Dec., 1935
THEPHOTOCHEMICAL DECOMPOSITION OF NITRICOXIDE
2641
Stage of Development
parativelp easy to maintain atiy desired temperaAn idea of the range of usefulness of the ma- ture gradient between the air above and belbw the chine in the present state of its developmerit can cell. Devices are contemplated whish pfomise to be gairied from the follomfiing. Without heating ihake the temperature of the air above, within h and below the rotor, wholly iadependent of the the air or attempting in any way to control t natural temperature gradient existing in the rotor, t o m temperature. Stlmary the second type of rotor herein described sedimented a t every trial such heavy molecules as the 1. The air-driven spinning top has been derespiratory proteih of the earthworm (5% WIU- veloped as a transparent convectionless ultration; 1% in potassium chloride)." With the wfitrifuge. first crude air heating arrangement tried (manu2. The three types of instability of the airally adjusted electric heater inside the air supply driven rotor are described, with their causes and tubing) which enabled us to keep the templera- remedies. tures above and below the cell at the same value 3. Means are given for measuring, increasing to within * 0.3', the same rot.tn: easily sedi- and controlling the speed of the top, with a dismented molecules as small as the hemp seed cussion of the best materials of construction. protein edestin (mcaletlulat weight about 208,000). 4. Spinning top centrifuges are described in With the driving air arbitrarily thermostated to which it is possible to photograph the cell by within * O 0 . O 2 O and the room temperature con- transmitted light of any monochromatic wave trolled, the same rotor will sometimes sediment length, visible or ultraviolet. Molecules as small as purified egg albumin (molecu5. Two different types of cells are described. 1st weight about 3(5,00Q). 6. Means are discussed for measuring and More remains to be dune before tbe instrument controlling the temperature gradient existing is aiformly reliable for the smallest molecules. between the top and bottom of the cell. The authors ate corifideflt that once the exact 7. The present stage of the development of the relations between the temperatures recorded by instrument is described. the themocourtples afid the temperatures of the (18) Mr. H. J. Fouts in this Laboratory has since found that in fully to equalize the tefhpeyature inside the fbtbf with the terncell walls have been deteffnirled, the problem o?deT perature of the slip stteam from the driving air (on account of the will have been solved completely, for it is com- friction caused by the periscope upon the air within the rotor) it is (17) The best avetage +&lite we have obtained for this protein (uncorrected for viscosity ahd density) is s = 61 8 X lO-*s. The S S7, 780 value of 72 X 10-13 published by us in T ~ I JOURNAL, (19Sb),proved to be exceptionally high.
advisable to draw the air entering the periscope, as well as the air playing upon the top of the rotor, through a much colder thermostat instead ot taking it from the guard ring as described above
STANFORD UNIVERSITY, CALIF. RECEIVED AUGUST5,1935
[CONTRIBUTION FROM THE CHEbkICAL LABORATORY OF THEOHIO STATE UNIVERSITY]
The Photochemical Decomposition of Nitric Oxide' BY PAULJ. F L ~ RAND Y HERRICKL. Bertbelot2 observed that nitric oxide decomposed when irradiated with light from the quartz metcury arc and concluded that nitrogen and oxygen are the final produets of the decomposition. M&cdoaaldJS as part of his investigation of the aitrous okide decompitfon, Made a bAef study of the decomposition of nitric oxide under iriiadiatisri from the altirniritlifl spask, between pressures of 50 and 650 mm. He observed that (1) Presented before the Division of Physical and Inorganic Chemistry at the Cleveland Meeting of the American Chemical Society, Sept., 1934. (2) Berthelot, Cornat. rend., 166, 1317 (IglO). (3) Macdonald, J Chcm. SOC.,1 (1828)
JOHNSTON
the rate seemed to depend entirely upon the amount af light absorbed and reported a quantum yield of 0.75. He interpreted his data in t e ~ s of a primary mechanism of activation. Noyes4 found good evidence for the decomposition of nitfic oxide through the agency of excited mercury atoms. Under the conditions of his experiments he observed no appreciable amount of direct photochemical reaction. We have investigated the photochemical decomposition of nitric oxide both with the mercury arc and with sparks between various metal elec(4) W. A. Noyes, Jr., THISJOURXAL, S3, 514 (1931).
2642
PAUL J. FLORY AND HERRICK L. JOHNSTON
range frorn 0.02 to 7 mm. We have been successful in correlating the primary process, which proves to be an act of predissociation, with the absorption spectrum of the gas. This interpretation appears, also, to provide a better explanation than that given by Macdonald for the decomposition at the higher pressures. Experimental Reaction System.-The reaction system consisted of a cylindrical quartz cell, 75 mm. long by 25 mm. in diameter, with plane end windows. The cell was connected through a graded quartz to Pyres seal, to a calibrated McLeod gage and to mercury cut-offs which communicated to a high vacuum system and to a preparation line. No stopcocks were contained in the reaction system.
