Decontamination of VX from Silicone: Characterization of ...
Decontamination of VX from Silicone: Characterization of Multicomponent Diffusion Effects Mark J. Varady,1 Thomas P. Pearl,1 Shawn M. Stevenson,2 and Brent A. Mantooth2* 1
OptiMetrics, Inc., a DCS Company, 100 Walter Ward Boulevard, Suite 100, Abingdon, MD,
21009, United States 2
U.S. Army Edgewood Chemical Biological Center, 5183 Blackhawk Road, Aberdeen Proving
Ground, MD, 21010-5424, United States *[email protected], Tel. (+1) 410.436.0967
S1
Numerical Simulations Four different types of simulations were performed in this work: (1) uptake of pure solvent into silicone, (2) VX contamination of silicone from a sessile droplet, (3) liquid-phase extraction of VX from silicone, and (4) decontamination of VX contaminated silicone. In all cases, the species transport equations in silicone were implemented in COMSOL v 5.1 using the “General Form PDE” interface. The solver settings were consistent across all simulations, and departures from default settings are listed below: •
Time dependent solver: enable nonlinear controller, max BDF order = 5
•
Fully coupled solution technique: Update Jacobian on every iteration, tolerance factor = 0.01, max iterations = 25
To simulate the absorption of pure methanol in silicone, a 3.1 mm line was created in 1D geometry to represent the thickness of the silicone disc. Eq. 4 was implemented using the “General Form PDE” interface. Dirichlet BCs were imposed at the ends of the line, setting ϕm = 0.033, consistent with the value reported in Table 1 of the manuscript. The predefined mesh size of “Extremely Fine” was used to create a uniform mesh of 100 Lagrange 2nd order elements in the domain. To simulate the absorption of the VX droplet in silicone, a rectangle 3.1 mm × 25.4 mm was created in 2D-axisymmetric geometry to represent the silicone disc. A point was placed at r = 1.75 mm from the centerline at the top of the rectangle to represent the edge of the VX droplet. Eq. 4 was implemented using the “General Form PDE” interface. A Dirichlet BC was imposed along the line 0 < r < 1.75 mm representing the VX/silicone boundary, and the time varying value of ϕVX along the boundary was computed according to eqs. 5 and 11 in the manuscript. Equation 5 was implemented using the “Global ODE and DAE” interface. A mesh size of 0.31
S2
mm was used on the domain and transitioned to a mesh size of 0.0875 mm at the VX/silicone interface. Triangular Lagrange 2nd order elements were used. To simulate the extraction of VX from silicone, rectangles 3.1 mm × 25.4 mm and 20 mm × 25.4 mm were created in 2D-axisymmetric geometry to represent the silicone disc and extraction solvent, respectively. Eq. 3 was implemented using the “General Form PDE” module for both VX and the extraction solvent (methanol or water). Eq. 1 was implemented using the “Domain ODE” interface to solve for the fluxes of VX and extraction solvent in silicone. The fluxes were resolved to their scalar components along the coordinate axes to facilitate this implementation. Eq. 3 was implemented along the boundary between the silicone and liquid domains using the “Boundary ODE” interface to provide the Dirchlet BCs for the volume fraction of VX and the extraction liquid in silicone. Transport of VX in the liquid-phase, eq. 6, was implemented using the “Transport of Diluted Species” interface. The concentration BC along the liquid/silicone boundary was calculated from eq. 13. The other BC at the boundary equated the normal flux in the liquid to the normal flux of VX in silicone computed by the “Domain ODE” interface. For the simulation of decontamination of silicone, the procedure was almost identical to that described above for the extraction of VX from silicone. The only differences were that the rectangle representing the liquid-phase was 8 mm × 25.4 mm, and reaction in the liquid-phase was included by adding a “Reaction” node to the “Transport of Diluted Species” interface.
VX Molar Volume by Group Contribution Method A group contribution method1 was used to compute the molar volume of VX, VVX , according to: VVX = ∑ N iVi
(1)
i
S3
where Ni is the number of molecular groups i contained in VX and Vi is the corresponding molar volume of a single molecular group. The molecular groups contained in VX and their molar volumes are identified in Figure S1, with the values taken from Barton,1 Chapter 6, Table 3.
