Chem. Res. Toxicol. 1994, 7, 568-574
568
Kinetics of the Reaction of Nitric Oxide with Oxygen in Aqueous Solutions Randy S. Lewis and William M. Deen* Department of Chemical Engineering, Massachusetts Institute of Technology, Cambridge, Massachusetts 02139 Received March 24, 1994"
An understanding of the rate of reaction of nitric oxide (NO) with oxygen in aqueous solutions is needed in assessing the various actions of NO in the body. A novel approach was developed for studying the kinetics of this reaction, which permitted simultaneous and continuous measurements of the concentrations of NO and the principal product, nitrite (NO2-). Nitric oxide was measured using a chemiluminescence detector, with continuous sampling achieved by diffusion of NO through a membrane fitted into the base of a small, stirred reactor. The results with various initial NO and 0 2 concentrations confirmed that the rate of reaction is second-order in NO and first-order in 0 2 . T h e rate of reaction of NO was described by the expression 4kl [N0I2[O2],where kl was (2.1 f 0.4) X lo6 M-2 s-1 a t 23 O C and (2.4f 0.3) X lo6 M-2 s-l a t 37 "C. The value of klwas the same at p H 4.9 and 7.4. The rate of formation of NO2equaled the rate of reaction of NO (within experimental uncertainty of a few percent), and there was no detectable formation of nitrate (Nos-). This confirmed that NO and NO2- were the only NO, species present in significant amounts and supported the validity of pseudo-steady-state assumptions for NO2 and N2O3, which are intermediates in the conversion of NO to NOa-. A kinetic model was developed to predict temporal variations in the concentrations of NO, 02, NOz-, NO2, and N2O3 in the system studied, which were due in part to transport across the gas-liquid interface and across the membrane a t the base of the reactor. T h e analysis emphasizes the importance of quantifying interphase mass transfer when studying reaction rates in open systems, a point which has not always been recognized in previous studies with nitric oxide.
Introduction Nitric oxide (NO) is synthesized in the body by many types of cells, including endothelial cells, macrophages, neutrophils, neurons, and hepatocytes. Important physiological roles of NO include blood pressure regulation, inhibition of platelet aggregation, and neurotransmission ( I ) . Nitric oxide also has cytotoxic and mutagenic effects, which have been attributed largely to trace amounts of nitrosating agents and oxidants which are formed by the reactions of NO with molecular oxygen or superoxide (2, 3). The specific effects of NO in a given tissue are undoubtedly related to the local concentrations of NO and its oxidation products, although there tends to be considerable uncertainty about actual concentration levels. An understanding of the potent and varied biological actions of NO clearly requires a knowledge of the amounts of NO and its products present, which in turn requires an accurate description of the kinetics of NO oxidation in physiological solutions. The overall stoichiometry of the reaction of NO with 0 2 in aqueous solutions is usually expressed as1 4N0
+ 0, + 2H20
ko
4H'
+ 4NO;
(1)
* Address correspondence to Department of Chemical Engineering, Room 66-509, Massachusetts Institute of Technology, Cambridge, MA 02139. Tel: (617) 253-4535; FAX: (617)258-8224. 0 Abstract published in Advance ACS Abstracts, June 15, 1994. 1 Abbreviations: k, mass transfer coefficient; A, mass transfer area; V , aqueous volume; kA/ V, volumetric mass transfer coefficient; a, gas solubility in silastic; D , diffusivity; S,gas solubility in solution.
The kinetics of this reaction have been the subject of a number of studies (4-7), in which the usual approach in measuring the rate has been to monitor the appearance of nitrite (NO2-). The available results generally support the view that the rate of nitrite appearance is secondorder in NO and first-order in 0 2 (4-6), although there have been inconsistencies in the reported values of the rate constant (ko)and in the temperature dependence of ko. In the one study in which the rate was measured by following the disappearance of NO, the results were interpreted as showing zero-order kinetics in NO (7). In none of these kinetic studies were NO and NO2- concentrations measured simultaneously, which would permit a check on the overall nitrogen balance implied by eq 1. In view of the biological importance of NO oxidation kinetics and the aforementioned inconsistencies in the literature, we developed a novel reactor configuration for studying the reaction of NO with 0 2 . A key feature of the design is that it permits NO and nitrite concentrations to be monitored continuously and simultaneously. The detection of NO was by the well-established chemiluminescence method, with continuous sampling accomplished by permeation of NO through a membrane fitted into the base of the reactor. This sampling method is similar to those described recently for NO detection in gas mixtures or aqueous solutions using a mass spectrometer (8). The kinetic results are interpreted in terms of a set of elementary reactions which is consistent with eq 1, and the estimation of the concentrations of the likely intermediates nitrogen dioxide and nitrous anhydride (NO2 and N203) is discussed.
