and organic phases. The model is applied to three typical process vessels used in solvent extraction purification of plutonium e a countercurrent aqueous/organic plutonium stripping column, an oxidation column used for HAN and hydrazine destruction, and a plutonium rework tank. Both normal and off-normal process scenarios are modeled. Two of the offnormal scenarios lead to the rapid autocatalytic reaction of HAN with nitric acid where heat and gas are generated and that could lead to damage of the process equipment and/or release of hazardous plutonium solution from the vessel. In these two cases, stationary aqueous solutions containing HAN, Pu(III), and nitric acid were allowed to slowly react until conditions for the autocatalytic reaction were reached. © 2015 Elsevier Ltd. All rights reserved.
Keywords: Hydroxylamine stability Plutonium Kinetics Nitric acid Model Solvent extraction Process equipment
1. Introduction A mathematical model for determining the stability of aqueous hydroxylamine nitrate (HAN) (NH2OH$HNO3) solutions used in solvent extraction purification of plutonium was described in an earlier publication (Barney and Duval, 2011). The model calculates rates of the reactions involved in oxidation of HAN under conditions existing in a typical plutonium purification solvent extraction process. Conditions for which the hazardous autocatalytic reaction of HAN with nitric acid occurs were predicted successfully with the model. This paper describes application of the model for predicting HAN stability in solvent extraction vessels where heat generation from chemical reactions, heat loss through vessel walls, gas generation rates, distribution of Pu(IV) and HNO3 between aqueous and organic phases, volume changes, and solution flow into and out of vessels must be considered in the stability calculations. The
* Corresponding author. E-mail address: [email protected] (G.S. Barney). http://dx.doi.org/10.1016/j.jlp.2015.03.004 0950-4230/© 2015 Elsevier Ltd. All rights reserved.
organic phase extractant considered in this paper is a 30% by volume tributyl phosphate (TBP) e branched paraffin hydrocarbon (~C12) solution. Examples of application of the model to three types of process vessels are described, along with normal operations and effects of off-normal conditions.
2. Heat and gas generation Chemical reactions at different stages of the process will produce heat and generate gases. These reactions will occur mainly in vessels where reagents are added to a solution in order to change oxidation states of plutonium or to remove reducing agents (mainly HAN or hydrazine) from the system. These vessels include the plutonium stripping column [where Pu(IV) is reduced to Pu(III) by HAN and removed from the organic phase], the oxidation column (where excess HAN and hydrazine are oxidized) and tanks downstream from the stripping column. The overall HAN reactions involved in heat and gas production are reduction of Pu(IV) to Pu(III),
G.S. Barney, P.B. Duval / Journal of Loss Prevention in the Process Industries 35 (2015) 12e18
2Pu ðIVÞ þ 2NH3 OHþ /2Pu ðIIIÞ þ N2 þ 4H þ þ 2H2 O;
(1)
the autocatalytic reaction,
NH3 OHþ þ 0:5ðx þ 1ÞHNO3 /xHNO2 þ 0:25ð3 xÞN2 O þ H þ þ ð1:75 0:25xÞH2 O; (2) and the nitrous acid scavenging reaction,
NH3 OHþ þ HNO2 /N2 O þ H þ þ 2H2 O:
(3)
Other important heat producing reactions are reoxidation of Pu(III) with nitric acid,
2Pu ðIIIÞ þ 3H þ þ NO 3 /2Pu ðIVÞ þ HNO2 þ H2 O; oxidation of
N2Hþ 5
(4)
and HN3 with nitrous acid,
N2 H5þ þ HNO2 /HN3 þ Hþ þ 2H2 O
(5)
HN3 þ HNO2 /H2 þ N2 O þ H2 O;
(6)
and absorption of NO2 gas in aqueous solutions, þ 2NO2 ðgasÞ þ H2 OðlÞ /HNO2 ðaqÞ þ NO 3 ðlÞ þ HðaqÞ
Kd and Beta terms are the dissociation constant for the PuNO3þ 3 complex and the formation constant for the PuNO2þ complex, 3 respectively. This equation is used in the model to calculate solution temperature as the reactions proceed. The heat capacities of both aqueous and organic phases in the vessels are accounted for in the calculation since they will sometimes both be present and in thermal contact. The heat capacity of aqueous nitric acid solution and organic (30% TBP-hydrogenated propylene tetramer) phases are 0.946 kcal/K-L (for a 1.0 M HNO3 solution at 25 C from Weast, 1980 p. D-119) and 0.387 kcal/K-L [for a 30% (vol.) TBP-hydrocarbon diluent mixture at 25 C, Dean, 1979, pp. 7-206, 7-374, 9-77], respectively. Since the heat capacity of the stainless steel vessel is not accounted for in these calculations, the actual temperature increase is less than the calculated values. Gas generation in process vessels during normal operations occurs by reactions that produce gases such as Reactions (1)e(3) and (6) as HAN solutions or NOx gas (a mixture of NO2, N2O4, and NO gases) are added to the vessel. These reactions will normally occur quite rapidly in the vessels so that gas generation rates will depend only on the feed rate of the aqueous solution or NOx gas to the vessel and the concentration of reactants in the feed and in the vessel. However, the gas generation rate can be very high when the autocatalytic reaction occurs, depending on solution acidity, flow into and out of the vessel, and temperature conditions. This reaction is very fast once started and is strongly exothermic.
