Effects of Relative Velocity upon Gaseous Ion-Molecule Reactions

velocity and in the macroscopic cross section Q upon thekinetic energy £„ of the ions at the exit slit. Charge transfer from various ionsto wfo-C6H...
0 downloads 0 Views 599KB Size
N. BOELRIJKAND W. H. HAMILL

730 [CONTRIBUTION FROM

THE

DEPARTMENT O F CHEMISTRY O F THE UNIVERSITY

OF

Vol. 84 NOTREDAME,NOTRE DAME,INDIANA]

Effects of Relative Velocity upon Gaseous Ion-Molecule Reactions ; Charge Transfer to the Neopentane Molecule' BY N. BOELRIJKAND W. H. HAMILL RECEIVED JUNE 30, 1961 The dependence of ion-molecule reaction cross section upon relative velocity g is not adequately described for ion-induced dipole forces by expressions previously used, which are presumed valid for point particles. Introducing an energy-independent component in the microscopic cross section u leads to discontinuity in the functional dependence of u upon ion velocity and in the macroscopic cross section Q upon the kinetic energy E., of the ions a t the exit slit. Charge transfer from various ions to neo-CsHlz to give CaH12+ appears to be energy-limited, for which Q a Ee-l is predicted and observed.

Introduction Results In order to understand radiation chemistry in The dependence upon ion repeller field of the terms of ionic processes it is necessary to consider, cross section Q of the reaction among others, the phenomenon of charge transfer. Ar+ + HP+ArH+ H A recently reported example2is CzHz+ C6H6 --+ CzHz f C B H ~ + was measured on the Notre Dame instrument to alwhich accounts for the previously observed retard- low comparison with published results. We find, ing effect of benzene on the alpha-induced poly- in terms of the equation Q = 2 a ~ / E ' / z Uk t h a t merization of a ~ e t y l e n e . ~The same study2 also UL = 67 x 10-l6 cm.2ev.'/Z and U k = 11 x 10-16 indicates that charge transfer may not occur when cm.2 From our graph of the data of Stevenson and Schi~sler,~ who used a Westinghouse Type LV, an efficient ion-molecule reaction also is allowed. The present study was undertaken with neo- ir/2-sector analyzer mass spectrometer, we obtain pentane as the common electron donor in various U L = 71 x cm.2ev.'/z and CTk = 13 x binary mixtures in a conventional mass spectrom- cm.2 The theoretical value of UL is 70 X 10-16 eter. This choice of donor molecule was dictated ev.'l2for a reaction probability of unity. Pressure Dependence.--A necessary condition by the impossibility of detecting an increment in the ion current of the molecular (unfragmented) ion for ion-molecule reactions is a second order preswhen i t also was produced, even in fairly low abun- sure dependence for the secondary ion, whereas a dance, by electron impact. The ion C5H12+ from primary ion exhibits first order pressure dependneopentane has a relative abundance of only ca. ence. We have found that a t fairly low repeller O.Olyo,and even small contributions from bimolec- field strength this is not the case. Thus, for CD4+ and CD5+ions by first and second order processes ular processes can be detected efficiently. Incidental to the study of neopentane in various in CD4, and a t 4 v. cm.-', the order is 0.88 and 1.66, mixtures for evidence of charge transfer, it was respectively. B t 8 and 40 v. cm.-l, however, the observed that the yield of C5H11+ ions (0.04y0) also expected integral order dependences actually were was enhanced by bimolecular processes. This observed. Measurements of cross sections on effect also has been observed by Field and L a r n ~ e . ~instruments of the type used in this work often are not dependable a t less than 10 v. cm.-l repeller Experimental field strength. Measurements were made in a CEC 21-103A mass specNeopentane.--iln examination of many subtrometer with a model 31-402 ion source. The repeller potential was manually adjusted using dry cells and step- stances for evidence of charge transfer in appropripotentiometers adjustable by 0.1 v. increments over the ate mixtures resulted in choosing neopentane berange 0-24 v. The ionizing voltage was continuously cause of its particularly small 72-ion abundance. adjustable over the range 5-73 v. and in steps of 0.1 v. Incidental t o this study we found, as did Field over the range 5-28 v. Electron current was 10 p-amp. The distance from the center of the electron beam, which and Lampe14evidence for secondary 71-ion formais 0.018 cm. thick, to the exit slit is 0.130 cm. and to the tion from neopentane. Such parent-mass-minusrepeller plates is 0.120 cm. The electron energy therefore one secondary ions are of common occurrence, both is increased by one-half the applied repeller voltages. The leak consists of two fine holes in a thin gold foil, by unimolecular decomposition and by "hydride and observation shows gas flow to be effusive within the ion" transfer. range of pressures used. Pumping of gas from the source I n the present work i t was observed that the 71also is effusive and the ratio of pressures between the reservoir and the source is constant and independent of the ion current increased as the 1.66 power of the presgas used. By calibration,5 the concentration of gas in sure of neopentane, both a t 12 and a t 40 v. cm.-' the ion source is 1.58 X 1O1O mol./cc. per micron of reservoir This suggests a combination of primary and secpressure. ondary processes. The ijl-ion current may be referred arbitrarily to the primary &,-ion current, as (1) Presented at the 136th National Meeting of the American Chemical Society, Atlantic City, September, 1959. This work was a matter of convenience, since they are of the same performed under the auspices of the Radiation Laboratory, University magnitude. For later reference we note that of Notre Dame, supported in part by the U.S. Atomic Energy Comi E 6 / i 5 7 = 0.0354%. It is to be expected then that mission under contract AT(ll-1)-38. i 7 1 = a' PI b' p12 and i65 = u" pl where pl is the (2) P. S. Rudolph and C. E. Melton, J . Chem. P h y s . , 32, 586 (1960). (3) S. C. Lind and P. S. Rudolph, ibid.. 26, 1768 (1957). pressure of neopentane. It was found that i ? l / i 6 5 (4) F. H. Field and F. W. Lampe. J . A m . Chem. SOC.,80, 5587 = 0.63 1.03 X pl a t a repeller field 12 v. (1958). cm.-l and &/i85 = 0.58 3.91 X 10-3p1 a t 40 V. ( 5 ) D. P. Stevenson and D. 0. Schissler, J . Chenz. P h y s . , 29, 282 cm.-1, expressing p in microns reservoir pressure. (1958).