0S93-228~/94/2707-0568$04.50/0 0 1994 American Chemical Society
Chem. Res. Toxicol., Vol. 7, No. 4, 1994 569
Oxidation Kinetics of Nitric Oxide KS outlet and samplelpurge access (via needles)
r/ c from spectrophotometer
+ to spectrophotometer
5
4
gas inlet +
3
2 6.2 cm
n
1
0 silastic sheet (0.13 nun) teflon sheet (1")
0
--
1
to chemiluminescence detector
5
15
10
20
25
30
35
40
NO Concentration (&M)
-
Figure 1. Schematic of apparatus used to study NO oxidation kinetics. The reactor was a modified 200-mL stirred ultrafiltration cell. A composite membrane at the base of the reactor, consisting of a Silastic membrane laminated to a Teflon sheet, allowed continuous entry of NO into a chemiluminescence detector. The flow loop connected to a spectrophotometer was used for continuous monitoring of NO2-.
Figure 2. Response of the chemiluminescence detector as a function of aqueous NO concentration. At the attenuation used (256X) the noise in the detector response was approximately 0.05 V. Ar purge
NO addition
Materials and Methods Note: All experiments were performed in a certified hood due to the potential toxicity of NO. Reagents. Nitric oxide was passed through a column containing 10 M NaOH to remove NO. impurities. Argon, after passage through an oxygen trap, was mixed with NO using controlled gas flow meters (Porter Instrument Co., Hatfield, PA) to obtain the desired NO gas concentration. Air and 5% 0 2 in N2were the 0 2 sources. Buffer solutions at 0.01 M ionic strength and pH 4.9 or 7.4 were prepared from the acids and sodium salts of malonate and phosphate, respectively. Reactor. As shown schematically in Figure 1,the reactor was a modified 200-mL stirred ultrafiltration cell (Amicon, Danvers, MA, Model 8200), with the stirrer replaced by that from a CYTOSTIR stirred bioreactor (Kontes,Vineland, NJ). A septum port, two ports for a flow loop connected to a spectrophotometer (for NO2- monitoring), a gas inlet port, and a thermometer were added. A hypodermic needle was inserted into the septum port for periodic aqueous sampling of nitrate (NO3-) as well as gas purging, and another hypodermic needle was inserted to provide a gas outlet. The base of the stirred cell was fitted with a composite membrane (6.2-cm diameter) consisting of a 0.13 mm thick Silastic sheet (Mempro, Troy, NY) on top of a 1mm thick Teflon sheet. The Teflon sheet contained four symmetrically positioned holes of 0.6-cm diameter, as shown. The purpose of the Teflon layer was to limit the amount of NO which could diffuse across the composite membrane and enter the chemiluminescence detector. Studies were performed with the reactor exposed to room air (23 1 "C) or immersed in a water bath (37 1 OC). Nitric Oxide Analysis. The external side of the composite membrane was exposed to high vacuum and connected to a chemiluminescence detector (ThermedicaDetection Inc., Woburn, MA, Model TEA-502) for monitoring of NO. The response time in this system is governed by the time required for transport of NO through the liquid boundary layer at the base of the cell and for diffusion of NO through the membrane. The half-time for the response of the detector to a step change in NO concentration was -2 s, similar to the response time measured using mass spectrometry (8). The output of the detector was linear up to NO concentrations of at least 40 pM, as shown in Figure 2. The minimum aqueous NO concentration measurable by the chemiluminescence detector was estimated at -0.01 pM (signal-tonoise ratio of 3). Nitrite and Nitrate Analysis. The flow loop attached to the reactor consisted of 1/8-in. tubing and had a total volume of -8 mL. The aqueous solution was circulated continuously by
*
*
7 0
20
40
60
80
100
120
Minutes
Figure 3. Response of chemiluminescencedetector as a function of time during a typical NO oxidation experiment. The signal remained at base line until NO was introduced into the aqueous solution (t = 40 min). After achieving a steady-state level of NO (t = 70 min), oxygen was introduced to initiate the NO oxidation reaction. a pulseless pump (Cole Parmer, Chicago, IL, Models 000-305 and 184-000) at 45 mL/min through the 10-mm flow cell of a spectrophotometer (Shimadzu Model UV16Ou). The NOzconcentration was proportional to the absorbance measured at 209 nm. Nitrate (Nos-) was analyzed in 1-mL aqueous samples withdrawn from the reactor at 5-min intervals, using the Griess procedure (9). The samples were rapidly purged with an inert gas to remove NO, which if present would contribute to the Griess reaction. Both NO2- and NOS- had minimum detectable limits of -1 pM. Nitric