(7)
The heats of reaction for each of these reactions are calculated from published standard heats of formation (DH0f ) for the reactants and products (Dean, 1979). These reaction heats are used to calculate a temperature rise for the aqueous phase in processing vessels. The reactions occurring in a solution containing HNO3, NH3OHþ, HNO2, NO2, N2Hþ 5 , HN3, Pu(IV), and Pu(III) are considered. The heat produced by these reactions is 50.78 kcal/mole of Pu(IV) (Reaction (1)), 23.76 kcal/mole of NH3OHþ (Reaction (2)), 57.71 kcal/mole of NH3OHþ (Reaction (3)), 10.23 kcal/mole of Pu(III) (Reaction (4)), 44.1 kcal/mole of N2Hþ 5 (Reaction (5)), 82.49 kcal/ mole of HN3 (Reaction (6)) and 12.8 kcal/mole of NO2 gas absorbed (Reaction (7)). The rate of each of these reactions is calculated by the model so that the contribution of each reaction to the total temperature rise of the solution is:
3. Heat loss from process vessels Some of the heat generated by chemical reactions occurring in the process solutions will be lost to the air in the process cell/glove box atmosphere by three heat transport mechanisms: (1) conduction of heat through the solutions and the walls of the stainless steel vessel, (2) natural convection with heat being transferred from the outer surface of the vessel wall to a flowing layer of air in the cell/glove box, and (3) radiative heat transfer from the outside surface of the vessel to the cell/glove box. The assumption of instantaneous mixing of the entire contents of the vessel is used in the model and results in a uniform temperature throughout the vessel solutions. Heat conduction in the solutions is, therefore, not considered in the heat loss calculation. Heat transfer from the inner
0 1 0 1 2 kcal C kcal C h i þ 2 B B 50:78 23:76 ½PuðIVÞ OH NH dT 3 2B þ moleC moleC ¼k1 h i2 B @ C kcal A þ k2 NH3 OH ½NO2 @ C kcal A dt 2 þ 4 + p p L C L+ C Kd þ NO3 ½PuðIIIÞ H 1 0 0:5 h i0:5 kcal C h i h i ½PuðIIIÞ½HNO2 0:5 Hþ NO3 B 57:71 0 þ moleC i h þ k3 NH3 OHþ ½HNO2 B @ C kcal A þ k4 þ k4 H p L+ C 1 þ Beta NO3 0
1
0
13
1
0
1
(1)
kcal C h i h iB44:18 kcal C h iB82:49 kcal C B10:23 mole þ þ B þ B moleC moleC C B @ C kcal A þ k5 N2 H5 ½HNO2 H @ C kcal A þ k6 ½HN3 ½HNO2 H @ C kcal A p L C p L C p L C
kcal moles NO2 gas absorbed=secondÞ 12:8 mole þ Cp kcal=L C ðsolution volume; LÞ
The Cp term in this equation is defined as the total heat capacity (in kcal/K) of the aqueous and organic solutions in the vessel. The
surface of the stainless steel column to the outer surface will be rapid due to its high heat conductivity and relatively small
14
G.S. Barney, P.B. Duval / Journal of Loss Prevention in the Process Industries 35 (2015) 12e18
thickness and is ignored in the calculation. Only natural convection and thermal radiation are considered in the heat transfer calculation. The rate of heat loss from the vessel due to natural convection depends on the temperature difference between the outer surface of the vessel and the cell/glove box air. The cell/glove box air temperature is conservatively (higher than expected and thus retards heat flow) set to be 40 C except for the oxidation column which was set to 25 C. Another conservative assumption is that there is no forced convection cooling. In fact, cell/glove box ventilation will provide additional cooling capacity not considered in the calculation. The heat loss rate also depends on the surface area of the vessel. The surface area is conservatively calculated from vertical cylindrical portions of the vessels with dimensions assumed for these calculations. This surface area does not include any horizontal surfaces and is thus a conservative estimate. These estimated surface areas are used in all of the heat loss calculations. McAdams (1964) has defined the heat transfer rate from a solid surface to a fluid by Newton's equation