+

+

+

+

+

+

March 5, 1962

VELOCITY

EFFECTS ON

GA4S€30US ION-MOLECULE REACTIONS

I n neopentane mixtures it may be expected that b' p12 c' pl p2, where p2 is the pressure of additive. The ratio i 7 1 / i 6 6 should be linear in p 2 . Ethane, propane and carbon disulfide give such results. Keeping pl = 105 microns and Ee, the ion energy at the exit slit, at 1.8 ev., the results for added ethane are described by the equation i71/iC& = 1.51 3.8 X pz and for added propane by iTl/iee = 1.55 3.4 x 10-3 p2 I n mixtures with carbon disulfide i t was not possible to refer i71 to because of interference. The results for these mixtures a t E, = 1.56 ev. and 418 microns constant total pressure may be expressed by

i,l = alpl

+

+

+

+

= 0.42

&/pi

+ 2.1 x 1O-'P1

Since the parameters of this equation are not derivable from those for pure neopentane by a common factor, i t appears that an ion from carbon disulfide (shown to be CS2+) also induces the formation of C6Hllf.

731

ev. enhanced i72. Propane is an apparent exception, but for this and all other hydrocarbons tested there. is a fairly good correlation between A i72/&6 and $28. The largest effect observed for any additive was produced by carbon disulfide. Secondary Ion Current-Repeller Dependence.Carbon disulfide increases rather less than &, but the dependence of upon E, is unlike that for any other additive; the results for carbon disulfide and for ethane-& are compared in Table 11. TABLE I1 EFFECTOF REPELLER FIELD Added

subs.