Oxide Oxidation Experiments. The buffer solution (150 mL) was first added to the reactor, stirring initiated at 100 rpm, and recirculation begun through the flow loop. The solution was then bubbled with Ar for at least 40 min to remove 0 2 , after which the NO/Ar gas mixture was bubbled into the solution for at least 30 min to obtain the desired aqueous NO concentration. The aqueous NO concentration was calculated from the known NO gas concentration and the solubility (IO). After a steadystate NO concentration was obtained in the liquid, the bubbling of the NO/Ar mixture was terminated and the OdNz mixture was introduced via the gas inlet at flow rates of 350 and 300 sccm at 23 and 37 OC, respectively. Diffusion of 02 into the liquid initiated aqueous NO oxidation, which was then monitored for at least 30 min. A typical chemiluminescencedetector signal for an NO oxidation experiment is shown in Figure 3. Mass Transfer Coefficients. Because there was transport of NO and 02 to and from the aqueous solution, interpretation of the results required knowledge of the corresponding mass transfer rates. The liquid-phase mass transfer coefficients' at the gas-liquid interface and at the base of the stirred cell are , the corresponding surface denoted by k~ and k ~respectively;
570 Chem. Res. Toxicol., Vol. 7, No. 4,1994
Lewis and Deen
areas1 available for mass transfer are A0 and AB. The specific quantities which had to be determined were k&/V and k & j / V, where Vis the aqueous volume.' We will refer to these lumped quantities as the "volumetric mass transfer coefficients".l Two independent measurements were used, one which determined the sum of the two volumetric mass transfer coefficients,and one which determined the value at the base. Accordingly, the value at the gas-liquid interface was calculated from the difference. Once the values were determined for NO, they could be calculated for 02, as will be described. In the absence of reactions (i.e., without 02present), the massbalance equation which describes the depletion of NO from the liquid is
provided that there is no NO in the gas phase. The s u m of the volumetric mass transfer coefficients for NO was obtained by following the NO oxidation protocol described above, except that after a steady-state NO concentration had been achieved, Ar rather than an Oz/N2mixture was flowed through the head space to cause NO depletion. The measurements were done at both 23 and 37 OC. Integration of eq 2 shows that a plot of ln([NO]/ [Nolo) versus time ( t ) ,where [NO10 is the initial NO concentration, will have a slope with magnitude equal to the sum of the volumetric mass transfer coefficients for NO. Characterizing mass transfer at the base of the stirred cell was more complicated. A consideration of series resistances indicates that k&/V is related to the corresponding coefficients for the aqueous boundary layer (kAAB/V) and the composite membrane (kM&/V), such that
--
1 (kBA$ V)
--
(k&
l
+ V)
1 (k&$ V)
(3)
The membrane resistance for NO and 0 2 was calculated from the known properties of Silastic, while the aqueous boundary layer resistance was measured using dissolution of benzoic acid, as described below. Regarding the membrane term in eq 3, the value of k M is equal to aD/SL, where aD is the Silastic permeability, comprised of the gas solubility (a)'and diffusivity (D)l in Silastic, S is the gas solubility in aqueous solution,' and L is the Silastic thickness. The permeabilities (aD)of NO and 02 in Silastic are equivalent, with values of 1.7 X 1 k 1 6 and 1.9 X 1kl6mol cm-l s-1 Pa-l at 23 and 37 OC, respectively (11). The solubilities (S)at 23 and 37 OC are 0.019 and 0.016 M/MPa for NO, and 0.013 and 0.011 M/MPa for 02,respectively (10). Therefore, with a volume (V) of 150 mL, a Silastic membrane thickness (L) of 0.013 cm, and a mass transfer area (AB,the total area of the holes in the Teflon sheet) of 1.13 cm2, kMAB/V for NO at 23 and 37 OC is approximately 0.5 X lo-' s-1 and 0.7 X lo-' e-', respectively. The corresponding values for 0 2 are 0.8 X lo-' s-1 and 1.0 X lo-' s-l. The aqueous boundary layer term in eq 3 was evaluated by studying the dissolution of benzoic acid from the base of the stirred cell at 23 OC. In these studies the Silastic membrane was omitted. Benzoic acid pellets were rapidly melted and poured into the four holes of the Teflon sheet. After the benzoic acid crystallized, it was sanded for smoothness. The sheet was placed in the reactor with 150 mL of deionized water and stirring maintained at 100 rpm. The reactor and a saturated solution of benzoic acid were placed in the same water bath. The concentrations of saturated benzoic acid solutions are very temperaturesensitive at room conditions (12),thus requiring care in keeping the saturated solution exactly at the reactor temperature. The reactor solution was sampled at various time intervals, and the saturated solution was sampled at the end of the experiment. Each sample was assayed with a spectrophotometer at 271 nm, where the absorbance was shown to be linear in the benzoic acid concentration. The mass balance equation which describes the dissolution of benzoic acid (PhCOOH) is
