The Initial Decomposition Rate of Nitrogen Pentoxide Citation White, Ernest Croel (1925) The Initial Decomposition Rate of Nitrogen Pentoxide. Dissertation (Ph.D.), California Institute of Technology. doi:10.7907/PVPC-1N90. https://resolver.caltech.edu/CaltechETD:etd-07162004-145619 Abstract No Abstract. Item Type: Thesis (Dissertation (Ph.D.)) Subject Keywords: Chemistry Degree Grantor: California Institute of Technology Division: Chemistry and Chemical Engineering Major Option: Chemistry Thesis Availability: Public (worldwide access) Research Advisor(s): Tolman, Richard Chace Thesis Committee: Unknown, Unknown Defense Date: 1 January 1925 Record Number: CaltechETD:etd-07162004-145619 Persistent URL: https://resolver.caltech.edu/CaltechETD:etd-07162004-145619 DOI: 10.7907/PVPC-1N90 Default Usage Policy: No commercial reproduction, distribution, display or performance rights in this work are provided. ID Code: 2904 Collection: CaltechTHESIS Deposited By: Imported from ETD-db Deposited On: 22 Jul 2004 Last Modified: 03 Oct 2019 23:21 Thesis Files Preview PDF (White_ec_1925.pdf) - Final Version See Usage Policy. 1MB Repository Staff Only: item control page

THE INITIAL DECOMPOSITION RATE OF NITROGEN PENTOXIDE Thesis by © Ernest C. White In partial fulfillment of the requirements for the degree of Doctor of Philosophy. California Institute of Technology Pasadena, California. 1924 THE INITIAL DECOMPOSITION RATE OF NITROGEN PENTOXIDE. -Recent developments in chemistry afford fev problems more interesting than that of reaction rates. The science of thermodynamics has become a most valuable tool in that it enables the chemist to distinguish between reactions that are possible and those that are impossible, as well as to determine in advance the optimum conditions under which will cperate any process involving a system at equilibrivwn. Unfortunately it doas not explain why certain theoretically possible reactions actually show little or no tendency to take place, nor can it predict how fast a reaction will proceed. The solution of these problems of reaction velocity cannot be accomplished by the methods of thermodyneanlce, but will necessitate the deeper insight provided by the methods of statistical mechanics, and it is hoped that the work on the initial rate of decomposition of nitrogen pentoxide, described in this article, may be of usa in furnishing experimental material which will contribute to the ultimate solution of the problem. The decomposition of Nz05 is a reaction which 1s peculiarly well adapted for study in connection with theories as to reaction velocities and their temperature coefficients. It is perhaps the only known unimolecular gas reaction which is not appreciably catalyzed by the walls of the containing vessel and it proceeds at moderate temperature very conveniently for measurement. The chief disadvantage, its corrosive action on mercury, rubber and other materials ordinarily used in handling gases, was obviated by means of an all-glass pressure (1) measuring device employed by Daniels and Johnston,’ who demonstrated the true unimolecular character of the reaction over the temperature (2) found that the decomposition range from 0° to 65°. They further was accelerated by blue Light, but only in the presence of the dioxide, NO.. In the course of some later work by Daniels, Wulf and Kerrer 2?) on the oxidation of nitrogen tetroxide, 4a phenomenon wes observed which raised some question as to the behavior of the pentoxide when entirely free from its decomposition products. After a quantity of NOp + MoO, had been oxidized by a quantity of ozone just sufficient to give a colorless mixture, it was expected that the brown color of NOp would reappear in a very short time due to the decomposition of the NoQ5 thus formed. From a yet unkmowm cause the mixture remained Colorless for a great length of time even after being raised to temperatures where the ordinary decomposition is extremely rapid. Attempts to explain this curious fact brought forth the suggestion that the decomposition was retarded by the complete absence of NO2 which might be a catalyst whose presence was necessary. No evidence to support this hypothesis had been add:ced in the previous measure- tents of the reaction velocity, for in every case the rate appeared (1) Daniels ani Jomnston, J. 4. C. S. 43, 53 (1921). (2) Daniels end Johnston, J. A. C. 8. 