Ee

CSz

Ai72/ie5 Ai?i/ies

c% C2D6 CzDs

Ai7z/i65 Ai7l/'&

0.42

1.04

1.56

2.60

5.20

2.40 0.99 0.67 0.38 0.18 .26 .18 .50 .40 -0.20 .064 .029 .I38 ,085 0.282 .59 .35 1.31 .95 3.10

The dependence of the 71-ion current upon E,, the terminal ion energy, relative to the 65-ion current, is reported in Table I11 for pure neopentane and for several mixtures. The parameters M and C refer to the empirical equations i71/if,b = C ;ZfE,-'/2, etc., as indicated. 72-IONABUNDANCES

TABLE I UPON71- AND EFFECTSOF ADDITIVES RELATIVE TO THE PRIMARY 65-ION AT E, = 0.4 e.V. Additive (MI

A ? ZEi

0.0

He

A?! 165

0.0

id A

+-166ZM

I(M)

24.58 15.44 Nz .5 15.56 a" .o 10.23 coz .3 13.79 13.26 CD4 2.7 0.0 11.41 CzHz .10 0.16 10.51 C2H4 1.7" CZH6 2.7 .21 .23 11.65 CZD6 2.8 .28 .27 11.65 .o 9.7 CaH, 3.7 .o 10.09 c-C~H~ 4.8 10.6" .18 11.21 CsHs .o 9.35 i-C4Hs 7.1 .o 9.72 1-CdHs 4.5 .1 10.8 n-C4Hl0 4.8 n-CaD10 4.8 .1 10.8 .o 11.22 CHsCl 2.0 CH3Br 3.4 .26 .26 10.54 1.1 .o 9.54 CHII ,064 10.97 CzHsCl 2.5 .03 C&Br 4.5 .45 .65 10.29 .o 9.33 CzHsI 5.9 .o 10.85 CH30H 1.9 .o 9.8 Pyridine 1.9 1.3 10.08 CSZ -0.2 2.40 5 Measured a t E , = 0.25 ev. i ~ is+ the current of the molecular ion of additive except for C&, C& when it refers to CzH2, C2D4+.

.o

Dz

-

.o .o .o .o .o .o

+

TABLEI11 DEPENDEKCE OF SECONDARY-TO-PRIMARY ION CURRENTS FIELDSTRENGTH UPONREPELLER Reservoir 9 , microns

Additive

Range E., e.v.

C

M

7.36c 1.3- 5 . 2 -0.97 418 None - .52 4.18' 217 1.9- 5 . 2 None .09 3.41"*' 217 1.6- 8 . 3 None .;5 4.4",e 217 5.2-12.5 None - .85 5.65' CZH6 209 209 1.3- 5 . 2 - .35 4.14c CZH6 105 313 1.6- 6 . 2 .61 4.44' 105 313 5.0-12.5 CzHe 1.0- 5 . 2 - .4'i 1.7gd CZD6 209 + 2 0 9 - .36 1.5yd 1.0- 5 . 2 CZD6 209 + 2 0 9 .O 0.14b*1 209 209 1.0- 5 . 2 CZD6 209 209 1.0- 5 . 2 . o 0 . Eb*' CZD6 - .78 3.13d 1.0- 5 . 2 n-CsHe 209 + 2 0 9 -1.28 7.77' C3Hs 209 209 1.6- 4 . 3 -0.25 1.3Td CHsCl 209 209 1.0- 5 . 2 -0.02 1.06b-f CSZ 209 209 1.0- 5 . 2 The accelerating voltage was 1500 v. for this series and 600 v. for all others. This may account for the discrepancy between the intercepts for this and the preceding series in Corrected for isotopic contributions the plot us. Ee-'/2. from CsHu+. Parameters refer to i71/i65 = C hfE,-'/2. d Parameters refer to Aiil/zSs = C 11f&-~/2. e PararnME,-'. f Parameters refer to eters refer to i 7 l / i 6 6 = C Ai72/&5 = C ME,-'.