d[PhCooH3 = *([PhCOOH]* dt V
- [PhCOOHl)
(4)
where [ PhCOOH]* is the saturated benzoic acid concentration adjacent to the solid. Integration of the above equation shows that a plot of ln{l - ([PhCOOH]/[PhCOOH]*)J versus time will yield a slope equal to k&/V for benzoic acid. Laminar boundary layer theory (13)was used to obtain values of k&/Vfor NO and 02,based upon that measured for benzoic acid. These results and the calculated values of kMAB/V were used in eq 3 to obtain k&/V for NO and 0 2 . The value of k&/V for NO, together with the results of the experiment discussed in connection with eq 2, were used to obtain k&G/V for NO. Finally, boundary layer theory was used again, to calculate kdG/Vfor 02 from that for NO. For chemical species i and j subjected to identical hydrodynamic conditions (same stirred cell and same stirring rate), boundary layer theory indicates that
where Dj and Dj are the liquid-phase diffusivities. The required diffusivities for NO, 02,and benzoic acid in water were obtained from the literature (14-16). In addition to extrapolating results from one chemical species to another, eq 5 was used where necessary to derive temperature corrections; it ignores the minor effects of viscosity differences at the two temperatures. Kinetic Model. Eq 1was replaced by a more detailed kinetic scheme (4, 6, 17): 2N0 + 0,
-
2N02
(6)
+ NO, eks N203
(7)
ki
NO
N203+ H,O
k4
2H'
+ 2NO;
(8)
The volumetric rate of formation of chemical species i (in M 8-1) is denoted as Ft+ The rate constants in eqs 6-8 are defined such that the net rates for NO, 02,and NO2- are given by RNo = -2k1[N012[0,1 - k,[NOI[NO21
+ k,[N20,1
(9)
Order-of-magnitude estimates of the rate constants kl through 4 (5, 18) suggest that in our experiments NO2 and NzOs were present only in very small amounts, justifying pseudo-steadystate approximations for those species (i.e., R, cz 0 for i = NO2 or N203). Using those approximations to eliminate [NO21 and [NzOsl as independent variables, the mass balance equations for the dominant species are
d[N03 = -4kl[N0l2[O21 + ( ykGAG ) N o ( [ N O ] * - [NO]) dt ~BAB
(12) (7 NO' )
-d[021 - -kl[N012[0,] + ( ~ ) 0 1 ( [ 0 2 1 *- [O,]) dt
(13)
Chem. Res. Toxicol., Vol. 7, No. 4, 1994 571
Oxidation Kinetics of Nitric Oxide 1.00
L
80
$
0.90
8 2
-a 0.80 L
0
, 20
1
1
40
60 Minutes
I
80
, 100
0.1
'
0
120
I
I
5
10
, 15
I 25
20
35
30
Minutes
Figure 4. Rate of change in benzoic acid (PhCOOH) concentration resulting from convective diffusion of solid benzoic acid from the base of the stirred cell into the aqueous solution. The concentration of a saturated solution of benzoic acid is denoted by [PhCOOH]*. The best-fit line is based on eq 4. The data were obtained at 100 rpm using 150 mL of deionized water at 23 "C.
Figure 5. Rate of change in NO concentration resulting from mass transfer alone, in the absence of chemical reactions. Depletion of NO occurred by diffusion across the gas-liquid interface and across the membrane at the base of the reactor. The initial NO concentration is denoted by [Nolo. The beet-fit lines are based on eq 2.The data were obtained at 100 rpm using 150 mL of 0.01 M phosphate buffer at pH 7.4.
(14)
where [NO]* and [02]* are the aqueous concentrations of NO and 02 in equilibrium with their respective gas-phase concentrations. On the basis of the solubilities of NO and 02 in water (lo), [NO]* was negligible in the oxidation experiments and [02]* was 0.0013~M at 23 OC and 0.0011y M at 37 "C, where y is the 02 mole fraction in the gas phase. The rate constant kl was obtained by minimizing a weighted sum of the squared differences between the measured NO concentrations and the NO concentrations predicted by solving eqs 12and 13. The concentration differenceswere weighted using the inverse of the measured NO concentration. The differential equations were solved numerically using a semi-implicit R u n g e Kuttaalgorithm (19). Only measured NO concentrationsbetween 10% and 100% of the initial NO concentration were used in obtaining the value of kl .