Vol. 57
visual estimates obtained with a small ultraviolet fluorescing spectroscope, and is considered reliable to within 25 A. The curves for cellophane* and for the Corex glass were determined by means of a Bausch and Lomb spectrograph equipped with a rotating sector comparator. An under-water spark was the source of continuous radiation. To prevent photolysis the filter liquids were made to flow through a thin iilter cell (one or two mm. liquid thickness) cemented to the reaction cell in such a way that the window of the latter formed one window of the filter cell.
Preliminary Observations
The Gage Reaction.--From considerations of stoichiometry alone nitrogen dioxide, nitrous oxide, nitrogen and oxygen are possibilities as ultimate products of the dissociation.b 100 However, Macdonald3has shown that nitrous oxide is photochemically decomposed ten times as rapidly as 10 nitric oxide while NorrishlO has found that nitrogen dioxide is decomposed 50 rapidly into nitric oxide and oxygen a t wave lengths shorter than 4000 A. The reverse thermal reaction is third 25 order and it can be shown from the data of Bodensteinll that the rate of 0 recombination of nitric oxide and Fig. 1.-Absorption curves of filters ( d , in centimeters; c, in moles per oxygen is negligible a t the pressures liter). of our experiments. The reaction system was evacuated to a pressure of From these considerations it would appear that 10-6 mm. prior to each run, and was usually baked out nitrogen and oxygen should be the final products during evacuation. Kitric oxide, prepared and purified of the dissociation, regardless of the primary by the method employed by Johnston and Giauque6 was introduced to the desired pressure and the reaction fol- mechanism, and that the final and initial pressures should be equal. Contrary to this expectation lowed by means of pressure measurements. In some of the experiments a condenser-diaphragm preliminary experiments showed that the pressure manometer, similar to the one described by Olson and decreased rapidly to less than one-half of the initial Hirsts was used. pressure (and went through a minimum a t about Sources of Illumination.-Most of the experiments l/qP0).l2 Furthermore, the pressure in the reaction were done with a quartz mercury arc which was relatively weak in unreversPd X 2637 resonance radiation. A num- system rose slowly when the arc was shut off but ber of experiments were also made with condensed spark dropped rapidly again, for a short period, when the discharges between suitable. metal electrodes. I t was cell was re-illuminated. As the result of several found necwsary t o electrically shield the cell, with a tin such alternate periods of light and of dark (cj. Fig. foil covering. to prevent decomposition under the influ2 ) the pressure eventually approaches a value ence of an induced field set up within the cell. Filters.---Filters were used to locate the active regions which is very near to one-half of the initial pressure. *F
of the spectrum. The absorption characteristics of these filters are shown in Fig. 1. The curves for water and for aqueous ainmonium chloride and the solid portion of the curve for :ammonium hydroxide are constructed froin the The dotted portion absorption data of Ley and Arends
'
( 5 ) Johnston and Ciauque, T H I S JOURNAL, 61, 3194 (1929,).
( 6 ) Olson and Hirst, ibid., SI, 2378 (1929). (7) Ley and Arends, Z. p h y s i k . Chem., B6, 243 (1999); B l 7 , 177 11932)
(8) The cellophane was placed between two pieces of quartz plate to protect it from the action of ozone. Although the cellophane did show some solarization the effect was not great. (9) Nitrogen tetroxide and nitrogen trioxide may be excluded since at the low pressures of our experiments these molecules would, themselves, dissociate into the simpler oxides. (10) Norrish, J . Chon. S u i . , 761 (1927). ( I t ) Bodenstein, Z. Electrochem., 24, 183 (1918). (12) The position of the minimum is based on subsequently more careful measurements with the technique finally adopted (if.rep.) for removing the accumulation of oxygen.
Dec., 1935
THEPHOTOCHEMICAL DECOMPOSITION OF NITRICOXIDE
A clue to this behavior was found in the observation that a film was deposited on the mercury
Additional confirmation was obtained in later experiments with the McLeod gage, in which the reaction was allowed to continue for five or six hours without running up the gage. Pressures measured a t the end of this period were about of the initial pressure, in good accord with the diaphragm gage curves. The presence of liquid mercury, either on the walls of the connecting tubing or in the reaction cell itself, did not affect the results in either of these experiments. We interpret this as evidence for the following series of reactions1*
0.140
0.120 vi
b
+-'
-
.-E .e
.-;0.100 2! g
2 0.080
_ _ __..- - -
h
L
H
I
,_
V . J = *
0
0
10
20
30 Time, hours.
I 40
2643
+
+
~
NO hv = '/zNz '/zOz NO f ' / ~ 0 2= NO? H g f NO2 = HgNOi
,&L
50
(I) (2) (3)
60 of which ( 2 ) and (3) occur only in
the gage. The increase in pressure after the illumination is shut off results from the slow, dark reaction
Fig. 2.-Anornalc~us behavior following the initial period of illumination.