43, 72 (1921). (3) Teniels, Wolf and Xerrer, J.A.0.S. 44, 2402 (1922). to ta independent of thse concentration of the dioxide. Nevertheless . the photochemical acceleration, already mentioned, and in addition gome results obtained by L Lueck! 4) in a study of the rection in carton tetrechloride solution offered support to the argument for autocatalysia. The present investigation was undertaken for the purpose of testing the possibility that nitrogen dioxide may act as a catalyst in the initial stages of the gaseous reaction t>05 —> MoO, + $ 0p \t 2 N0o Ags will be seen in the sequel, this study has shown that the reaction retains ite unimolecular character even at the lowest concentrations of dioxide that it has been possible to measure. It is felt that this discovery considerably increases the theoretical interest and importance of this reaction. AT RU mame el bk el = ow Although the glass manometer used in the previous investizations by Daniele and Jehnston 4s capable of measuring moderate pressures with considerable accuracy, it was necessary for the purpeses of this experiment to devise @ more sensitive method which would give a ueacure of the N05 independently of the ‘otal pressure. The colorimetric principle offered a very simple means of estimating concentrations of this reddish trown gas, end accordingly there was (4) Lueck, J. A. C. S. 44, 757 (1922). } OT Y suoijoas-sso. ) ” TTTAETHIANTANY = i a IN ait iit | it (Il i | rH ee In| hil ID NTALL Constructed an apparatus which was essentially « gas colorimeter. Fig. 1 shows the plen and three vertical cross sections of the colorimeter. The reaction chesber consisted of the tubes a end b, S83 and 8.2 rm. leng respectively, connected by the side arm oc. Their ends were formed of plane microscope slide glass about 1 mm. thick, fused on by a special process. These tubes were jacketed as shown and kept at censtent temperature by circulation of wasar from and to the thermostat d, which was provided with a screw pimp eo and she conventional epparatus for automatic ta:perature regulation (not shown in the diagram). The tubes 1, 2, 3, 4, 5, 6, 7 will be referred to as the "one-phase standards." Hach had plane glass ends and a side arm connected to the manifold f, thus forming a single gas chamber with seven branches, whose inside lengths ranged from 37 mu. to 259 ma. by successive inltiples of 37. This system was evacuated, then charzed with a measured quantity of NOp and sealed off. It was securely Pastened to a cradle-shared metal frame capable of being rotated about the axis g by means of the lever h and arrested in the desired position by meane of a spring-latch. ‘Thus each of the tubes could be brought into the horizontal plane of the reaction tubes and parallel to them. The open-ended tubes 1, Jj and k wers of the same dianeter (15 mm.) as the standerd and the reaction tubes, and served to equalize reflections along the paths of the three light beams, which entered through the circular apertures 1, m, n, after diffusing through the opal-glass plate p, the source being a filament lasp. The shutter o excluded the light during intervals between reazings. After pessing through the tubes the light beams entered the mirror-box g through glass-windowed apertures 6 mm. square and were reflected at right angles by the sharp-edged speculum metal mirrors Yr, &, t, adjusted in such a position that the images of the Square apertures were contiguous when viewed through the eye~plece u and the lens v focused upon them. Green or blue color screens were sometimes introduced to aid the eye in obtaining a sensitive match. Cross section 3 , Fig. 1, illustrates the "two-phase standard", consisting of a short, plene-ended tube w connected by a narrow tube to the bulb x containing liquid NOp. This system was evacuated at a low temperature before being sealed, so as to remove all foreign gases. The pressure of NOp in the compartment w could be varied between wide limits, being calculable from the temperature of the liquid and that of the vapor, with the aid of the experimental vapor pressure curve and Schreber's equation for the equilibrium between NO. and NoOy. ‘This "two-phase standard" could be inserted in the middle, as illustrated in the drawing, or at the left. In the former position it was used for comparison with tube a or tube b, and in the latter position for comparison with the "one-phase standards" in order to check their titre, in which cese it was necessary to remove the tube b or to compensate lts absorption of light by interposing extra glass slides in the middle path. V jt