q ¼ hm AðTs TÞ
(2)
where q is the heat transfer rate in Watts, hm is the mean coefficient of heat transfer from surface to fluid in units of W/m2K, excluding any radiation, Ts is the temperature of the surface, T is the bulk temperature of the fluid and A is the surface area of the solid surface. The mean coefficient of heat transfer for vertical cylinders in air at ordinary temperatures has been estimated by McAdams to be
hm ¼ 1:31ðTs TÞ1=3
(3)
so the overall equation for heat transfer by natural convection for the vessels is
q ¼ 1:31ðTs TÞ4=3 A:
(4)
The net heat loss by thermal radiation from a convex surface (a gray surface having an emissivity less than a black body) to a larger enclosure can be estimated by the equation given by Perry and Green (1997).
q ¼ Aεs T14 T24
(5)
where ε is the emissivity of the stainless steel vessel (0.3 for stainless steel having a coarse surface (Weast, 1980), s is the StefaneBoltzmann constant (5.67 108 W/m2K4), T1 is the temperature of the vessel surface, and T2 is the temperature of the black surroundings (cell/glove box temperature). The net rate of temperature change (in K/s) for the solutions in process vessels due to heat loss to the process cell/glove box surroundings by natural convection and by thermal radiation can be estimated by the equation
4=3 dT A 4 4 ¼ 1:31 Tsurf Tcell þ εs Tsurf Tcell dt Cp
(6)
where A ¼ surface area in m2, Cp ¼ the total heat capacity of aqueous and organic solutions contained in the vessel in W-s/K ¼ (kcal/K) (4184 W-s/kcal), ε ¼ 0.3, s ¼ 5.67 108 W/m, Tsurf is the surface temperature of the vessel in K, and Tcell is the temperature of the process cell/glove box air in K.
An estimate of the magnitude of the heat loss from a vessel, for example, a vessel having a total volume of 40 L (20 L aqueous and 20 L organic phases) is readily calculated by the above equations by adding the contribution by natural convection and by thermal radiation with a cell/glove box air temperature (Tcell) of 40 C. For surface temperatures (Tsurf) of 45 C and 50 C and a surface area of 1.0 m2 the heat loss by natural convection is 11 W and 28 W, respectively and for thermal radiation, is 11 W and 22 W, respectively. 4. Distribution of Pu(IV) and HNO3 between aqueous and organic phases The model normally assumes instantaneous partitioning of Pu(IV) and HNO3 between the organic solvent and the aqueous phase. This assumption is conservative and reasonable for cases where the aqueous and organic flows are near normal flow rates. When the aqueous flow is stopped (for example in a Pu stripping column) or the rate lowered significantly, and the organic flow rate remains near the nominal rate, a significant amount of Pu(IV) and HNO3 will be removed from the column in the outgoing organic phase. This removal will not allow high concentrations of Pu(IV) and HNO3 to build up in the aqueous phase as the flow continues. To calculate the concentrations of Pu(IV) and HNO3 in the aqueous phase, the organic and aqueous distribution coefficients for these components must be calculated. Equilibrium distribution coefficients are used in the model to calculate these distributions. The distribution coefficients, DPu ([Pu(NO3)4$TBP]o/[Pu(IV)]a), DH1 ([HNO3$TBP]o/[Hþ]a) and DH2 ([HNO3$2TBP]o/[Hþ]a) (where o and a subscripts refer to the organic and aqueous phases, respectively) depend on the composition and temperature of the aqueous phase and on the volume fraction of TBP in the organic solvent. The equations used for calculating these coefficients were obtained from the SEPHIS computer program (Gronier et al., 1971) for solvent extraction calculations. In this program, the pseudo-mass-action equilibrium expressions for the extraction reactions are correlated using a function of total nitrate salting strength (na) where