Results Mass Transfer Coefficients. The data from the experiments designed to measure the mass transfer coefficient in the boundary layer at the base of the stirred cell are given in Figure 4, which showsthe change in benzoic acid concentration with time at 23 "C. The slope of the line in Figure 4 yielded a value of (0.24 f 0.01) X 10-4 s-1 for kAAB/V of benzoic acid. Figure 5 shows rates of depletion of NO from deoxygenated solutions, due to the combined effects of transport of NO into the gas phase and across the membrane at the base of the cell. The slopes at the two temperatures yielded values for (k&iG/ V) + (k&/V) of (7.5 f 0.1)X lo4 s-l a t 23 OC and (10.2 f 0.2) X lo4 s-l at 37 "C. The measured and calculated volumetric mass transfer coefficients for NO and 0 2 are summarized in Table 1. At the base of the cell, the aqueous boundary layer contributed somewhat more to the overall mass transfer resistance than did the membrane (i.e., kAAB/V < kMAB/V). In each case the value of k d G / V greatly exceeded that for kBBBIV, indicating that transport of NO and 0 2 across the gas-liquid interface was much more rapid than that from the base of the reactor to the chemiluminescence detector. In other words, the main physical process for removal of NO from the liquid was transport into the gas. Nitric Oxide, Nitrite, and Nitrate Analyses. Figures 6 and 7 show concentrations of NO and NO2- measured during NO oxidation at 37 OC,with initial NO concentra-
\
- - - - - - __ -------
0-
0
5
10
15 Minute3
20
25
30
Figure 6. Change in NO concentration with time during the reaction of NO with 02 at 37 OC. The symbols show data for initial NO concentrations of 10 or 30 pM,at pH 7.4 and 21 % 02 in the gas. The solid curves were calculated using the kinetic model (eqs 12-14). The model predictions without chemical reactions are shown by the dashed curves. Table 1. Volumetric Mess-Tranrfer Coefficients (le8-l) species O C kMAdV kAA,ABIV k d $ V kddV NO 23 0.5 0.4 0.2 7.3 37 0.7 0.6 0.3 9.9 02 23 0.8 0.4 0.3 6.7 37 1.0 0.5 0.3 7.9
tions of 10 or 30 pM,an 0 2 gas composition of 21 %5 ,and a pH of 7.4. Although the concentrations were recorded continuously, the data are depicted as discrete symbols to distinguish them from the curves, which were computed using eqs 12-14. As seen in Figure 6, the solid curves computed by adjusting the rate constant kl (see below) provided excellent fits to the NO concentration data for both initial concentrations and at all times. The dashed curves show the predicted depletion of NO by physical processes alone (primarily transport to the gas phase). Thus, the differences between the corresponding dashed and solid curves reflect NO depletion due to oxidation. It can be seen that the rates of physical and chemical loss of NO in our system were roughly comparable. Using the same parameter values, the kinetic model accurately predicted the changes in NO2- concentrations, as indicated by Figure 7. Figures 8 and 9 illustrate the results obtained at 23 "C, with initial NO concentrations of 10 and 30 p M , 02 gas compositions of 5% and 21 % ,and pH values of 4.9 and 7.4. A single best-fit value of kl at 23 "C represented all
572 Chem. Res. Tonicol., Vol. 7, No. 4, 1994 40 I
I
0
Lewis and Deen 1
I
i
21% 0, pH 7.4
30
Table 2. Rate Constant for NO Oxidation
kl (10s M-2 8-1) 15 OC ref present study Wink et al. (4) Pogrebnaya 1.9 f 0.1 et al. (5) Awad et al. (6) 1.9
22-25
2.1 i 0.4 (23) 2.4 f 0.3 (37) 1.5 i 0.4 (22)a 0.9 i 0.2 (37) 2.2 k 0.1 (25) 2.1 (25)
2.2 (35)
I
0
,
I
10
I
30
20 Minutes
35 0 0
[NO1(PM) 15
,, ,
21% 0, pH 4 9 21% 0, pH 7 4
- _
0
5
10
15 Mmutes
(no r e a c t l o 9
20
30
25
Figure 8. Change in NO concentration with time during the reaction of NO with 02 at 23 "C. The symbols show data for initial NO concentrations of 10 or 30 KM,at pH 4.9 or 7.4, and 5 % or 21 % 02 in the gas. The solid curves were calculated using the kinetic model (eqs 12-14). The model predictions without chemical reactions are shown by the dashed curves. 40 I
25
I
I
0
1
5 % 0 , pH7.4
[NO],= 30pM [NO],= 30pM
15 -
10 [NO],= lOpM
5 -
w 0
*
the NOz- concentrations reported in Figures 7 and 9 are increments above the initial value. The initial NOz- was traced to a small air leak in the flow loop (possibly in the flow cell of the spectrophotometer), which could not be eliminated. The air leak led to a base-line 02concentration estimated at -2 pM (1% of air saturation), which permitted some reaction of NO and 02to occur during the time it took NO to reach ita initial steady-state concentration. The inclusion of an air leak term in eq 13 did not significantly affect the calculated results. Initial NO concentrations were reduced 10% due to the air leak. We were unable to detect any formation of NOS-during NO oxidation. The absence of noticeable amounts of NO3formation, and the consistency of the NO and NOzconcentrations shown by the results in Figures 6-9, indicates that the only NO, species present at significant concentrations were NO and NOz-. Rate Constant. Values of the rate constant kl calculated at 23 and 37 "Cwere (2.1 f 0.4) X lo6and (2.4 f 0.3) X lo6M-z s-l, respectively, as shown in Table 2. As already mentioned, these values were found to be independent of the pH and the initial NO and 02 concentrations. The slight increase in kl at the higher temperature implies a small, positive activation energy, Ea&.= 2.0 kcal/mol. Also shown are values of kl and Ea&. derived from the literature (4-6). Comparisons of the literature values with the present results are discussed below.