surface in the stem of the McLeod gage, which indicated that something was being removed from the reaction system during compressions in the gage. This indication was confirmed in a series of runs in which the McLeod gage was replaced by the diaphragm manometer. The results are shown graphically in Fig. 3, for several initial pressures. The broken curve, plotted to the same scale, is from data obtained with the McLeod gage and shows the true course of the decomposition, It is clear that our expectations regarding the constancy of pressure during the early course of the photochem.ica1 decompo~ition'~ are verified. (13) The marked pr(sssure drop which comes two hours or more after the reaction is initiated is interesting although i t appears t o he irrelevant to the subject of our investigation since i t comes after the decomposition of the nitric oxide is ended (comparison with the broken curve). This fact itself is significant since i t seems to indicate t h a t the change which produces the decrease in pressure is either prevented, or compensated, through the agency of nitric oxide. i t also appears t h a t the change is produced through the action of light, or of excited mercury, on oxygen since we were able to duplicate the pressure decrease, t o the correct order of magnitude, with the cell filled with oxygcff alone. In this experiment the oxygen pressure decreased to about a,/d Poiri seven and one-half hours and was still falling. Experiments in which a Blue Purple Corex A filter, which prevented decomposition of the nitric oxide, was placed in the path of the radiant energy prevented the appearance of this delayed pressure drop. The presence of a yellowish-brown deposit (probably mercuric oxide) on the cell window and on the walls for a few millimeters hack from the window provided evidence for some reaction, either of oxygen or of ozone, with excited mercury. There is probably some relation between these results and those observed by Neujmin and Popov [Z. p h y s i k . Chcm., 8 2 1 , 1 6 (1934)j for illumination of oxygen with light of wave length shorter than X 1750 but we did not investigate the reaction further.
HgNOz = HgO
+ NO
200 400 Time, minutes.
(4)
600
Fig. 3.--Experiments with the membrane manometer;
- - _ _ _ , decomposition curve measured with the McLeod gage. (14) We refer here only to the stoichiometrical equations for the successive steps. The arguments advanced against the nonoccurrence of (2) in the reaction cell do not apply to the gage, where the pressure reaches 200 millimeters during the compression. Our data do not permit us to assign the formula HgNOz with certainty although (cf. s e q . ) they do indicate t h a t the nitrogen/oxygen ratio in the compound is very close to 1 / ~ .
PAULJ. FLORY AND HERRICK L.JOHNSTON
2644
I
Haxs of lllununation
0.00 f
4
j
;1
Hours InDa-k
+ ,~, T
Vol. 57
filter. (3) Insertion of a U-tube cooled
lo-* per hour. This figure is based on data ob- quartz cell (previously baked out thoroughly) and tained in a run, illustrated in Fig. 4, which was the rest of the reaction system, to keep the cell taken primarily to study this phase of the process. free of mercury vapor, had no effect on the rate Pco represents the limiting pressure t o which the of reaction. curve of Fig. 4 (dark reaction) extrapolates I 1 I 1 asymptotically and is indicated in that figure. Although the reactions which occur in the McLeod gage may appear to be an encumbrance to the study of the photochemical reaction, they il -2.0 actually are an aid since it becomes possible to follow the reaction by measurements of pressure. 8 The expedient was adopted, in all of our later, more careful, quantitative work, of compressing GI -2.2 the gas in the gage to a pressure of 200 mm. or more every five minutes, whether pressure read8 a, v ings were wanted or not. Since the gage con3 stituted about 75% of the total volume of the 14 -2.4 reaction system three-fourths of the accumulated oxygen was thus removed every five minutes. Preliminary trials showed that compression for a period of thirty seconds was more than ample time -2.6 to remove the oxygen from the gas, and also that 0 40 80 120 the rate of diffusion was sufficient to accomplish Time, haws. removal of the accumulated oxygen by this proFig. 5.-The first order gage reaction. cedure. We conclude that the reactions, in our experiElimination of Photosensitization as a Mechaments, are a t least 98% non-sensitized.ls nism for the Decomposition.-The absence of (15) Our results in no wise contradict those of Noyes (Ref. 4) who a mercury sensitized step in the decomposition, reported a mercury sensitized reaction with negligible photochemi. under the conditions which pertained to our ex- cal reaction. Noyes used a water-cooled arc which doubtleso pave a periments, was proved by three independent tests. highintensityof A2537 while the cooling water, if tap water wereused, must have removed almost all of the radiation which we find to be (1) A mercury vapor filter formed by placing a effeotive for the photochemical process. Our own air-ew4cd YC c)
$
h
u
THEPHOTOCHEMICAL DECOMPOSITION OF NITRICOXIDE
Dec., 1935
Dependence on Light Intensity.-The influence of the light intensity on the rate of nitric oxide decomposition was determined by the use of calibrated wire screens16placed in the light path. The results of these experiments, with three different screens, are shown in Table I. These runs were all made at about 0.12 mm. initial TABLE I DEPENDENCE OF R.EACTION RATEON Screen Change in no pressure,
1
2
3
11.1 16.9 8.9 12.7 19.1 7.5
t,
t.
l
min.
min.
30 50 40 60 100 150
19.8 31.6 15.6 22.8 36.4 15
TABLE I1 VARIATION OF REACTION UTE WITH PRESSURE Initial pressure
Po, mm.