na ¼ mHþ þ 2mUO2þ þ 4mPuðIVÞ þ mHAN 2
(7)
and mHþ , mUO2þ , mPuðIVÞ , and mHAN are molal concentrations of acid, 2 uranium, plutonium(IV), and HAN in the aqueous phase. The empirical equilibrium expressions are
KU* ¼
UO2 ðNO3 Þ2 ,2TBP h i UO2þ ½TBP2o 2
o
7:3 ¼ 3:7n1:75 þ 1:4n3:9 a a þ 0:011na
a
* KPu
PuðNO3 Þ4 ,2TBP ¼ Pu4þ a ½TBP2o
o
¼ KU* 0:2 þ 0:55F 1:25 þ 0:007n7:23 a
½HNO ,TBPo * ¼ 3 ¼ 0:135n0:82 þ 0:005n3:44 KH1 a a Hþ a ½TBP2o ½HNO ,2TBPo * * KH2 ¼ 3 zKH1 Hþ a ½TBP2o where F ¼ volume fraction of TBP in dry, solute-free solvent. Correction factors in the SEPHIS code allow use of TBP concentrations other than 100% TBP. These equations are
* 4F 0:17 3 e200t KPu þ KPu
G.S. Barney, P.B. Duval / Journal of Loss Prevention in the Process Industries 35 (2015) 12e18
and
* 1 0:54e15F e340t KH1 zKH2 ¼ KH1
where t ¼ 1/t0 -1/298 and t0 ¼ temperature, K. Assuming no uranium is present, free (uncomplexed) TBP concentration (in molar units) is given by
n ½TBP0 ¼
15
Examples of equations used in the HAN safety model for estimating the concentrations, volumes, and temperature changes due to solution flow are derived in the paragraphs below. The chemical components that can be extracted to the organic solvent (a 30% TBP-branched paraffin hydrocarbon solution) were considered in these equations [Pu(IV), Hþ and NO 3 ]. Equations for total Pu in the aqueous phase [Pu(IV) plus Pu(III)] and temperature are also
i1=2 o h ðKH1 mHþ þ 1Þ þ ðKH1 mHþ þ 1Þ2 þ 8 KPu mPuðIVÞ þ KH2 mHþ ð3:651FÞ 4 KPu mPuðIVÞ þ KH2 mHþ
The distribution coefficients are
DPu ¼ KPu ½TBP20 DH1 ¼ KH1 ½TBP0 and
DH2 ¼ KH2 ½TBP20 These equations were used in the model to calculate distribution coefficients for Pu(IV) and HNO3 so that the concentrations in the aqueous and organic phases in the vessels could be calculated.
included. Only examples of the equations for calculating changes in the Pu(IV) concentrations in the aqueous phase are given here. The terms used in these equations are defined as follows: t ¼ time (s), T ¼ temperature ( C) V ¼ volume (L), F ¼ flow rate (L/s), In and Out ¼ flows into or out of the vessel Tk ¼ tank, Ex ¼ extractor org ¼ organic phase, aq ¼ aqueous phase Dx ¼ equilibrium distribution coefficient for x between organic and aqueous phases Cp ¼ heat capacity of solutions (kcal/L- C) Rz ¼ rate of Reaction z (moles/L-s) DHz ¼ heat of reaction for Reaction z (kcal/mole) For each time step Dt,
5. Solution flow into and out of vessels The rate of change for concentrations of chemical species, solution volumes, and temperatures in process solutions over time will depend on solution flow to and from the process vessels. This is in addition to the change in concentrations and temperatures due to chemical reactions in the solutions. The calculation becomes more complex when two liquid phases are present (aqueous and organic phases in some of the cases considered here) and some of the chemical species will be distributed between these phases. The flow can be a simple addition of solutions to a tank or counter-
Daq ½PuðIVÞTk Faq ¼ Dt
aq ½PuðIVÞIn
aq
5.1. Aqueous þ organic flow into a tank (with no outflow) The changes in concentrations of Pu(IV) over time in the aqueous phase of a tank receiving flows of aqueous and organic solution are given by the equation,
þ Forg org ½PuðIVÞIn DPu aq ½PuðIVÞTk 1 aq org V2 þ DPu V2
current flows of aqueous and organic solutions through an extraction vessel (pulsed column or mixer-settlers). These flow situations require different equations for calculating the effects of flow on the concentrations of important solution components and temperatures.