i“ FIG. 2 a San NG, introduction Mixing device Colorimeter SA A Se pe fi \ TT ee DA\\ pee \ AMS WR RAN Ozone —_ unit Sy a et gt in) WRSRNEE ES BR ee nhansdeneianlan fy FaeRaReaE Ee wl oui : OF low meter Ozonizer Cross sections A and 0 need little further explanation. They show the inlet y and the outlet g of the reaction chamber, and mors Clearly illustrate the arrangement of the "one-phase standards", the water circulating system, and the light-tight cabinet with removable lid, Initially several variations were tried in the method of preparing the N50s5 free from NOs, and of introducing it into the reaction chamber. As all of these gave similar results, it was decided to adopt the most convenient procedure, which will be described here with the aid of Fig. 2. ‘This final method as described was used in all work later than Run No. 4, . Oxygen from a commercial cylinder was ariea in contact with P50, and passed into the train at the left of the figure, its flow being regulated by a blow-off immersed in liquid petrolatum and registered by a capillary flow-meter employing the same liquid. A Siemens ozonizer ylelded a mixture of about 4% ozone by volume when operated Steadily on 20,000 volts at room temperature and a flow of 150 c.c. per minute. This oxygen-ozone mixturs passed thence through a device for anelvzging it by the pressure chenge attending the subsequent decomposition of the ozone on heating, the pressure being measured with ccncentrated sulphuric acid in a capillary manometer, an arm of which was sealed to a closed buffer tube containing dry air at a known pressure and equal in volume to the analysis chamber, namely that portion of the train between the two constrictions @ and b. The analytical procedure consisted simply in fusing eat the constrictions and taking readings of tamperature and of pressure before and after heating for about four hours at 150° in the oil bath, which was heated electrically and stirred by bubbling air from a tube not shown in the sketch. This analysis was not performed, of course, wmtil — after the desired mixture of gases had been introduced into the colorimeter. The N5O5 was produced by forcing Mp and 50,4 into the gas stream under its own vapor presaure from the liquid in the jacketed bulb, as illustrated in the figure. By carefully controlling the temperature of the water circulating through the Jacket, the liquid oxides could be vaporized at a steady rate, which could be varied in different runs 30 as to give relatively high or low concentrations of the NoO5 that formed upon contact with the stream of ozone. P05 driers were used as shown, and the resulting mixture of 205, Og and O3 passed thence into a mixing device, whare stirring was accomplished by means of a glase propeller, with hollow blades filled with iron filings, driven by rotating a horseshoe magnet, After leaving this chamber the gases passed through a tube having a constriction and a glass check-valve oc, beyond which a capillary lad to the reaction chamber of the colorimeter. The valve and the capillary served to retard diffusion into the colorimeter of the slight quantity of zas decomposed in the process of fusing off the constricted tube. The voluns of the entire train was approximately a liter, and usually the zas flow was maintained steadily at about 150 c.c. per minute for half en hour, by which time the reaction chamber contained a fairly homogeneous mixture antirely free from M0. and of a distinctly bluish color from the excess of ozone. The partial pressure of the Nz05 was determined from subsequent colorimetric analysis of the NOs present after the complete decomposition, as will be described in a later section. In all except the first four runs the partial pressure of the ozone was also estimated from the combined data. The temperature of the water bath having been adjusted to the desired point and kept constant within about 0.02°, the system was now ready for the main experiment and the reaction chamber was quickly sealed off by fusion of the inlet and outlet tubes. ‘The exact time of this operation was recorded, but in the calculations five minutes was allowed for the interval during which the gas had been traversing the mixing bulb. In the later rune the procedure from the time of sealing off the reaction chamuber was to take brief observations of the color in the mein tubs a (Fig. 1), by comparison with a colorless blank at frequent intervals. As the excess ozone was being converted to oxygen by the