0.0215 .a28 .0677 .0907 ,1258 .1258 LIGHTINTENSITY ,1315 ,1354 lo/t I/Io ,1990 0.681 0.66 .235 .63 .306 .39 .361 .467 .38 ,694 .364 .977 .10 .13 2.24 (cf. seq.) where, 4.37 7.12
pressure, which falls in the region for a given light intensity, the percentage decomposition is a function of time and almost independent of the pressure. “t” and “to” refer to the times required for the recorded percentage changes in pressure, with and without the screens in the light path, respectively. The last column in the table is obtained from the screen calibrations. The Primary Process
2645
(dP/dT)o mm./min:
1.12 X 2.40 3.58 4.54 6.14 6.00 6.50 6.72 9.4 10.7 13.6 19.4 22.6 26.8 36.8 39 44
%
Error
6 4 2.5 2 2 2 2 2 2 2 2 2 2 2 3 4 4
Obsd.
5.2 X 5.6 5.28 5.00 4.88 4.76 4.94 4.96 4.72 4.52 4.44 4.12 3.26 2.74 1.64 0.90 .62
Calcd.
5.22 X 5.16 5.06 4.98 4.86 4.86 4.84 4.84 4.62 4.52 4.30 3.88 3.36 2.86 1.58 0.84 . 31
ship between rate and light intensity (Table I), suggests a t once that the main factor in determining the rate may be the amount of light absorbed. This suggestion is confirmed by the agreement of the experimental values for (d In P/dt)o in column four with those, in column five, which are calculated from the relationship
Influence of Pressure on the Rate of Reacorder to obtain information as to the which is derived1’ on the assumption that the general character of the primary process the de- decomposition rate is directly proportional to the pendence of reaction velocity on the pressure of rate of light absorption and that the absorption of nitric oxide was determined for a series of pres- effective radiation obeys Beer’s law. The fit is (17) Here, k is a constant in which is incorporated the quantum sures varying between 0.02 and about 7 mm. So yield and the relation of the pressure decrease t o the number of as to minimize the possible influence of the de- molecules which disappear in unit time; Io is the intensity of the composition products on the reaction velocity, the incident light of effective wave length and a is its absorption coefficient which includes the fixed length of light path applicable to our comparisons were made between the initial rates experiment. POis the initial pressure. The derivation is as follows. of decomposition. The values of (dP/dt)o for dPldt = dPno/dt dPN,/dt (5) the several individual runs are recorded in Table = -3/2 kI,b, kaA (6) 11, in which the third column gives the probable by the stoichiometry of equations 1-3. Here A represents the preserror in the several determinations, as deduced sure equivalent oE the NO withdrawn by equation (3)and kr is the from a consideration of the accuracy of the Mc- first order velocity constant of reaction (4). Introduction of Beer’s law gives Leod gage, the completeness of oxygen removal, dPldt = - 3/2 k I o (1 - C ” ~ N O ) k,A (7) etc. which reduces, in the limit of (1 = O), to equation (8) above. The table indicates that, for very low pressures, The several factors 3/2kaIo were constant under the experimental which we employed and were evaluated collectively, by (dP/dt)o is about linear with the pressure and, a t conditions extrapolation of the experimental values of (d In P/dl)a (cf. Fig. 6) high pressures, approaches a constant value. to zero pressure. I n this limit t h e quantity in parenthesis, on the of equation ( 8 ) , becomes unity. This behavior, together with the linear relation- right The values in t h e 6fth column of Table I1 were then computed
tion.-In
+
+
gave out very little unreversed radiation of A 2537, due largely t o self reversal in t h e arc, and accounts for our low mercury sensitized yield. Subsequent experiments with the same arc operated under conditions t h a t very greatly increased the output of X 2537 reSOnMce rndiation gave evidence of a very considerable mercury sensitized reaction. (18) Forbes, J. Phy.?.Chcm., 82, 482 (1928).
through t h e choice of an arbitrary “best fit” for t h e absorption coefficient a which proved t o be, for these experiments, 1.44 in reciprocal millimeters of pressure. Justification for regarding this fit as a confirmation of the dependence of initial rate on the amount of absorbed light is provided in the excellent agreement with the experimental results over a thirty-five fold change in the initial pressure, obtained with only this one arbitrary constant.