Daq ½PuðIVÞEx Faq ¼ Dt
Dt ¼ t2 e t1 DV ¼ V2 V1 ¼ DtF.
aq ½PuðIVÞTk 1
aq
V ðR20 R19 Þ þ 2aq org V2 þ DPu V2
5.2. Aqueous þ organic flow into an extraction vessel The changes in concentrations of Pu(IV) over time in the aqueous phase of an extraction vessel receiving flows of aqueous and organic solution are given by the equation,
½PuðIVÞIn aq ½PuðIVÞOut þ Forg org ½PuðIVÞIn DPu ðV aq þ DPu V org Þ
aq ½PuðIVÞOut
þ
V aq ðR20 R19 Þ ðV aq þ DPu V org Þ
16
G.S. Barney, P.B. Duval / Journal of Loss Prevention in the Process Industries 35 (2015) 12e18
It can be seen that the equations derived for calculating the rate of change of concentrations of Pu(IV), PuTotal, Hþ, and NO 3 and temperature have the same form when used for adding aqueous and organic solutions to a tank or to an extractor. The only difference is that the concentration or temperature in the tank or extractor calculated for the previous time step is labeled [X]Tk or [X]Ex or [X]Out. They are all calculated in the same manner, so that it is not necessary to modify the computer code when applying the model to an extractor or a tank. 6. Examples of model results from application to process scenarios Results from model calculations for three different process scenarios are presented in this section. A number of assumptions are required in these calculations because of a lack of data and a need for simplification of the computations. The assumptions made are conservative and lead to results that give higher temperatures and greater quantities of gases generated than would be the case for actual conditions in the process vessels. These assumptions are listed below along with justifications for making them: 1) Extrapolation of rate constants to a temperature higher or lower than the temperature at which the rate constants are measured using the Arrhenius equation requires making the assumption that the reaction mechanism remains constant over the relevant temperature range. This is justified since the temperature range is generally small (a few tens of degrees centigrade) and changes in reaction mechanism are rare for reactions in aqueous solutions over this temperature range. Many of the rate constants used in the model have been shown to follow the Arrhenius law over the entire required temperature range. 2) The heat capacity of the stainless steel vessels is ignored in the calculations due to the complexity of the vessel configuration and heat flow. This results in calculated temperature increases that are higher than actual values, which is a conservative assumption for chemical reaction rates. Heat conduction through the walls of the vessel is assumed to have no temperature gradient and no forced air convection cooling is assumed outside the vessel walls. 3) Instantaneous partitioning of Pu(IV) and HNO3 from the organic phase to the aqueous phase in the extraction column and tank is assumed in order to simplify the calculations. Partitioning of these components in PUREX-type solvent extraction systems is known to be rapid and equilibrium is assumed in most models of these systems. This assumption is conservative since it allows more time for reaction in the aqueous phase. 4) Instantaneous mixing of the entire volume of aqueous solution in the vessels with input solution is assumed (uniform composition throughout the solution volume). This assumption does not account for the multistage countercurrent operation of the pulsed columns (stripping column), which leads to more complete reactions of the solution components. In all cases run with the model, however, the Pu(IV) reduction reactions with HAN are 80e95 percent complete after steady-state operation is reached. 5) For process vessels that contain both aqueous and organic phases, it is assumed that both phases are at the same temperature because of good mixing of the phases, 6) Constant volumes for the aqueous and organic phases are assumed in the pulsed column (stripping column) calculations and input/output solutions to all the vessels are assumed to be additive (no volume changes due to mixing). The model output for major reaction product gases (N2 and N2O)