+ + + +

+ + +

+

+

+

+

+

Appearance Potentials.-Appearance potentials were measured relative to xenon, a t 0.7 to 0.8 repeller voltage. The results appear in Table I V . TABLE IV

The 72-ion current for neopentane, as received, APPEARANCEPOTENTIALS OF NEOPENTANE, ALONE AND IN was not altered by gas chromatographic purificaMIXTURES tion of the material. Neither does this current Additive: None CS? CzHs CiHa arise from a secondary reaction, since 22 measureCt"i 12.5 11.8 10.7 ments of i 7 2 / i 6 6 (corrected for the isotopic comCsHiz' 12.4 10.1 12.1 10.6 ponent from i71) a t various pl and repeller fields The appreciable primary 71-ion current made i t showed no significant variation. The results of tests with many additives are reported in Table I. impossible to establish the primary ion from neoNo ion with ionization potential I greater than 11 pentane responsible for the second order contribu+

N. BOELRIJK AND W. H. HAMILL

732

tion t o i,l/&. Field and Lampe' have reported on the basis of appearance potential measurements, that the 41-ion was responsible. They emphasized that such "hydride ion" reactions are probably rather general, and the present results support this view for neopentane. The appearance potentials of C&+, CS+, S+ and C+ from CS2 are 10.10, 14.8, 14.1 and 25.1, respectively.6 The charge transfer process in mixtures with carbon disulfide is considered to be C8+ C6Hii +C 8 CsHiz+ The only appearance potentials below 13.5 ev. for ethane and ethylene are: C2H8+, 11.65; C2H6+, 12.84; C2H4+, 12.09 from ethane and7.* C2H4+, 10.51 from e t h ~ l e n e . ~ The indicated reactions in mixtures with ethylene are

+

+

CzH4+ CzHd+

+ C6H12 +CsH12+ + C2H4 + C6Hu +CzHs + CsHii+

(A) (B)

A second break in the ionization efficiency curve of CsHll+ for the neopentane-ethylene mixture was observed a t 13.0 ev. This may be attributed to a reaction of C3H6+ from neopentane itself. The reported appearance potential for the primary 41-ion'' is 13.13 ev. I n mixtures with ethane i t is difficult to assign the primary ions with assurance but the best agreement is found for the processes GHo+ CzHs+

+ CsH12 +CsHil+ + CzHa + H + CsHu +CsHii+ + CzHs + Hz

(C) (D)

The reactions A and B probably also take place in these mixtures. Discussion Ion-induced dipole forces lead t o energy-dependent collision cross sectionsl1.l2 u ( E ) = abo2 = 2 ~ ( e * a / p g ~ ) ~ / *

(2)

Since we also have Q

=

Ee-l

$"' u ( E ) d E 0

(3)

it follows that Q = 2re(2rnla/,xEe)'/r = 201 E,-'/%

where E, is the ion energy a t the exit slit and UL represents the collected constants or the value of o(E) a t one ev. This equation has been used to describe the results of several ion-molecule reactions.sJ2 There are, however, many more reactions which depart considerably from the predicted linear dependence of Q upon Ee-'Ir, in particular reactions of hydrocarbons for whichla Q varies as Ee-'.

Re-examination of Previous Work.-ln the derivation of equation 4 by Gioumousis and Stevenson12 the most serious restriction imposed is that the gas kinetic collision cross section OK shall be small compared to nbo2. "Thus the analysis is most likely to be valid for some such reaction as that between a noble gas ion and hydrogen, or hydrogen with hydrogen."12 On the other hand, if the preceding description is to be consistent with short range repulsive forces, it is to be expected that as the relative energy of the collision partners increases the Q of equation 4 becomes eclipsed by a nearly energy-independent cross section, CK, whose value should approximate the gas kinetic collision cross section. Some of the relevant parameters are listed herewith in Table V for a few reactions. TABLE V COMPARISON OF UL

x

mi

AND UK

+ m: rLX1fl" e v , l / l x ioin cm.2 cm.3 UK

Ion

Molecule

cc.

mp

Hz+(Dz+) Ar+ Kr+ H1+ CHh' C2H4 +

Hz(D2) Hz Hz Kr CHp CzH4 D2

0.78 .79 .79 2.48 2.6 4.26 0.79 1.60

2 21 43 1.0

0 2

+

D2+

0 2

21

21 69 97 33 38 49 45 22

2 2 9

1.1

24 26 26 33 40 24 24

' ~ { z

1.0 8.2 14.0 1.6 1.4 1.5 6.6 0.8

(1)

where CY,p and g are the electric polarizability of the molecule, the reduced mass and the relative velocity of the collision pair. As an adequate approximation the molecule may be considered to be a t rest and the energy E of the ion of mass ml taken as E = 1/2m1 g2. The current of secondary ions is formed from zp primary ions over a path do a t a concentration of N molecules/cc. is expressible in terms of the phenomenological cross section Q Q = i./i,Ndo