PAULJ. FLORY AND HERRICK L. JOHNSTON
2646
VOl. 57
As the graph shows, the experimental values are a little high in the final period of the reaction. Although the discrepancy is definitely greater than the probable error of the rate measurements themselves, l9 the results are, nevertheless, in sufficiently good agreement with the general shape of the calculated curve to show that the rate-controlling factor throughout the decomposition is the absorption of light and that the accumulated nitrogen has no marked influence on the reaction. Dissociation as the Primary Step.-It is possible from the above results to deduce the general character of the primary process. At 0.1 mm. pressure of nitric oxide the average time between collisions is about seconds. If the photoreaction involves the collision of an excited and a normal molecule, then a t low pressures TABLE I11 where only a small fraction of the active radiation COMPARISON OF THE DECOMPOSITION RATES WITH AND is absorbed and where most of the excited moleWITHOUT NITROGEN AS A DILUENT cules would lose their energy by fluorescence, the rate should be proportional to the square of the 0 0.1121 4.4 X 10-4 3 . 9 (*0.1) X 10-3 min-1 nitric oxide pressure. The experimental observarnin-' .0916 4.0 X lo-' 4 . 3 (*0.4) X 0.360 tion (Table I1 and Fig. 6) that the rate becomes Figure 7 shows the course of a complete run linear with pressure a t low pressures therefore (carried out under the more careful conditions of rules out the collision process. A primary disour later experiments) in which the experimental sociation of the molecule is the only alternative. pressures are shown in their relation to a curve Failure of nitrogen, either as an accumulation calculated from the same basic assumptions that product or as an initial diluent, to modify the were employed in the preceding section and with rate also indicates the absence of primary activathe same value of the absorption coefficient which tion. We may also conclude that any secondary procwas derived from the initial rates (namely, 1.44 esses which occur subsequent to the dissociation mm. -') . are little influenced by variation in the pressure (18) This curve was calculated from the equation either of nitric oxide or of nitrogen.
shown graphically in Fig. 6 where the smooth curve is calculated by equation (8) and the circles represent experimental points. Influence of Nitrogen on the Rate of Reaction.-We .tirid that nitrogen has no effect on the rate of reaction. This was tested both by studying th.e influence of an excess of nitrogen added a t th.e beginning of an experiment and by studying the effect of nitrogen accumulation on the progress of a full reaction. The latter study also provides additional confirmation of the dependence of rate on the light absorption. Table IIL shows the results obtained in two experiments which were run to test the effect of an initial excess of nitrogen. Within the limits of experimental error the rates of nitric oxide decompositioni are the same.
This equation is derived from (7) as follows. As a first approximation reaction (4) is neglected so far as i t inEuences either the total absorption of light or the amount of H g N a remaining a t time t . Then, i t follows from the stoichiometry of reactions 1-3 t h a t we may rewrite (7) in the form
d P / d t = - 3 / 2 kIO(1
- e-a(4P-Po)13)+
2/3 h ( P o
- P)
(9)
which rearranges t o give
dP
+ (3a/4) dln (1 - e--0(4p-p0)/3)-
Expansion of (9) by MacLaurin's theorem with disregard of higher powers of (P Po) and of the term in k4 yields, on integration, the following approximate relationships
-
4P and
Po
-P
- P o = 3 Poe-"'&t = 3/4 Po(l - e-"'&t)
(11)
(12)
These values are then substituted in the third term only (which is a
The Active Region of the Spectrum Effective Mercury Arc Radiation.-It is difficult to be quite certain of the lower limit of the small correction term) of equation (IO) and the denominator of this term is then expanded by MacLaurin's theorem, and powers of the exponent higher than the first are neglected. This treatment yields the equation
dP
+ 3a - d In ( 1 - e-o(4P--Po) 4 k4(e2aktro - 1 ) dt + 3 / 2 k l o dt = 0 2a /3)
(13)
which integrates to give (14)above. The several approximations introduced in this derivation are exact in the limit of 1 = 0 and remain good throughout the greater portion of the reaction. We cannot expect this equation to reproduce accurately the experimental data in the immediate neighborhood of the pressure minimum (shown in Fig. 4), where reaction (4) plays a relatively great inEuence. (19) The formation of a little mercurous nitrate (or other compound with an 0 to N ratio greater than Z ) , in the gage may be the cause of the discrepancy.
Dec., 1935
THEPHOTOCHEMICAL DECOMPOSITION OF NITRICOXIDE
2647
about 50% corresponds to the absorption of water for wave lengths in the neighborhood of X 1830. The only mercury lines which fall within this active region are those a t XXX 1775, 1832 and 1849, respectively (cf. Fig. 8). Hg 1775 falls in a region which is free of nitric oxide 5 absorption. Furthermore, it should be removed completely by the water 2 filter. The results of the several experiments with the mercury arc suggest x4 * I therefore that the primary process is d either the absorption of unreversed 33 radiation from the strong line a t X 1849 .-EI in the (9,O) band of the P-system or the 3 absorption of the weak line at X 1832 G2 (together with continuous radiation 9 'c1 surrounding X 1849) in the (LO) band of -1 the &system. To gain further information on this point an approximate meas0 urement of the absomtion coefficient 0 2 4 6 of X 1849 was carried out spectrographiPressure in millimeters. cally. An absorption cell 5 cm. long Fig. &-Dependence: of the initial decomposition rate on the pressure: (approximately the mean light path 0, obwwed: , calculated. of the divergent light which passed
wave lengths which may have entered our reaction cell but in view of the intense continuous absorption which sets in with air below X 1750 we are safe in concluding that no radiation below
I).