Figure S1. Identification of molecular groups and corresponding molar volumes in VX for calculation of molar volume by the group contribution method.
VX Absorption in Silicone Pretreated with Methanol A silicone disc was immersed in a bath of pure methanol for 60 min, excess liquid methanol on the surface of the disc was removed, and a 2 µL droplet of VX was deposited on the pretreated disc. The mass of VX absorbed over a 60 min contamination time was determined via extraction of the silicone disc in 2-propanol and subsequent chemical analysis. The absorbed mass of VX for the pretreated silicone disc was 220 µg compared to 332 µg for a clean silicone disc.
Liquid-Phase Reaction Kinetics To obtain kinetic data for the reaction between VX and hydroxide dissolved in different water or methanol solutions, time-resolved measurements of VX consumption and reaction byproducts formation in the liquid phase were recorded. Reaction product concentrations were evaluated at a hydroxide solution concentration of 100 mM in water/methanol mixtures of varying volumetric
S4
ratios. Glass vials were filled with 4.9 mL of a NaOH solution and then dosed with 100 µL of a dilute VX solution (250 mM in methanol). The liquid was stirred to keep the reacting species well-mixed. Samples (10 µL) were taken at predetermined time points and placed in 5 mL of 2propanol to quench the reaction by dilution. The dilute samples were analyzed by LC/MS/MS. The liquid-phase disappearance rate of VX was modeled as a second order global (first order in each the agent and decontaminant) kinetic expression,
dCVX ,l dt
=
dCOH ,l dt
= −rl = −kl CVX ,l COH ,l
(2)
where rl is the reaction rate between the agent and decontaminant active species, kl is the reaction rate constant, and CVX,l and COH,l are the concentrations of the agent and reactive species of the decontaminant, respectively. Figure S2 shows fits of the kinetic expression to experimental data for sodium hydroxide in solution with pure water and with pure methanol.
Figure S2. Experimental results (symbols) from well-stirred liquid-phase reaction of VX with sodium hydroxide in solutions of water and methanol. Also shown are best-fit models (lines) and corresponding rate constants.
S5
VX Reaction with Sodium Hydroxide in Silicone Generally, the reaction between VX and hydroxide could occur in both liquid and polymer phases if both agent and reactive species are present. To assess the extent of chemical reaction in each phase, a silicone disc was contaminated with a 2 µL droplet of 89% high purity VX, with impurities that include EMPA and EA2192. After a 60 min contamination time, the surface of the silicone disc was rinsed with water to remove the bulk liquid VX from the surface of the silicone, followed by decontamination for 30 min by submersion of the substrate in 8 mL of 1000 mM NaOH in four different methanol/water mixtures (0%, 30%, 70%, and 100% methanol by volume). The silicone disc was then removed from the decontaminant solution and extracted in 2-propanol to quantify the remaining absorbed VX and reaction byproducts.
Both the
decontaminant solution and the silicone extractant were analyzed using LC/MS/MS to determine the quantity of VX and two common VX reaction byproducts EMPA (ethyl methylphosphonic acid) and EA2192 (ethyl methylphosphonothioic acid) in each phase. The results shown in Figure S3 demonstrate that the ratio of total moles of EMPA and EA2192 to moles of VX within the polymer after decontamination was 1–2 orders of magnitude lower than in the decontaminant solution, indicating that the decontamination reaction occurred primarily in in the liquid-phase over the 30 min decontamination time. Also, the molar ratio of byproducts to VX in the polymer was approximately equivalent to that in the VX droplet deposited on the silicone surface at the start of the contamination process. Thus, reaction between VX and hydroxide in the silicone was neglected in the model over the 30 min decontamination time. Note that a slower degradation of VX occurs in the polymer that does not involve reaction with hydroxide, and does not proceed to a measurable extent over the 30 min decontamination time. Also of note is the apparent decrease
S6
of the reaction rate in the liquid-phase as the water/methanol ratio was decreased. This is consistent with the observed trend in kinetic rate constant as shown in Figure S2.
Figure S3. Ratio of total moles of reaction byproducts EMPA and EA2192 to VX within the polymer (gray bars), and in the liquid-phase (green bars) after 30 min decontamination using 1000 mM NaOH in four different methanol/water mixtures.