Discussion
-
o
1.Ob
-
model
...
-
2.0 -6.5b 2.8
40
Figure 7. Nitrite formation resulting from the reaction of NO with 02 at 37 O C . Nitrite concentrations are expressed as incrementa above the initial value (see text). The symbols show data for initial NO concentrationsof 10 or 30 p M , at pH 7.4 and 21% 0 2 in the gas. The solid curves were calculated using the kinetic model (eqs 12-14).
25
E&.
(kcal/mol)
14 applied to the NOz-concentration (absorbance)data of Wink et al. gave a best-fit value of 2.2. Estimated from reported rate data. Exact temperature ("C)given in parentheses. a Equation
0
35-37 OCC
OCc
,
,
,
10
20
30
40
Mmutes
Figure 9. Nitrite formation resulting from the reaction of NO with 02 at 23 O C . Nitrite concentrations are expressed as incrementsabove the initial value (see text). The symbols show data for initial NO concentrations of 10 or 30 p M , at p H 4.9 or 7.4,and 5 % or 21 % 02 in the gas. The solid curves were calculated using the kinetic model (eqs 12-14). results a t that temperature, as shown by the agreement between the data points and the curves. Overall, the accuracy of the fits of the kinetic model to the data in Figures 6-9 shows that the rate of NO oxidation is secondorder in NO and first-order in 02,as assumed in eqs 1214. In all of the NO oxidation experiments NOz- was found to be present in solution prior to the deliberate introduction of oxygen, reaching concentrations 5-60 pM just before the gas was switched to the 02 mixture. For this reason,
Our determination of the rate constant kl for the reaction of NO with 02 was based on the direct measurement of NO concentrations, together with pseudo-steady-state approximations for NO2 and Nz03. The novel NO sampling and detection system combined with spectrophotometry permitted NO and NOz- to be measured simultaneously, which had not been done in previous studies of NO oxidation kinetics. Kinetics which were second-order in NO and first-order in 02 were shown to be consistent with our concentration data for both NO and NOz-. Our finding of second-order kinetics based on NO depletion is consistent with that of previous studies which monitored only NOz- formation (4-6). The consistency of our NO depletion measurements with our own and with previously reported data on NOz- formation suggests that our method for monitoring NO, based on membrane sampling and chemiluminescence detection, may be more reliable than the electrode method employed by Taha et al. (7).Those authors concluded that NO oxidation was independent of (zero-order in) NO concentration. Aside from possible differences due to NO detection methods, it is noteworthy that Taha et al. did not account for interphase mass transfer, although their experimental apparatus involved both a gas and an aqueous
Oxidation Kinetics of Nitric Oxide
Chem. Res. Toxicol., Vol. 7, No. 4, 1994 573
phase. As already emphasized, quantifying physical losses of NO was very important for the interpretation of the present results. Moreover, certain inconsistencies of the data of Taha et al. with their conclusion of zero-order kinetics have been pointed out recently by Ford et al. (20). There seems little doubt that the rate of NO oxidation in aqueous solution is second-order in NO and first-order in
Concentrations of Nz03 are of particular interest because of the potential importance of N2O3 as a nitrosating agent (2,3). On the basis of the pseudo-steady-state assumptions applied to eqs 6-8, N203 is related to NO and 02 by
(
[Nz031= zk,)[N012[Ozl k4
02.