0 h
arc a number of runs were carried out with the filters whose absorption characteristics are shown in Fig. 1. Reference has already been made to the results obtained with the Corex and the cellophane filters which showed that radiation below X 2300 is responsible for a t least 98% of the reaction, under the conditions which pertained to our experiments. The results obtained with the other filters are given in Table IV, which also includes comparative rates which show the influence of filters on the reaction promoted by radiation from
' I-_\-__c_l o
0.04
(20) Quantitatively, the 3% residual rates with these filters agree, within limits of error, with the 1% rates obtained with the cellophane and Corex filters and indicate: (1) that none of the mercury arc radiation between X 1900 and X 2300 is effeetive in promoting the decomposition and (2) that the small effect due to radiation of longer wave length is due to unreversed radiation of the X 2637 line.
~
b
I
with Eastman ultraviolet sensitizing solution) the absorption of the 1849 h e was determined. Application of Beer's Law, which Lambrey21 has ( 2 1 ) Lambrey. Ann. Phys.. 14, 96 (1030).
2648
PAUL
J. FLORYA N D HERRICK L. JOHNSTON
VOl. 57
TABLE IV almost complete removal of the line. This result INFLUENCE ‘OFLIGHTFILTERS ON THE RATES OF NITRIC is not incompatible with an absorption coefficient OXIDEDECOMPOSITION of 1.4. Comparative rates (per cent. pressure decrease per minute) H20 NH&I “,OH Empty cell d = 0.15 cm. cd = 0.0013 cd 0.25
I
I
0
0
11
~
0
0
A
.v2 H g .
I
. . -! L
I
I
I
*
A
e
1
3
d
I
* A . , ?
L
p
0
Y
a
From this we conclude that most, if not all, of the reaction is due to absorption in the 6 (1,O)
-
Source of radiation
6
I
It
II
$Zn i
I
I
I
2 AI 1
I
I
I
II Ill
Cd
I
1 I I I Ill
1
I’
iI
I
.___
I
I I
l
I
l I
*I
~
I
I1I
I
/ / I /I
Ni
1
I I
**
I
I
III I
11
I I
some photochemical influence of P-band absorption. The influence of water and am.monium chloride filters on reactions promoted by the aluminum and zinc sparks is included in Table IV. Comparative results obtained with six different spark sources with a 2-mm. water filter are shown in Table V.22 The most general observation that we can draw from
Dec., 1935
THEPHOTOCHEMICAL DECOMPOSITION OF NITRIC OXIDE
2649
169,350 calories per mole) with that of the heat of DIFFERENT dissociation of oxygen2’ and with the heat of of nitric oxide a t the absolute zero,28we formation Rate (% pressure Spark source decreases per miaute) obtain 121,950 calories as the molal heat of dissociaA1 0.30 tion of nitric oxide into atoms, a t zero degrees Zn .23 Kelvin. The energy associated with radiation of Cd .21 wave length 1830 A. exceeds this dissociation Ni * 19 energy of nitric oxide by approximately 33,600 cu .11 Sn .25 calories. Since this is less than the energy necessary to put either atom into an excited electronic the mercury arc--namely, absorption in the 6 (1,O) band of nitric oxide, together with some absorption in the hands of the fl-system, of which weak absorption of Hg 1849 (and perhaps of Zn 2025) constitute special examples. The nature of the dependence of the reaction rate on nitric oxide pressure and its independence of nitrogen pressure indicated that the primary step is dissociation. The fact that the active radiation falls in a banded region of the spectrum proves that the dissociation process is one of predissociation. This conclusion receives support in the emission spectrum of nitr:ic oxide. Thus KaplanZ4 attributes the absence of bands with v’ greater than 4 (and the unexpected weakness of (v‘ = 4) bands) to predissociation to a repulsive level which must intersect the upper potential energy curve (cf. Fig. 9) somewhere in the neighborhood of v’ = 4. A similar explanation may be offered to explain I 0 the fact that no 6 emission bands have been observed25 with v’ greater than zero, although 1.00 1.50 2.00 2.50 Internuclear distance, A. strong emission bands would be expected for higher vibration states, both on the basis of the Fig. 9.-Potential energy curves for the known states of nitric oxide: - - - -, estimated position Franck-Condon principle and of the distribution of the 4X repulsive level responsible for predisin absorption. sociation in the region X 1800-2000. We expect that radiation absorbed by any of the bands which are marked with open circles in state the dissociation products must be the norFig. 8 would be photochemically active while mal atoms. And since momentum, as well as radiation absorbed by bands marked with black energy, must be conserved in the dissociation, it is circles would be inactive (except for some activity a simple matter to compute the kinetic energies in the p (4,O) band). given to the respective atoms in the process. These prove to be: for nitrogen, 17,630 calories Secondary Processes per gram atom; for oxygen 15,690 calories per Nature of the Dissociation Products.-The gram atom. dissociation energy of nitrogen now seems to be Atomic Reactions.-Stoichiometrically pospretty definitely fixed in the neighborhood of 7.34 sible secondary processes are : volts.2s Combining this value (equivalent to TABLE V
COMPARATIVE RATESOF DEeOMPOSITION WITH SPARKS
I
(24) Kaplan, P h y s . Ret?.,87, 1406 (1931). (24) Knauss, i b i d . , S2, 417 (1928); Schmid, 2. PhysiR, 6% 42, 860 (1929); ibid., 64,279 (1930) (26) Herzberg and Sponrr, Z. p k y s i k . Chem., B26, I (1934); Mulliken, Phys. Rev., 46, 144 (1934). C j also Vegard. NaZurL, 194, 696 (1934), and Lozier, .Phys. Rm.,44, 476 (1933); ibid., 46, €30 (1934).