S7
REFERENCES: (1) Barton, A. F. M. CRC Handbook of Solubility Parameters and Other Cohesion Parameters. CRC Press: Boca Raton, FL, 1991.
S8
OptiMetrics, Inc., a DCS Company, 100 Walter Ward Boulevard, Suite 100, Abingdon, MD,
21009, United States 2
U.S. Army Edgewood Chemical Biological Center, 5183 Blackhawk Road, Aberdeen Proving
Ground, MD, 21010-5424, United States *[email protected], Tel. (+1) 410.436.0967
S1
Numerical Simulations Four different types of simulations were performed in this work: (1) uptake of pure solvent into silicone, (2) VX contamination of silicone from a sessile droplet, (3) liquid-phase extraction of VX from silicone, and (4) decontamination of VX contaminated silicone. In all cases, the species transport equations in silicone were implemented in COMSOL v 5.1 using the “General Form PDE” interface. The solver settings were consistent across all simulations, and departures from default settings are listed below: •
Time dependent solver: enable nonlinear controller, max BDF order = 5
•
Fully coupled solution technique: Update Jacobian on every iteration, tolerance factor = 0.01, max iterations = 25
To simulate the absorption of pure methanol in silicone, a 3.1 mm line was created in 1D geometry to represent the thickness of the silicone disc. Eq. 4 was implemented using the “General Form PDE” interface. Dirichlet BCs were imposed at the ends of the line, setting ϕm = 0.033, consistent with the value reported in Table 1 of the manuscript. The predefined mesh size of “Extremely Fine” was used to create a uniform mesh of 100 Lagrange 2nd order elements in the domain. To simulate the absorption of the VX droplet in silicone, a rectangle 3.1 mm × 25.4 mm was created in 2D-axisymmetric geometry to represent the silicone disc. A point was placed at r = 1.75 mm from the centerline at the top of the rectangle to represent the edge of the VX droplet. Eq. 4 was implemented using the “General Form PDE” interface. A Dirichlet BC was imposed along the line 0 < r < 1.75 mm representing the VX/silicone boundary, and the time varying value of ϕVX along the boundary was computed according to eqs. 5 and 11 in the manuscript. Equation 5 was implemented using the “Global ODE and DAE” interface. A mesh size of 0.31
S2
mm was used on the domain and transitioned to a mesh size of 0.0875 mm at the VX/silicone interface. Triangular Lagrange 2nd order elements were used. To simulate the extraction of VX from silicone, rectangles 3.1 mm × 25.4 mm and 20 mm × 25.4 mm were created in 2D-axisymmetric geometry to represent the silicone disc and extraction solvent, respectively. Eq. 3 was implemented using the “General Form PDE” module for both VX and the extraction solvent (methanol or water). Eq. 1 was implemented using the “Domain ODE” interface to solve for the fluxes of VX and extraction solvent in silicone. The fluxes were resolved to their scalar components along the coordinate axes to facilitate this implementation. Eq. 3 was implemented along the boundary between the silicone and liquid domains using the “Boundary ODE” interface to provide the Dirchlet BCs for the volume fraction of VX and the extraction liquid in silicone. Transport of VX in the liquid-phase, eq. 6, was implemented using the “Transport of Diluted Species” interface. The concentration BC along the liquid/silicone boundary was calculated from eq. 13. The other BC at the boundary equated the normal flux in the liquid to the normal flux of VX in silicone computed by the “Domain ODE” interface. For the simulation of decontamination of silicone, the procedure was almost identical to that described above for the extraction of VX from silicone. The only differences were that the rectangle representing the liquid-phase was 8 mm × 25.4 mm, and reaction in the liquid-phase was included by adding a “Reaction” node to the “Transport of Diluted Species” interface.
VX Molar Volume by Group Contribution Method A group contribution method1 was used to compute the molar volume of VX, VVX , according to: VVX = ∑ N iVi
(1)
i
S3
where Ni is the number of molecular groups i contained in VX and Vi is the corresponding molar volume of a single molecular group. The molecular groups contained in VX and their molar volumes are identified in Figure S1, with the values taken from Barton,1 Chapter 6, Table 3.
Figure S1. Identification of molecular groups and corresponding molar volumes in VX for calculation of molar volume by the group contribution method.