simultaneous reactions NoOs = Nyy + #0 and NoOy + O3 = N505 + Op the blue color gradually faded out. The first appearance of yellow therefore marked the exhaustion of the ozone and ths beginning of the unreterded decomposition of the pentoxide. Since a pressure of NOo as low as 0.01 um. could be detected easily, an effort was made to get a precise observation of the time at thie point of "zero" concentration. Next the “one-phase standard" tube No. 1 was brought into position and. rapid observations were taken at intervals of increasing frequency until the intensity of its color was matched by that of the reaction mixture. After the intensity had exceeded that of standard No. 1, tha latter was replaced by No. 2 and similar observations were taken until this gaa matehed, and so on tntil all, or aa ‘nany as desired, of the stendaris had been matched and the tines noted. Tue to the linited sensitivity of the eye, there was usually a period of from one to ten or more minutes - depending mainly on the rate at which M0. wa: being formed = during which no difference could be distinguished between the intensities of the two fields. The middle point of such a period was selected as the "time" of equal intensity, and was rrobably correct to within a fifth of the length of the period, corresponding usually to about 5% of the interval between the match of one standard and that of the next highest. In many of the runs 4 later point on the time- pressure curve was obtaining by matching against the "two-phase" tube, Finally the temperature of the reaction mixture was raised to about 4o°, vinere the decomposition is 99% completed in five hours, and upon cooling again the MNO was analyzed by matching cne or the other of the reaction tubes a or b against the two-phase atandard. This and the othsr analyses will be the special subject of a later section. oO 10 Preparation of Materials. The source of the NoOs had no noticeable effect on the results. In the earlier rana fuming nitric acid was dehydrated by forming a paste with phosphorus pentoxide, from which the gases were distilled off in a current of ozone and dried over powdered phosphorus pentoxide in a long tude. Pure white crystale of NjO5 were condensed in a U-tube at about -S50° and these were later vaporized and ewept with an excess of ozone into tha mixing bulb, whence the mixture was later transferred to the previously evacuated reaction chamber by means of @ glags connection like the illustration in Fig. 3. This was opened by a smart tap from beneath, which knocked the glass bead against the thin, file scratched tip, and was closed by fusing the constriction. ‘The same device was used in place of a stopcock in other operations as well, In all the later runs, as described above, the sources of the NO, and No0y, was the liquid, a supply of which waa prepared by the destructive distillation of lead nitrate crystals (Kehlbaum), water vapor being carefully removed by passing through three tubes of phosphorus pentoxide before condensing in the jacketed container shown in Fig. 2. When not in use the container was sealed by fusing the constricted opening at the top, and to open it the tip was melted in a torch flare while the pressure inside was slightly above atmospheric. Connection was made to the train by inserting into the beveled tube, which had been ground to fit snugly over the neck of the container. The ozone was formed from cormercial cxygen (Linde), containing no impurities recognized as affecting the decomposition of NpQs, with the possible exception of water vapor, for which drying tubes ware provided. That the ozone was reasonably stable was proven by teste of the slow rate at which it decomposed spontaneously at room temperature. The concentrations in various runs ranged from 2.5% to 5.2% by volume. Analyses and Computations. The methods of analysis of the NO. were based on Zeer's Law that equal amounts of light are absorbed by passage through equal quantities of a substance independently of the concentration. Although deviations from this law are known to occur in gases (>) they are not of such a nature as to introduce serious errors in this work. Analysis of NOp in Reaction Tube during Run. - ‘The gradually increasing partial pressure of NOp in the reaction tube was analyzed by matching against the various "one-phase standards." The pressure of NOg first introduced into this standard system wag 11.8 mm. Hg at 22.0%, so that the shortest tube, No. 1, absorbed the sae auount of light as would be absorbed by 0.495 mm. NOp pressure in the long reaction tube at the same tempereture, the ratio of the lengthe being 37:688. Similarly the pressure equivalent of No. 2 was 0.990 mm. and so on. After Bun No. 7 the standard was evacuated and refilled with a smaller quantity (6.45 mm. at 21.2°) of gas, sc that the lowest pressure measurable was reduced to 0.209 mm. at the same eee a eee Renn te ee nae (5) Re'% ood, "Physical Optics", Second Ed., p. 44K; Macmillan 1921. 