+ + 20 +
N NO = NI 0 +KO = 0 2 2N W Nz mi =
0,
+0 +N
+ 47,400 -
4,6CO 4-169,350
+ll7,350
(15)
(16) 07) (18)
(27) Herzberg. 2.physik. Chena., B10,189 (1930). (28) Giauque and Clayton. THIS JODRNhL, 66,4886 (1Q33).
PAULJ. FLORYAND HERRICK L.JOHNSTON
2650 N 0 0 N N W
+ 0 + W .= NO + Nz = NO + N + Nz = N20 + = NO + 0 + W = NO2 + NO -+- W N20
+121,950
(19)
-
(20) (21) (22) (23) (24) (25)
+ + +
47,400 12,600 4,600 53,250 34,800 48,650
Vol. 57
Correlation with the Decomposition at Higher Pressures
M a ~ d o n a l d ,who ~ made a brief study of the decomposition for the pressure range 50 to 650 millimeters, found an extinction coefficient for the 0 + N O -t- W = NO9 active radiation which is only 2% of that we have where W represents reaction with an adsorbed computed for the active radiation a t low pressures layer of atoms (or molecules) on the walls. and about double that which we have measured Of these, (20) and (21) are ruled out com- for the absorption of X 1849. From this we conpletely, anti (16) reduced to an improbable occur- clude that p-band absorption, followed by prerence, from energy considerations al0ne.~9 dissociation, is the primary mechanism of the high A simple calculation, based on the measured pressure decomposition. This is reasonable since, rate constants, indicates that the accumulation of through this range of pressures, absorption in the oxygen in the reaction system rarely exceeded 1 X p-bands must increase in almost direct proportion mm., and was, as a rule, considerably lower to the pressure while &absorption, already subthan this. This eliminates reaction ( 2 2 ) , in com- stantially complete even at four or five milliparison with (17) or (19), as a contender for the meters, can show little gain. removal of nitrogen atoms. This conclusion differs from that of Macdonald, At the lowest pressures (0.02 mm.) a t which the who pictured the primary mechanism as one of photodecomposition was studied wall recombinaactivation, followed by independent bimolecular tion of atoms by reactions (17), (18) and (19) reactions between activated and normal moleshould become a competing process with any gas cules. However, we find nothing in Macdonald's phase reaction, such as (15). If this latter reacwork (which was restricted almost entirely to the tion shoulld occur to an appreciable extent the measurement of quantum yield and to the relation rate would fall off, a t these low pressures, more between reaction rate and light absorption) which rapidly than with the first power of the pressure, precludes the predissociation mechanism. which is contrary to the experimental results (Fig. Although the details of the secondary processes 6). Furthermore, the quantum yield a t high appear less certain, most of the arguments adpressures, where bimolecular gas phase reactions vanced in favor of atomic recombinations a t the would be favored, is less than unity while reaction walls, for the low pressure decomposition, appear (15) would, itself, lead to a quantum yield of two. applicable also a t the higher pressures. The one Therefore we conclude that the recombinations of objection to this conclusion is hlacdonald's puratoms on the walls, by reactions (17), (18) and ported observation of appreciable nitrous oxide (19), constitute nearly the entire secondary' procamong the reaction products. This may be ess. If the specific velocity constants of (17), accounted for in terms of reaction (24) although (18) and (19) are numerically equal there should to produce 10% nitrous oxide among the products be a quantum yield of 0.5, but values from 0 to 1 would require a very considerable amount of this are possible depending on the ratio of the mean of reaction since nitrous oxide itself decomposes therates of (17) and (18) to that of (19).30 much faster than nitric oxide. This would re(29) I n collision of a fast oxygen atom, with either nitrogen or quire a quantum yield approaching 1.5, which is nitric oxide in the gas phase, conservation of linear momentum exacts a minimum toll of 35% of the kinetic energy of the oxygen. contrary to the facts. On the basis of what de0 2
4 - 0 2
+
This leaves ouly 10,000 calories, a s a maximum, for activation. In only a small traction of the collisions will this maximum energy be available. The great majority of collisions will be elastic due to steric and statistical factors and will dissipate the high energy of the oxygen in successive small losses. A. further argument against (IS) ir the fact t h a t nitrogen as a diluent did not affectthe late. (30) Reactions (23), (24) and (25) are not excluded on any of the grounds advanced above, although wall reactions of this character appear considerably less probable than (17), (18) or (19). In any case (23), (24) or (25) must be considered in connection with the ultimate products which result from the photodecomposition of the NO1 or N20, which are more rapidly decomposed than the KO. Norrish (ref. 10) has shown t h a t NO, decomposes photochemically, by a bimolecular process, to yield NO and 0 2 . The net effect of (23) would thus bc: the complete reversal of the initial photodecomposition of nitric oxide, and a consequent quantum yield of zero if this
process represented the normal method of removing the nitrogen atoms. I t is clearly not of much relative importance in the actual decomposition. The net influence of (25) would be the substitution of a more complex process than (IS) for the formation of molecular oxygen. Reactions (17) or (19) would still be required to account for the removal of atomic nitrogen. hfacdonald (ref. 3) finds t h a t the photodecomposition of Nz0 yields N r , N O and 0 2 in the proportions of 3,2 and 1. Thus the net effect of (24) is the substitution of a more complex process than (17) or (19) for the removal of nitrogen atoms and would lead to a quantum yield of 1.5, which is double the value observed by Macdonald. Any of these mechanisms (23, 24 or 25) require t h a t one or more of the simple reactions (17 18 or 19) be operative for the removal of one species of atom. I t appears reasonable t o credit the entire secondary process, in the main, to the simpler set of reactions.