VX Absorption in Silicone Pretreated with Methanol A silicone disc was immersed in a bath of pure methanol for 60 min, excess liquid methanol on the surface of the disc was removed, and a 2 µL droplet of VX was deposited on the pretreated disc. The mass of VX absorbed over a 60 min contamination time was determined via extraction of the silicone disc in 2-propanol and subsequent chemical analysis. The absorbed mass of VX for the pretreated silicone disc was 220 µg compared to 332 µg for a clean silicone disc.
Liquid-Phase Reaction Kinetics To obtain kinetic data for the reaction between VX and hydroxide dissolved in different water or methanol solutions, time-resolved measurements of VX consumption and reaction byproducts formation in the liquid phase were recorded. Reaction product concentrations were evaluated at a hydroxide solution concentration of 100 mM in water/methanol mixtures of varying volumetric
S4
ratios. Glass vials were filled with 4.9 mL of a NaOH solution and then dosed with 100 µL of a dilute VX solution (250 mM in methanol). The liquid was stirred to keep the reacting species well-mixed. Samples (10 µL) were taken at predetermined time points and placed in 5 mL of 2propanol to quench the reaction by dilution. The dilute samples were analyzed by LC/MS/MS. The liquid-phase disappearance rate of VX was modeled as a second order global (first order in each the agent and decontaminant) kinetic expression,
dCVX ,l dt
=
dCOH ,l dt
= −rl = −kl CVX ,l COH ,l
(2)
where rl is the reaction rate between the agent and decontaminant active species, kl is the reaction rate constant, and CVX,l and COH,l are the concentrations of the agent and reactive species of the decontaminant, respectively. Figure S2 shows fits of the kinetic expression to experimental data for sodium hydroxide in solution with pure water and with pure methanol.
Figure S2. Experimental results (symbols) from well-stirred liquid-phase reaction of VX with sodium hydroxide in solutions of water and methanol. Also shown are best-fit models (lines) and corresponding rate constants.
S5
VX Reaction with Sodium Hydroxide in Silicone Generally, the reaction between VX and hydroxide could occur in both liquid and polymer phases if both agent and reactive species are present. To assess the extent of chemical reaction in each phase, a silicone disc was contaminated with a 2 µL droplet of 89% high purity VX, with impurities that include EMPA and EA2192. After a 60 min contamination time, the surface of the silicone disc was rinsed with water to remove the bulk liquid VX from the surface of the silicone, followed by decontamination for 30 min by submersion of the substrate in 8 mL of 1000 mM NaOH in four different methanol/water mixtures (0%, 30%, 70%, and 100% methanol by volume). The silicone disc was then removed from the decontaminant solution and extracted in 2-propanol to quantify the remaining absorbed VX and reaction byproducts.
Both the
decontaminant solution and the silicone extractant were analyzed using LC/MS/MS to determine the quantity of VX and two common VX reaction byproducts EMPA (ethyl methylphosphonic acid) and EA2192 (ethyl methylphosphonothioic acid) in each phase. The results shown in Figure S3 demonstrate that the ratio of total moles of EMPA and EA2192 to moles of VX within the polymer after decontamination was 1–2 orders of magnitude lower than in the decontaminant solution, indicating that the decontamination reaction occurred primarily in in the liquid-phase over the 30 min decontamination time. Also, the molar ratio of byproducts to VX in the polymer was approximately equivalent to that in the VX droplet deposited on the silicone surface at the start of the contamination process. Thus, reaction between VX and hydroxide in the silicone was neglected in the model over the 30 min decontamination time. Note that a slower degradation of VX occurs in the polymer that does not involve reaction with hydroxide, and does not proceed to a measurable extent over the 30 min decontamination time. Also of note is the apparent decrease
S6
of the reaction rate in the liquid-phase as the water/methanol ratio was decreased. This is consistent with the observed trend in kinetic rate constant as shown in Figure S2.
Figure S3. Ratio of total moles of reaction byproducts EMPA and EA2192 to VX within the polymer (gray bars), and in the liquid-phase (green bars) after 30 min decontamination using 1000 mM NaOH in four different methanol/water mixtures.
S7
REFERENCES: (1) Barton, A. F. M. CRC Handbook of Solubility Parameters and Other Cohesion Parameters. CRC Press: Boca Raton, FL, 1991.
S8