12 temperature. The temperature of the "one-phase standar’" was always that of the room while that of the reaction tube usvally differed fro it, hence corrections were calculated and applied for the effacts of variation in temperature both on the pressure-volume product and on the equilibrium between NOp and Noy. Calibration of the One-phase Standard.- The two-phase tube constituted the ultimate standard, and the calibration of the one- phase standard is here illustrated by reproducing an actual example. Observations Tate: May 8, 192u. Tube No. 7 (length 259 um.) matched two-phase tube (length 28.5 mm.) Temperature of liquid in two-phase tube: -16° r " vapor . ft " fos +20.7° Teaperature of one-phase system: 21.2° Derived Data . (6) Total pressure in two-phase tube; Pyo, + Py, = 102 mi. 2 (7) Partial pressure of NoCy; Py,0y = 0,014ug PHO. (6) 108 mm. is the vapor pressure at -16°, obtained by interpolation from the data given in Tables Ann. Int., 2, 113 (1911). (7) 0.01442 is the value of Ry at °0.7° obtained from Schreber's equation; log K = log Tyne 2ees PN. 24, 651, 1897). - log T- 9.132% (2. physik. Chen, 13 Computations From the above equations, Partial pressure of NOs in two-phase tube; -1 +Vi + (4 x 0.01448 x 108) Pro, * 2x 0,014KE = 5&5 ne. 52.5 x 28.5 x (273 + 21.2) N _ . t = wt. Pressure NO. in one-phase system; Pro, 259 x (273 + 26-7) Determination of Final Pressure in Reaction Tube.- The analysie of the final pressure of NOp after the corpletion of a run waa usually obtained in a similer manner by matching the color in the reaction tube against the two-phase standard. For the higher pressures, the shorter tube b was matched against this standard. It was sometimes found convenient to wake use of the one-phase standard also in these analyses. accuracy of the Analysea.,- It is souwewhat difficalt to make an exact estimate of the accuracy of the various analyses. In the case of the comparison between the reaction tube and the various one-phase standerd tubes during the run, it was found that the apparent match persisted throughout a period in which the partial pressure of 05 mst have changad dy about £%. Since the time of the match was taken as the middle of this period, it is believed that Yobservaticnal™ error in these concentrations was less than 2%, In the use of the two-phase standard, the comparison was made by Ly Bijuating the temperature of the liquid or of the vapor in the two-phase tube until a match was secured. It was generally found that a change of about 1° in either temperature was sufficient to destroy the match. Calculation showed thet en error of 1° in either of these temperatures would correspond be an error of about 3% in the result. Since the temperatures were probably known to considerably batter than this, it is thought that the “observational” error in the concentrations thus measured was certainly less than 5%. In addition to the above "observational" errors, "systematic" errors were also undoubtedly present, owing to deviations from Boer's Law, and to incorrectness in the figures taken for the degree of association of the NOp and for the vapor pressure of the NO. in the two-phase standard. Effect on Computations.- It is evident from the foregoing that the relative values of the partial pressure of NOz obtained in the early stages of the run are fairly accurate, while the absolute values of these pressures and the values of the final pressure are more uncertain. Fortunately a greater degree of accuracy in these latter figures was nct neceasary for the main purpose of the experiment, which was to compere ths retes in successive stages cf the early part of the reaction, as represented by the familiar equation — = to - by ns ~ k where P, and Pp are pressures of the disappearing constituent No0s 15 at tines t, and t, respectively, and k is constant for a given temperature. If substitutions in this equation gave a constant values of k for every stage of the reaction, its unimolecular charecter would be substsntiated. In this reaction we obviously may set Pp _. 