PSEUDOCODEINONE
Dec., 1935
scription Macdonald gives of his analytical procedure we feel that the presence of nitrous oxide in the reaction products may not be fully established. Summary The photochemical decomposition of nitric oxide has been studied, as a function of the pressure, over the pressure range 0.02 to 7 mm. with irradiation both from the mercury arc and from sparks between electrodes of aluminum, zinc, cadmium, nickel, copper and tin, respectively. The final products of the reaction are nitrogen and oxygen although the latter is removed by a reaction which takes place during compression of the gas in the McLeod gage. This reaction produces a solid product, probably mercurous nitrite, which decomposes to liberate nitric oxide by a first order process which has a velocity constant of 1.5 X lo-* hr.-l. The rate of decomposition proves to be directly proportional to the rate of light absorption for light of the effective wave length and is, a t very low pressures, (directly proportional to the pressure. The absorption of effective radiation follows Beer’s law and has, for the reaction with the mercury arc, an extinction coefficient of about 2.9 X pe’r millimeter light path per millimeter of pressure. Nitrogen as a diluent does not appreciably influence the rate of decomposition. These results, alone, indicate primary dissociation as the first step in the process. The use of filters containing solutions of ammonium chloride and of ammonium hydroxide places the effective mercury arc radiation below
CONTRIBUTION FROM THE COBB CHEMICAL
2631
X 1900. The use of a water filter indicates that
the effective region is in the neighborhood of X 1330. The strong resonance line a t X 1849 is
ruled out, except for a relatively small influence at the higher pressures, by measurement of its absorption coeficient which proves to be one hundred times too small for the observed reaction. The alternative is the weak h 1832 line of mercury which is strongly absorbed by the (1’,0”) band of the &system of nitric oxide. The essential features of this interpretation are confirmed by the measurement of the decomposition under the influence of both filtered and unfiltered radiation from sparks between metal electrodes. The latter owe the greater portion of their effectiveness to a background of continuous radiation below X 1900. This interpretation of the primary process, which must be a predissociation since it occurs in a banded region of the spectrum, agrees well with the evidence for predissociation provided by the emission spectrum of nitric oxide. Possible secondary processes are considered. All are believed ruled out on theoretical or experimental grounds, in comparison with Combination of nitrogen and oxygen atoms on the walls. Macdonald’s work with high pressures of nitric oxide has been considered and it is concluded that the mechanism determined for the low pressure decomposition may also be applied to the results a t higher pressures. However, predissociation from the upper levels of the ,&bands, rather than the &bands, provides the primary step a t these higher pressures. COLUMBUS, OHIO
RECEIVED JUNE 25, 1935
LABORATORY, UNIVERSITY OF VIRGINIA ]
Reduction Studies in the Morphine Series. VII. Pseudocodeinonel BY ROBERT E. LUTZAND LYNDON SMALL Pharmacological studies which have been carried out a t the IJniversity of Michigan coordinate with the chemical investigations a t this Laboratory have shown that as regards general physiological action, the four codeine isomers may be grouped in pairs, codeine with allopseudocodeine, (1) The work reported in this paper i s part of a unification of effort by a number of agencies having responsibility for the solution of the problem of drug addiction. The organizations taking part are: The Rockefeller Foundation, the National Research Council, the U. S. Public Health Service. the U.S. Bureau of Narcotics. the University of Virginia and the University of Michigan.
and isocodeine with pseudocodeine.2 This pairing is based primarily upon the relative degree of convulsive and depressant actions exhibited by each of the drugs, upon their toxicity and upon their effective ratios in other respects when toxicity is taken into account. The same relationship holds for the four corresponding isomeric morphines, and suggests that the spatial arrangement of the alcoholic hydroxyl group may be of (2) Eddy. .7 Phnrmaco! , 45, 361 (1932).