is given in terms of aqueous solution molarity for ease in calculating the rate of gas generation. Obviously, their solubilities will be exceeded at some point and they will evolve from solution as gases. The solubilities will depend on a number of factors (solution composition, pressure, temperature, etc.). Since these relationships are unknown, the gas generation rate calculation conservatively assumes that the gases are initially saturated in the aqueous solution and any additional N2 or N2O formed will be evolved as gases. The model requires input of initial conditions in the process vessel of interest and input for any flows that are entering the vessel. For the three process scenarios considered here (a plutonium stripping column, an oxidation column, and a plutonium rework tank) values for the initial parameters and the input flow parameters are given in Table 1. 6.1. Stripping column The stripping column is used to separate plutonium from the organic extraction phase (a TBP-branched paraffin hydrocarbon mixture) by reducing the TBP-extractable Pu(IV) nitrate to lessextractable Pu(III) nitrate with aqueous HAN. The purified Pu(III) nitrate moves to the aqueous phase, leaving uranium and other impurities behind in the organic phase. The column operates in a countercurrent mode with the lighter organic phase loaded with Pu(IV) nitrate fed near the bottom of the column and the HANHNO3 stripping solution fed near the top of the column. The column is pulsed so that small droplets of the organic solvent can rise through the continuous aqueous phase and the Pu(IV) can be rapidly stripped and reduced to Pu(III). For the stripping column run simulated by the model, the concentration and temperature curves given in Fig. 1 show that calculated steady state concentrations of Pu(IV), Pu(III), and HAN in the aqueous phase are reached before about 300 min and are about 0.018 M, 0.18 M, and 0.32 M, respectively. This is for the normal operation of the column with feed concentrations, feed rates, and temperatures of the feeds given in the table. The temperature of the solutions in the column are somewhat elevated (about 0.1 C) compared to the feed, due mainly to the exothermic reduction of Pu(IV) with HAN. The effects of the off-normal condition of stopping the aqueous flow and continuing the organic flow are also shown in Fig. 1. When the aqueous flow is stopped and the organic flow continues, the acid concentration increases because of the Pu(IV) reduction with HAN and extraction of acid from the organic phase. The Pu(III) concentration rises as the HAN continues to reduce more Pu(IV) from the organic phase until the HAN is spent (along with the hydrazine and hydrazoic acid) and will be rapidly oxidized by the nitrous acid generated by the autocatalytic reaction. The concentration of Pu(IV) in the aqueous phase never reaches that of the Pu(III) maximum since the organic phase will continue to extract it. Stopping the aqueous flow results in a rapid drop in temperature because the main heat-producing HAN-Pu(IV) reaction slows as the HAN is used up. The autocatalytic reaction that occurs about 760 min after stopping the aqueous flow generates enough heat to raise the temperature of the column by about 0.12 C. The maximum gas generation rate during this reaction was 14 L/min of N2O. 6.2. Oxidation column The purpose of the oxidation column is to remove reducing agents (HAN, hydrazine, and hydrazoic acid) from the aqueous solution coming from the stripping column by oxidation so that they will not cause safety problems downstream. The solution in the column is contacted with a mixture of NOx gases (mainly NO2) and air that is fed to the bottom of the column as the solution flows
G.S. Barney, P.B. Duval / Journal of Loss Prevention in the Process Industries 35 (2015) 12e18
17
Table 1 Compositions of solutions in the vessels initially and solutions flowing into the vessels, volumes, flow rates, temperatures, cell temperatures, and wetted surface area of the vessels. Vessel Variable
Stripping column
[HAN]aq (M) [NO2]aq (M) NO2 gas flow rate (mole/s) [HNO2]aq (M) [HNO2]org (M) [NO]aq (M) [HNO]aq (M) [Hþ]aq (M) (M) [Hþ]org (M) [NO 3 ]aq (M) [NO 3 ]org (M) [N2Hþ 5 ]aq (M) [HN3]aq (M) [N2O]aq (M) [N2]aq (M) [Pu(III)]aq (M) [Pu(IV)]aq (M) [Pu(IV)]org (M) [Fe(III)]aq (M) [FeTotal]aq (M) Vaq (L) Vorg (L) Taq ( C) Torg ( C) Faq (L/s) Forg (L/s) Cell temperature ( C) Wetted surface area (m2)
0.5 4E-12 N/A 4E-10 0 2E-13 0 0.5 0.126 1.1 1.21 0.1 0 1E-5 1E-5 1E-6 1E-5 3.62E-5 1E-6 1.0001E-6 15.6 20.4 40 40 N/A N/A 40 2.40
Oxidation column
Pu rework tank
Fig. 1. Aqueous solution concentrations of chemical species and temperatures for normal and off-normal (aqueous solution flow stopped) process scenarios for the stripping column over time.