2Peon- 2 Pi Peo 7 Pi = = Po . 2Po- 2 Pe Poo = Pe where Py Pe Snd peo are pressures of the appearing product NOo, provided we take into account the equilibrium 2 NOp <= Mo0y and include the NpCy as equivalent N05 in the p terms. Thus, if the pressure of NOp were 0.552 moa., as indicated by matching against the one-phase etandard No. 2, the corresponding value of p ina 2° run would be (taking 0.01523 as the value of Kp at 2°, of. Note 7); _ P* Pyo, + 2Pr 0, = Pyo, + 2 (0.01523 Pace) = 0.561 rm. It will be observed that, for values of the final pressure relatively large in comparison with the pressures in early stages, a moderate error in the determination of p,. would affect the magnitude of k to a slight extent, but would not appreciably affect its constancy throughout the early stages of the run. The game is true for an error in the callbration of the one-phase standard system, for this weuld but slightly influence the relative values of p,, De ...p7, or of the k's. Typical Run.- Eelow are srrenged the summarized data and computations for a typical run (No. 9). Time is in minutes from the initial formation of the No0s5, and pressure is in mma. Hg under standard conditions. Run No. 9. Summary. Date: June 10, 192u. Tewperature of reaction: 20.0° Standard Time, Matched. | -~ Pressure. Period [Point | No. | Temp. Prog” + 2 Py, 0¢2) = PB 230 =| ELank 8) 0) 267 = 271 | 269 1 21.7 | 0.269 + 2(0.01523)(0.269)" = 0.271 303 - 309 | 306 349 - 357 | 353 389 - 399 | 394 was = uo | Usk “01.3 | 0.827 + 2(0.01523)(0.537)° = 0.545 21.5 | 0.207 + 2(6.01523)(0.807)° = 0.227 21.5 | 1.075 + 2(0.01523)(1.075)* = 1.112 LS: Cn en <) So) 21.3 | 1.339 + 2(0.01523) (2.339) = 1.39 \ phasa j{v 21.5 7.122 End of run| © | 7¥O- fF 18.5 | 6.019 + 2(0.01523) (6.019)° Unimolecular Tests. 2. ™ Interval — nae log a = k te to ty Be oH tog ES Oa = 0.00099 ty to t, Pees 5 log aco 2 ae = 0.00110 to to ts FeO 30 log aE = aa = 0,00093 ts to ty 3 353 log iit = 0.00112 ty to ts wits on log coe 1355 = 0.00120 t, to ts Wir> 365 tog D5 1355 = 0,00109 (8) These values are the vartial pressures of NOp at 20° in the reaction tube, as determined by the match. (9) These values are calculated with the help of Schreber's equation. LT Results. ——_ Runs tiade at 20°.- Of the fourteen runs at 20°, two were unsvnecessful or account of extremely low initial pressures of NOs, requiring uninterrupted observations over a prohibitive length of time. <A third run was discarded because NO> was found to be present at the start. Fun Mo. 5 was the earliest in which un attempt was made to note the time of the first appearance of N05 and to estimate the concentration of ozone. rior runs differed also in the method of preparing and introdwcing the N05, a3 noted above. The sleven completed runs at 20° are shown diazramaticully in Table 1 below. The short horizontal lines indicate values of Fr, as plotted on the scale in the column at the extreme left, while the intervening figures represent the value of k x 10° Calculated for the interval. The rune are arranged in columme in the order of diminishing values of po, which are shown at the botton together with the mean values of k x 10° for the entire interval between the second and the last observations. The low values of k for the first interval in Puns Nos. 5, 6 and 13 may be accounted for by a reasonable assumption. In addition to the ordinary experimental errors, thare covldi be a diatinct type affecting the earliest stage exclusivel:, e+ nanely a lack of homogeneity in the gas mixture; for it is quite conceivable that the ozone and the pentoxide wera not uniformly distributed throughout the long narrow tube, and consequently that all of the pentoxide did not begin decomposing freely at the same instant. In the other runs it would appear thet the mixing has been mors thorough. Table i, k x 10° | P No. 4) No. LY) Mo.13} No. 7/No.16|No. 5/No.11/Mo. 2]io. SIN. S|No. 6 fe) 102 | 70 35 T 103 99 96 72 o.5-|—| 204 | 208] gy | 88 | — | 226) __] nso | nue | 100 ait 106 101 gi | 100 | 109 33 193 | 105 1.07 ——| 107] 92}; —| 98] —-]| 9 | —| 112 | 1¢c0 98 | 100 | 10g | 112 cs) 97 | 120 1.57) — 117 3 — | 114 | —— — 113 107 | 192 | | OS 85 __ 204, |i] 2] | __ 8&7 113 gl 2.5- | — — 9g 3.0- 90 | 1e1 101 | 99 — 102 | —— 3.54 117 92 4.07 —_ — 4Wo5- 91 5.97 5.57 6.04 Mean k} 99 | 104 | 102] 102/ 102] 90] 105 | 99] 109} 105 | 112 Poo 110 42] 37.7] 35 | 19.7] 15.4) o.4O] 8.501 7.122) 5.025) 2.652 Lé 19 Varied Runs.