The present rate constants are compared with those from the literature in Table 2. Wink et al. (4) and Pogrebnaya et al. (5)reported values of ko corresponding to the rate law RNO,-
= -RNo = k ~ [ N 0 1 2 ~ 0 ~ l
(15)
Comparing eq 15 with the rate expressions in eqs 12 and 14 shows that k1 = ko/4;this relation was used to calculate the corresponding values of kl in Table 2. Awad et al. (6) calculated k l using pseudo-steady-state approximations, as done here. The results for kl at 22-25 OC are in excellent agreement. Indeed, the agreement among the four studies is even better than that implied by Table 2. On the basis of the NO2- absorbance data at 22 "C reported by Wink et al. (4), we calculate a best-fit value for kl of 2.2 X 106 M-2 s-l, indistinguishable from the other mean values shown. The results for kl at 35-37 "C likewise are in generally good agreement; this includes the value of 2.6 X 106 M-2 s-1 (not shown) obtained by extrapolating the results of Pogrebnaya et al. (5) to 37 "C. Except for the negative value implied by the results of Wink et al., there is no significant difference among the activation energies. Reactions omitted from the present kinetic model, but which may be relevant under other conditions (e.g., low pH), include
2N02 + HzO
+ NO; + 2H+ HNO, a NO; + H+ 2HN0, a N,O, + H 2 0 +
NO;
2N0,
G
N204
(16) (17) (18) (19)
The reaction in eq 16, in which NO2 (or Nz04) combines with HzO to form equimolar amounts of NOz- and NO3(211,was neglected because no noticeable NO3- formation was observed. This is consistent with previous observations (6,17). Nitrous acid (HN02) has a pK, of 3.4 (22), much lower than the most acidic pH studied here. Accordingly, HNOz was neglected as a potential source or sink for nitrite ion (eq 17) or N2O3 (eq 18; 23). Rather, the product of the hydrolysis of N@3 was considered to be nitrite ion (eq 8). In addition, N204 (eq 19)is expected to make a negligible contribution to the nitrogen balance (24). The pseudo-steady-state assumptions used for NO2 and N203 depend on these species being present only in trace amounts. These assumptions were justified by the overall nitrogen balances, which showed that essentially all of the nitrogen added initially as NO could be accounted for by the sum of the NO and NOz- remaining in the reactor, together with the amount of NO calculated to have left the system (using the measured mass transfer coefficients). Previous reports (4-7) did not utilize nitrogen balances to verify that NO and NO2- were the only major species present.
Similarly, the concentration of NO2 is predicted to be given by
[NO,] =
(t)(2)(
1 + ~ ) r N O 1 ~ O z l (21)
Wink et al. (4) observed an intermediate species in the NO oxidation reaction that had a second-order dependence on NO and a first-order dependence on 02.Equation 20 suggests that N2O3 is a strong candidate for this intermediate species, despite certain arguments to the contrary (4). Reported values Of k4 are 530 ~ ~ ( 2and 5 )2000 s-l(26) with a recommended value of 1000 ~ ~ ( 1 8The ) . value for the equilibrium constant k3/k2 is 3.3 X M at 23 "C (18). With a value for k2 of 1.1 X lo9M-l s-l a t 20 OC (18), this gives k3 = 3.7 X lo4 s-1 and k4/k3 = 0.03. Equations 20 and 21 are based on the reactions represented by eqs 6-8; whenever additional reactions involving N2O3 or NO2 exist, these results must be modified. In summary, the kinetics of the reaction of NO and 0 2 in water were studied using a novel reactor configuration, which made it possible to quantitate both NO and NOzin real time. The results confirmed that the rate of the oxidation reaction is second-order in NO, and the measurement of both of the dominant nitrogen species provided verification of pseudo-steady-state assumptions applied to NO2 and Nz03. The importance of considering mass transfer effects in this or other open systems was emphasized. The kinetic model presented should be helpful in predicting the concentrations of NO and of trace intermediates such as NzO3 and should thereby aid in understanding the biological roles of NO both as a messenger and as a cytotoxic or mutagenic agent.
Acknowledgment. The authors are grateful for the technical assistance provided by Shelly Sakiyama. This work was supported by a grant from the National Cancer Institute (Pol-CA26731).