down from the top. The column is filled with Lessing rings to allow efficient contact between the gas and liquid phases. The oxidized solution is collected in a 10-L pot below the column. During normal operation, the NO2 feed rate to the column is about 0.01 moles/s, as shown in the table. This oxidation by NO2 results in very low concentrations of HAN and hydrazine in the pot, as shown in Fig. 2. Pu(III) is also oxidized to Pu(IV), nitrous acid concentrations reach a high level, and the temperature increases (about 3.8 C) due to the oxidation reactions and absorption of NO2 by the solution. The off-normal condition of stopping the NOx flow after about 123 min is shown in Fig. 2. The temperature rapidly drops and the concentrations of HAN, hydrazine, HNO2, Pu(III), and Pu(IV) approach those of the aqueous feed solution after about 300 min. The HAN-HNO3 autocatalytic reaction conditions are not
Oxidation column
Pu rework tank
0.25 4E-12 0.01 4E-12 N/A 2E-13 0 0.75 N/A 1.6 N/A 0.06 0.03 1E-5 1E-5 0.185 0.00239 N/A 1E-6 1.0001E-6 N/A N/A 25 N/A 0.005 N/A 25 2.0
0.4 1E-10 N/A 1E-9 0.0001 5E-12 9E-7 0.7 0.2 1.43 0.3 0 0 1E-5 1E-5 0.15 0.00384 0.03 1E-6 1.0001E-6 N/A N/A 40 40 0.003 0.004 40 10.0
Input flow Variables
Initial conditions in vessel 0 4E-12 0.01 0.001 N/A 2E-13 0 0.78 N/A 1.65 N/A 0.06 0 1E-5 1E-5 0.0001 0.185 N/A 1E-6 1.0001E-6 10.0 N/A 25 N/A N/A N/A 25 2.0
Stripping column
0 6E-9 N/A 1E-5 0 1E-10 0 0.1 0.01 0.1 0.000552 0 0 1E-5 1E-5 1E-7 1E-6 1E-9 1E-6 1.0001E-6 30 30 40 40 N/A N/A 40 10.0
0.5 4E-12 N/A 4E-12 0.001 2E-13 0 0.14 0.22 0.71 0.80 0.1 0 1E-5 1E-5 1E-5 0 0.145 1E-6 1.0001E-6 N/A N/A 40 40 0.0055 0.073 40 2.40
Fig. 2. Aqueous solution concentrations of chemical species and temperatures for normal and off-normal (NOx flow stopped) process scenarios for the oxidation column over time.
reached during this process scenario. 6.3. Plutonium rework tank The plutonium rework tank is a 300 L slab tank that receives solutions from the stripping column and can recycle the organic phase to the stripping column. The aqueous solution from the stripping column, which normally contains HAN, hydrazine, and hydrazoic acid, is sent to a second tank where these reducing agents are oxidized with NOx gas. In this computer simulation, aqueous and organic solutions from the stripping column are fed to the tank that initially contains 30 L each of aqueous and organic solutions. The compositions, temperatures and feed rates of the initial and input solutions are given in Table 1. Fig. 3 shows a
+
G.S. Barney, P.B. Duval / Journal of Loss Prevention in the Process Industries 35 (2015) 12e18
10
1
Flows To Tank Stopped
40.020
-1
10
40.015
-3
10
[NH OH ] [Pu(IV)] [Pu(III)] [Pu(IV)] [HNO ] Temperature
-5
10
-7
10
40.010
Temperature, °C
[NH3OH ], [Pu(IV)], [PuIII)], [HNO2], M
18
40.005
-9
10
10
2
3
4 5 6
100
2
3
4 5 6
1000
2
3
4 5 6
Time, min Fig. 3. Aqueous and organic solution concentrations of chemical species and temperatures for normal and off-normal (aqueous and organic flows stopped) process scenarios for the plutonium rework tank over time.