- Since in all of these experiuents the analytical method required the admission of light at frequent intervals, ‘he reaction was exposed intermittently to photochamical acceleration, though only for @ gmall fraction of the time, probably never exceeding one-tenth. In the majority of the runs the intensity of illumination, after the light has passed through the opal glass, was approximately 570 ergs per sescond per om? at the aperture, vhose area was 2 one, 60 that any appreciasle photochemical action would have been surprising. However, to make certain that the effect was negligible, two runs, Nos. 20 and 21, were carried out at 20° with continuous illumination of the same inteneity. The results, shown in Table 2, indicate that the photochemical effect was scarcely appreciable. Table 2 also gontaing the results of three rungs at 25°, one at 35° and one at 40°, fn exact observation of "zero" pressure wag not taken in Run No. 1, end wes missed accidentally in Nos. 12 and 20, as was also the first match in No. 20; but approximations arrived at from neighboring observations indicated no abnormal behavior. Table 2. x x10 ace | P Continuous 25° 35° 4Oe Lllumination No.21 {| No.2O0} No. 1 | No.17/ No.1] No.19] No.12 ° 92 . 172 | 2ky 907 0.5 - 103 192 208 866 1540 101 $5 2hy 187 | 2@ll 665 | 136¢ 10-— 94 86 —. 226 296 741 1340 122 | 114 2h3 235 888 | 1860 15-7 105 | 120 235 1530 114 | agg | 238 226 | gue | —— 204 —— | J | 22 |) 3.04 106 22h | 197 836 3.5 a1 4,5 Mean k 106 108 233 202 222 837 Lugo Poo 12.78 18.548 | 58.1 35.5 | 4.965 | 5.895] 4.950 el Retariation by Ozong.- The phenomenon previously mentioned as having been ebserved by Daniele, Wulf and Karrer ©? namely the dslayed reappearance of 10, after the elapse of tina calculated for the disintegration of the ozone, was not encountered in the fifteen experiments of the present work in which the ozone concentration was analyzed, with the possible exception of one doubtful case (Run No. 13). In every other instance the calculated time excesied the observed time several fold, as shown by the results in Table 3. The value Co, in the third coluwm is the molal fraction of ozone befora mixing with M50,, a3 obtained from the analysis of the oxygen-ozone mixture by the method previously described. The fourth colwm’ gives the pressure of NzOs5 in the reaction chamber, namely one-half the value of pg determined after the run. In the fifth colum POs is the excess pressure of ozone immediately after mixing, and is derived from the corresponding values of Co, and PN205 by means of the relationship | Po; = Co,(P - pylog) - PiOs here P is the total, 1.4. approximately barometric, pressure at the moment of mixing. This is equivalent to stating that the resultant pressure of ozone is its pressure on dilution, diminished by one equivalent of ozone for each aquivalert of pentoxide formed, and hence assures that the reaction is that found by Wulf, Daniels and Farrer, NoQ, + 03 = NpO5 + Oo. The seme assumption ig also the basis on which were computed the values in column six, using the average observed rate kn, ag follows: Pp t (eale) = —~ = _-almutes. en" P05 Table 3. Run Mo. Temp. Co, Pr.05 Po, t (calc. ) t (obs.} 5 20 0,039 8.2 20.2 2,370 105 6 20 0.028 1.33 1.3 14,100 399 7 x 0,038 17.5 9.6 532 14 g rs 0.025 2.52 15.9 6,130 300 o 20 0.038 3.56 a4. 2 5,610 235 li 20 0.037 4.700 22.5) B50 125 13 20 0.0B 18.8 o* o- 32 14 20 0.052 21.0 16.4 758 6C 15 20 C.050 1.04 36.6 34,200 510 16 20 0.048 9. 85 au. 2, uko 125 17 25 0.050 17.8 18.6 5C2 50 16 0.047 2.48 32.4 6,000 215 19 35 - G.050 2.95 19.2 794 75 20 20 0.040 4.27 283.6 5,360 200 21 20 0.040 5.29 23.9 3,940 183 * The actual calculation gives a nezative valve (see text). 23 Discussion of Results. If the initial rate of the decomzosition were other than unimolecular, we should expect that the equation 1 Poo = 2p = tr-t “po - pt - * would give values of k differing in the initial stages from those in the later stages. In case the reaction had proved to be autocatalytic at the start and of the first order later on, as originally susvected, the velocity would have presuuably increased for a while as a function of the increasing concentration of the catalytic product, until a “critical concentration" of the catalyst has been established, after which it would have no longer been dependent on this variable. In such a case the value of k calcvlated in the unimolecular equation above would be low for the early stages of the decomposition, especially