Vol. 84

(4)

(6) H. D. Smyth and J. P. Blewett, Phys. Reo., 46, 276 (1934). (7) D. P. Stevenson and J. A. Hipple, J . A m . Chcm. Soc., 64, 1588 (1942). (8) R . Watanabe, J . Chcm. Phys., 26, 542 (1957). (9) W. C. price and W. T. Tutte, Proc. Roy. SOC.(London), 8174, 207 (1940). (10) F. W. Lampe and F. H. Field, J . A m . Chcm. Sac., 81, 3238 (1')59). (11) P. Langevin, Ann. chim. pkys., 6 , 245 (1905). (12) G. Gioumousis and D. P. Stevenson, I . Chem. Phys., 29, 294 (1958).

Values of CK are based on van der Waals radii. The last column lists values of the particular ion energy Et a t which C L / E ~ ' /= ' UK. I t is Et =

(UL/UK)~

=

+

2a2e2a(ml

m2)/m2uK2

(5)

It appears that equation 4 is inadequate to describe Q(E) for many reactions, even a t relatively low ion energy, and also for reactants as small as Hz+ and Hz. Let us represent the reaction cross section by u ( E ) = [ u i / E 1 / l- uc]

+

uk

(6)

The term in square brackets represents the effective area for glancing collisions; (TO and Ok correspond to head-on collisions. In the language of collision theory, each sigma involves the product of an area and an efficiency factor. Tentatively we may identify UI with PLCLand 00 with PLOK where UK is a gas kinetic collision cross section. Similarly, bk becomes P K C K ;the reaction probability factors P L and P K need not be equal, so that co and ffk may be quite different and either may well be zero. For lack of information 00 and U k are taken to be independent of energy. The equation is valid for 0 < E < Et for which the term in (13) F. H . Field, J. L. Franklin and F. W. Lampe, J . A m . C k m . Soc., 79, 2419 (1957).

March 5, 1962

733

VELOCITY EFFECTSON GASEOUS ION-MOLECULE REACTIONS

For E

brackets equals or exceeds zero. assume u(E)

> Et

we 120

(7) I

Uk

Integration over the low energy range gives Q = Ee-l

where

JOE'

UkO

=

u ( E ) d E = 2aEe-'/3 Uk

- UO.

Q = E*-1 JoEt u ( E ) d E

+

Uke;

E.

< Et

(8) 100

At higher energy

+

JOE'

+

Ee-'[2u~Et'/*- UOEJ

a(E)dE =

Uk

Ee-'L:

+

Uk

(9)

If the preceding considerations are valid, the dependence of Q upon E, for the reaction Dz+ + D2 +Ds+ 4- D should obey equations 8 and 9, rather than 4. The results of Stevenson and Schissler,6 for this and other reactions, appear in Fig. 1, all expressed in terms of Q us. E e - l I t . The cross section for ArHf obeys equation 8 rather than 4, while that for D3f and DOz+ indicate departure from equation 8 a t fairly low E,. Since we attribute greater precision to these measurements than did the observers, further consideration is desirable. There is a necessary relationship between the parameters of equations 8 and 9 for the particular functional dependence of Q upon E which we have postulated Considering that insufficient data are available to establish empirically the correctness of the postulated functions by graphing, we may take the concordance of the observed and calculated slopes as a limited test of the adequacy of equations 8 and 9 to describe the facts. The results of this comparison appear in Table VI. Values of Et, 01 and 2 are all in fair agreement with theory. The present data do not permit a critical test of the alternative descriptions. Other ~ o r k ' ~ - Pro'~ vides further evidence of the effects implied in equations 6-10. TABLE VI Et",

DI+

++ +

e.v.

c1.O

vko

c k P

obsd.

a,b

calcd.

2, obsd.

2.0

calcd.

D2 4

DI+ D 1 . 7 6 11 17 21 50 42 Do+ Oz+ DO*++ D 1.5 0 9 16 22 62 44 From Fig. 2. By eq. 4, identifying O'L and a. Assuming 00 = 0 and ulEt'/t = U . L ~ / U I C .