Fig. 4. Gas generation rates for N2O and N2 gases and solution temperatures during the autocatalytic reaction in the plutonium rework tank.
gradual increase in the concentrations of HAN, Pu(III), and Pu(IV) in the aqueous phase as the feed solutions are added to the tank. This is expected since the original concentration of each of these components in the tank is very low or zero. The gradual decrease in the HNO2 concentration is due to dilution of the initially higher concentration in the tank (1E-5 M) with the lower feed concentration (1E-9). The small temperature increase is mainly the result of HAN reduction of Pu(IV) from the organic phase of the feed solution. After about 400 min the volumes of aqueous and organic phases in the tank are about 102 L and 126 L, respectively, and the flows to the tank are stopped. Concentration of Pu(III) increases slightly and the concentration of Pu(IV) decreases initially in the aqueous phase as the reduction of Pu(IV) by HAN continues. HAN concentration decreases slowly at first and then more rapidly as the autocatalytic
reaction is approached at about 6300 min (4.38 days). Since HNO2 is a product of the autocatalytic reaction (Reaction (2)), its concentration slowly increases initially until the rapid autocatalytic reaction is approached. The temperature of the solutions rapidly increase by about 0.022 C due to the reaction. Pressurization of the tank is a safety concern when the autocatalytic reaction occurs because the reaction produces N2O gas as a major product. The rate of gas generation can be calculated using the model and the results for this tank are given in Fig. 4. The figure shows that the maximum gas generation rate for N2O is about 10 L/ min for the scenario used in the model run. The maximum pressure in the tank resulting from this gas generation rate is not calculated here, but will depend on the gaseous headspace in the tank and on the vent size and configuration and on the temperature. 7. Conclusions A mathematical model for predicting the stability of HAN in plutonium-containing process solutions is applied to three typical types of process vessels in solvent extraction plutonium processes and for normal and off-normal process scenarios. The model allows calculation of the effects of important process parameters such as temperature, volume, flow rates, gas generation rates, and concentrations of chemical components over time. Two of the offnormal scenarios led to the rapid autocatalytic reaction of HAN with nitric acid where heat and gas are generated and that could lead to damage of the process equipment and/or release of hazardous plutonium solution from the vessel. Both of these autocatalytic reactions resulted from the aqueous phase of the process solutions being allowed to remain stationary in the vessel while the Pu(III) in these solutions slowly generated nitrous acid by Reaction (4) and which used up the HAN by Reactions (3) and (1). Eventually, at some combination of temperature, concentration of nitric acid, nitrous acid and HAN, the solution becomes unstable and the autocatalytic reaction occurs. It can be concluded that all solutions containing nitric acid, plutonium and HAN are inherently unstable and will eventually lead to the autocatalytic reaction occurring. References Barney, G.S., Duval, P.B., 2011. Model for predicting hydroxylamine stability in plutonium process solutions. J. Loss Prev. Process Ind. 24, 76e84. Dean, J.A., 1979. Lange's Handbook of Chemistry, 12th ed. McGraw-Hill Book Company, Inc., New York. Gronier, W.S., Rainey, R.H., Watson, S.B., 1971. An analysis of the transient and steady-state operation of countercurrent liquid-liquid solvent extraction process. Ind. Eng. Chem. Process Des. Dev. 15, 385e390. McAdams, W.H., 1964. Heat Transmission, third ed. McGraw-Hill Book Company, Inc., New York. Perry, R.H., Green, D., 1997. Perry's Chemical Engineering Handbook, seventh ed. McGraw-Hill Book Company, Inc., New York. Weast, R.C. (Ed.), 1980. CRC Handbook of Chemistry and Physics, 61st ed. CRC Press, Inc., Boca Raton, Florida.