for the first interval examined, and would eventually reach & constant value coinciding with that found in. the previous investigations of Daniels and Johnston, “t? In every experiment of tha present research, the value of k proved to be reasonably constant throughout, though in a few instances thers were moderate deviations for which a plausible explanation was advanced ebove. Hence the results supply no confirmation of the _hypothesis that this reaction is autocatalytic. Tt is of some interest to note that in experiments with low initial pressures of Np05, the rate of formation of M0. would be au, souswhat retarded at firat, if the mechanism of the decosposition were that proposed by Bodenstein, namely No03 —> NO + NO, (Ir) 2NC +0, —> 2 M0. (IIT) Reaction II is assumed by Bodenstein to be very rapid, while reaction III has been studied separately and found to ve of the third order, the rate being proportional to the concentration of the Co and to the gquare of the concgentration of the NO. However, calculations based on the reported rate of reaction III have indicated that, under the conditions of the present experiments, the large excess of O05 would raise the velocity of this oxidation to a par with the velocity of reaction I in a comparatively short time. As a result the rate of differ beyond experimental error of cbservation from the rate of ite ~ production by the direct mechanism assured in thia work, There is substantial agreement between the values of k in the (1) present results and those found by Daniele and Johnston,’ as shown by the comparison in the following table. The agreement is mora striking when attention is directed to the contrast between the initial preseures in the to sete of exverinents, and the fact thus established that the rate is independent of the initial pressure furnishes additional confirmation of the unimolecular character of the reaction. 25 Table 4. Temp. Present Results. Results of Daniels end Johnston. (c.) | Pange of kz (Ove.) | 22nge of k (Ob8.) | & (Cale.) 20° | 1.3 to 55 ma. | 0.00103 275 imi. 0.00117 | 0.00099 25° | 2.5 to 29 um. | 0.00219 268 to 290 0.00203 | 0.00203 35° 2.9 0.00837 BG to 308 o.00¢0¢ | C.0079° Loe 2.5 0.0148 ~-- --- 0.0151 In regard to the rapid disappearance of the excess ozons, which was contrary to axpectations in view of the observations of Daniels, Walf and Karrer\>) 4% is felt thet no explanation should be attempted before the phenomenon has been subjected to a mora rigorous investigation, as it probably depends on conditions whose lmportance has been underestimated. In Run No. 13, the one instance in which appeargi an exception to the general character of thse present results, there was reason to suspect that the ozone concentration was greater than the value found by analysis, for the latter would have been insufficient for the comlets oxidation of the NoQy; moreover it is noteworthy that the most probable sources of error in the eanal--tical procedure, such as leaks or insufficient haating, would yield low reaults. Acknowladgments. The author wishes to express his gratitude to Professor F.C. Tolien for suggesting the subject of this ressarch and for aid in the form of surveillance and criticism, and to Hr. 0O.R. Walf for halpful detaile of his experience with the gases under investigation. Summary . That the thermal decomposition of gaseous nitrogen pentoxide is a unimolecular reaction after the initial stage has been shown by previous investigators. The rate of the decomposition in the initial stage has been found by the present investigation to be unimolecular. Evidence of an autocatalytic induction period did net appear at the lowest concentrations of NOp measurable. A colorimetric mathod of NOo analysia has been developed, capable of measuring as low as 0.27 mm. with an accuracy of about 57. The rate of the reaction was determined at temperatures from 20° to 40°, showing e good agreement with the reavlts obtained by Daniels and Johnston. The preliminary procedure consisted in mixing freshly prepared nitrogen pentoxide with oxygen and ozons, the latter serving to re-cxidize the decomposition products until it was exhausted. In contradiction to the results of Daniels, “ulf and Rarrer, the measured excess of ozone did not retard the decomposition of the nitrozen pentoxide for an unduly long period of time. The explanation of the difference in the behavior found in the two sete of experiments is an important problem for further investi zation.