When a(E) > 0 a t low values of Ee and a(E) = 0 above a critical limit, then for measurements in the interval above this limit the expression QE, = const. must apply. Thus, i t already has been found that Q a E e - ' for the formation of the persistent collision complex. l7 In such instances one expects the complex to be rather unstable and the rate of decomposition to increase rapidly with increasing ion energy. With fixed collection time in a mass spectrometer, this would be expressed by a limiting, critical energy E, beyond which u(E) rapidly approaches zero. If, for simplicity, (14) L. P. Theard and W.H. Hamill, J. Am. Chem. Soc., in press. (15) R. F. Pottie, A. J. Lorquet and W. H. Hamill, ibid., 84,529 (1962). (16) Don Kubose, Andree Lorquet and Thomas Moran, this Laboratory. (17) R. F.Pottie and

W.H. Hamill. J . Phys. Chcm., 63, 877 (1959).

"*

80

.j

Ei

x *o 01

40

20

0

0.0

0.8

0.4

1.2

1.8

En-'". Fig. 1.-Ion-molecule reaction cross section Q as function of ion energy from the work of Stevenson and Schissler.6 From top t o bottom, curves refer to ArH+, D8+and DO*+.

this limit is treated as a discontinuity, then equation 9 also should apply to this case. The data for charge transfer to the 72-ion in neopentane are of this type and all the facts strongly suggest that C6H12+is very unstable. We conclude that i t is susceptible to unimolecular decomposition following charge transfer when a critical ion-molecule relative velocity has been exceeded. We obtain from the preceding equations for ion-induced dipoles, a t fairly low energy

-

(2~iEo-'/* ~ o E o ) / E e ;Eo

< Et

(10)

It is evident that the relation QEe = constant

(11)

will apply regardless of the force law. Charge Exchange.-The qualitative dependence of A&/& upon the ionization potential of the added gas is evident. Concerning fragment ions we can, of course, eliminate a t once all for which the difference of ionization potentials is endothermic (e.g., C2H3+, CzH5+ and C3HT+). Comparison of results for CzH4 and CZHSand their ion abundances strongly indicates CzH4+ as the effective primary ion. This is supported by the results with added n-CaHloand n-C,Dlo but not by those with i-C4Hs and 1-C6He. The disagreement may arise from differences in structure of the CzHd+-ions. The possibility that the increased 7'24011 current in mixtures is due to reaction rather than simple charge exchange was tested with CzDs and with CdD10, for which the sum of all ion currents

734

ALBERT

T. BOTTINIAND

from m/e = 73 to 78 was 0.14 G2 and 0.52 i72 respectively. For comparison, in mixtures with C2H6 and C4H10 the values were 0.23 i72 and 0.45 i72. Also, such a hypothesis would not explain the dependence of Q upon Ee-I. TABLEVI1 ESERGYDEPENDENCE FOR REACTION CROSSSECTIONS 01

Primary

Secondary ion

ion

cs*+ CEH~+~ C2Ha C2DI +" CzH*+ f C3Hs-d CzH, +'

+'

C6Hi2+ CsHii + CsHii CjHiz' CsHiz C5Hizf

+ CjHIi+

X 10:6 cm. e.v,1/2

5.3 30 n

1

+

m

X 1016, cm.2

QEe X 1O216 cm. e.v.

3.1 15 4

1.5 1.5

1.0 csz CsHn 5.3 a For pure neopentane or added propylene. * Combined average for C2H4+and CzD4+ from runs with CzHe or with CzDo. "From CzDs. dCombined average for CzH4and C&+ from C3Hs, assuming equal efficiencies. e From

'

+

CZH6.

+

The combined current, A i72 A i71, should be a measure of the primary process F, provided decomposition does not proceed beyond C5Hll+. In fact equation 8 describes the results very well on this basis, and i t would be expected that the primary process F obey such an energy dependence. CS2+ CaHiz +CSz GHiz+ (F) H C6Hiz+ +CsHii+ (G) The results appear in Table VII. On the other hand, the net A i72obeys equation 11. Considering all the evidence, we conclude that A i 7 2 also obeys the particular law, 10. If one accepts this interpretation, the critical energy for the decomposition

+

[COSTRIBUTION FROM

+

+

THE

CHARLES

P.