References Moncada, S., Palmer, R. M. J., and Higgs, E. A. (1991) Nitric oxide: physiology, pathophysiology, and pharmacology.Pharmacol. Rev. 43,109-142. Tannenbaum, S. R.,Tamir, S., de Rojas-Walker,T., and Wishnok, J. S.(1993) DNA damage and cytotoxicity by nitric oxide. In Roc. ACS Symposium on N-Nitroso Compounds, Washington, D. C. (Loeppky, R. A., and Michejda, C. J., Eds.) (in press). Beckman, J. S., and Crow, J. P. (1993) Pathological implications of nitric oxide, superoxide and peroxynitrite formation. Biochem. SOC. Trans. 21,330-334. Wink, D.A.,Darbyshire, J. F., Nims, R. W., Saavedra, J. E., and Ford, P. C. (1993) Reactions of the bioregulatory agent nitric oxide in oxygenated aqueous media: determination of the kinetics for oxidation and nitrosation by intermediates genereted in the NO/Oz reaction. Chem. Res. Toxicol. 6, 23-27. Pogrebnaya, V. L., Usov, A. P., Baranov, A. V., Nesterenko, A. I., and Bez'yazychnyi,P. I. (1975)Liquid-phase oxidation of nitric oxide by oxygen. Zh. Rikl. Khim. 48, 954-958. Awad, H.H.and Stanbury, D.M. (1993) Autoxidation of NO in aqueous solution. Int. J . Chem. Kinet. 26, 375-381. Taha, Z.,Kiechle, F., and Malinski, T. (1992) Oxidation of nitric oxide by oxygen in biological systems monitored by porphyrinic
674 Chem. Res. Toxicol., Vol. 7, No. 4, 1994 sensor. Biochem. Biophys. Res. Commun. 188,734-739. (8) Lewis, R. S., Deen, W. M., Tannenbaum, S. R., and Wishnok, J. S. (1992)Membrane mass spectrometer inlet for quantitation of nitric oxide. Biol. Mass Spectrom. 22,45-52. (9)Green, L. C., Wagner, D. A., Glogowski, J., Skipper, P. L., Wishnok, J. S., and Tannenbaum, S. R. (1982)Analysis of nitrate, nitrite, and ISNO*- in biological fluids. Anal. Biochem. 126,131-138. (10) Lange, N. A., Ed. (1967)Lange’s Handbook of Chemistry, rev. 10th ed., p 1101,McGraw-Hill, New York. (11)Robb, W. L. (1968)Thin silicone membranes-their permeation properties and some applications. Ann. N.Y. Acad. Sci. 146,119137. (12) Windholz, M. W., Budavari, S., Blumetti, R. F., and Otterbein, E. S., Eds. (1983)The Merck Index, 10th ed., p 1093,Merck & Co., Rahway, NJ. (13)Sherwood, T. K., Pigford, R. L, and Wilke, C. R. (1975)Mass ?‘ram-fer, pp 214-224, McGraw-Hill, New York. (14)Wise, D. L, and Houghton, G.(1968)Diffusion coefficients of neon, krypton, xenon, carbon monoxide, and nitric oxide in water at 10-60 C. Chem. Eng. Sci. 23, 1211-1216. (15) Chang, S. Y. (1949) S. M. Thesis, Massachusetts Institute of Technology, Cambridge, MA. (16) Wilke,C. R.,andChang,P. (1955)Correlationofdiffusioncoefficients in dilute solutions. AIChE J. 2, 264-270. (17) Ignarro, L.J., Fukuto, J. M., Griscavage, J. M., Rogers, N. E., and B y ” R. E. (1993)Oxidation of nitric oxide in aqueous solution to nitrite but not nitrate: comparison with enzymaticallyformed nitric
Lewis and Deen oxide from Garginhe. Proc. Natl. Acad. Sci. U.S.A. 90,8103-8107. (18) Schwartz, S.E. (1983)Trace atmospheric constituents properties, transformations, and fates. In Advances in Environmental Science and Technology (Nriagu, J. O., Ed.) p 91,John Wiley & Sons, New York. (19) Michelsen, M. L. (1976)An efficient general purpose method for the integration of stiff ordinary differential equations. AIChE J. 22, 594-597. (20) Ford, P.C., Wink, D. A., and Stanbury, D. M. (1993)Autoxidation kinetics of aqueous nitric oxide. FEBS Lett. 326, 1-3. (21)Abel, E., and Schmid, H. (1928)Kineticsof nitrous acid. 111.Kinetics of nitrous acid decomposition. 2.Phys. Chem. 134,27*300. (22) Tummavuori, J. and Lumme, P. (1968)Protolysia of nitrous acid in aqueous sodium nitrate and sodium nitrite solutions at different temperatures. Acta Chem. Scand. 22,2003-2011. (23) Bunton, C. A., Llewellyn, D. R., and Stedman, G. (1959)Oxygen exchange between nitrous acid and water. J. Chem. SOC.1959,568573. (24) Licht, W. R.,Tannenbaum, S. R., and Deen, W. M. (1988)Use of ascorbic acid to inhibit nitrosation: kinetic and maee transfer considerations for an in vitro system. Carcinogenesia 9, 365-372. (25) Gritzel, M.,Henglein, A., Little, J., and Beck, G. (1969) Pulse radiolytic investigation of some elementary processes of oxidation and reduction of nitrite ion. Ber. Bunaenges. 74,488-492. (26) Treinin, A., and Hayon, E. (1970)Absorption spectra and reaction kinetics of N02,N203,and N20, in aqueous solution. J.Am. Chem. SOC. 92,5821-5828.