NASH

ITOI.

84

reaction G can be obtained from the measured values of the parameters in equations 8 and 10 if we assume that Uk = 0. We have Q E e = - u& = 5.3 X 10-l6 cm.* e.v. and 2u1 = 10.6 X cm.2 e.v.'/z; - 'JkO = (To = 3.1 x cm.2 One obtains E, = 0.39 ev. for C5Hlz+. Finally, using the same parameters we find Et = (Z1.3/1.3)~= 2.9 ev. for C5H1lf and CSH~:!+ combined which places an approximate upper limit on the range of validity of the measurements suitable for testing the preceding equation. The preceding data in Table VI also support the crude description by collision theory mentioned above, viz. ul/uo = U L / U . ~ , assuming that reaction P-factors cancel. For the first three entries in Table VI1 we find ratios of 1.7, 2.0 and 1.8. The corresponding calculated ratios of U L / U K are 1.e5, 1.1 and 1.0. Neopentane.-It appears that no single ionic species is responsible for the enhanced 71-ion formation, either in mixtures or in neopentane alone. Comparison of the results in Table I with the mass spectral patterns suggests that CHS+, C2H6+ and CzH4+ are effective. The parameters U I and uo for the more efficient reactions appear in Table VII. It is significant that every additive which enhances &* also enhances &, while .the converse appears not to hold. This suggests that, in addition to the mechanism previously proposed by Lampe and Field,Io there is also a spontaneous decomposition following charge transfer. The effect is most pronounced for carbon disulfide for which we postulate, as suggested by the data of Table 11, the consecutive reactions F and G. Acknowledgment.-We are grateful to Dr. D. P. Stevenson for critical comment.

DEPARTMEFT O F CHEMISTRY, CSIVERSITY

OF

CALIFORXIA, DAVIS,CALIFORFIA]

The Ultraviolet Spectra of N-Phenyl-substituted Cyclic Imines1 BY ALBERTT. BOTTINIAND CHARLES P. NASH RECEIVED JULY31, 1961 The ultraviolet spectra of N-phenyl-substituted cyclic imines ranging in ring size from 3 t o 6 have been examined. From comparisons of the ultraviolet spectra, base strengths and molar refractions it can be concluded t h a t the resonance energy due t o phenyl ring-nitrogen interaction increases a s the ring size of the imine is varied in the order 3 < 6 < 4 < 5. Comparison of the spectra of these imines with that of S,N-dimethylaniline ( I ) indicates that the configuration about nitrogen in I and the great majority of aromatic amines approaches the pyramidal configuration much more closely than the trigonal configuration and that ~-sp3-conjugationbetween a phenyl ring and an amine nitrogen is as effective as r-?r-conjugation in lowering the energy of the system. The effects of polar and hydrogen-bonding solvents on the spectra of the amines studied are discussed.

Introduction Wepster2has pointed out that the stereochemistry of nitrogen in aromatic amines is complicated, among other reasons, because the optimum configuration is a compromise between the intrinsic tendency of trivalent nitrogen to be pyramidal and the expected tendency in conjugated systems to assume the trigonal configuration in order to maximize the resonance interaction with the benzene ring. Thus, the lone pair of electrons on the

nitrogen atom in an aromatic amine may have sp3character, $-character or some intermediate character, depending upon how the various factors affecting the potential energy of the molecule vary with the valency angles about nitrogen. For example, in aromatic amines in which rotation about the aromatic carbon-nitrogen bond can occur, the valency angles of nitrogen would be expected to increase toward 120' as the "angle of twist" + 3 approaches O', provided that the resonance energy

(1) Acknowledgment is made to the donors of the Petroleum Research Fund, administered by the American Chemical Society, for support of this research. (2) B. M. Wepster, Rec. tiw. chim., 72, 661 (1963).

(3) The angle of twist + i s defined by the plane through the aromatic carbon-nitrogen bond perpendicular to the benzene ring and the plane through the aromatic carbon-nitrogen bond and the axis of symmetry of the lone pair of the nitrogen atqm.