摘要:

Tunable Laser Electronic Spectroscopy BY ROBERT W. FIELD Department of Chemistry, Massachusetts Institute of Technology, Cambridge, Massachusetts 02139, U.S.A. Received 12th January, 198 1 The crucial difference between classical and laser spectroscopies is the capability with the latter of performing separately the usual components of a spectroscopic experiment : sample preparation, signal detection, frequency resolution and measurement, and spectral line assignment. Incompatibi- lities, such as intensity us. resolution, are eliminated by this separation. Various laser experiments will be reviewed as illustrations of solutions to tactical spectroscopic problems. Excitation spectroscopy separates frequency measurement from signal detection, thus facilitating the recording of relatively sparse ( < 5 lines per cm-') spectra at a limiting precision (< cm-') far beyond what is possible with presently available and convenient standards.Selective fluorescence detection eliminates " unnecessary " inter-band and inter-branch blends and simplifies rotational assignment. Doppler-width-free techniques, such as intermodulation and optical-optical double resonance (OODR) spectroscopy, provide additional resolution at small cost in signal strength. OODR combines near-perfect prior state selection, the ability to access " special " levels, and a strong probe in which aN prepared molecules participate at a selectable time, rate and on a single probe transition. Optically pumped lasers (OPL) yield spectroscopic information analogous to OODR, but with one crucial advantage and disadvantage.These are, respectively: the OPL signal appears in a diffraction-limited beam, not over 477 steradians; certain 00DR-observable transitions have insufficient gain for OPL oscillation. Spectra acquired with tunable lasers are ideally suited to computer control over the recording, frequency calibration, and line assignment processes. All of the tedious measurement, trial-and-error and mass data-handling aspects of classical spectroscopy and the every-experiment-is-a-special-case character of high-resolution laser spectroscopy are eliminated. 1. INTRODUCTION This paper is intended as a survey of the tactical strengths of tunable laser tech- niques as applied to the traditional objectives of Molecular Electronic Spectroscopy. I hope that 1 have avoided stressing the obvious, that lasers promise almost embarrass- ing resolution, precision and sensitivity, and that, if only lasers were tunable from r.f.to X-ray, laser-free spectroscopy would be obsolete. By organizing this discussion of classical and laser spectroscopic techniques around an evaluation of the form and strength of the interaction between experimentalist and molecule, I hope to suggest novel ways to attack traditional Molecular Structural problems. In designing a spectroscopic experiment, decisions are made, often implicitly, about how to accomplish six basic objectives : sample preparation, sample interrogation, signal detection, frequency resolution, frequency measurement, and spectral line assignment. It is quite common that an experiment designed to optimize one of these functions unacceptably sacrifices performance of the others.For example, it is almost instinctive to regard resolution and detection as incompatible. However, for certain optical-optical double resonance experiments, detection sensitivity is enhanced as resolution increases.'112 TUNABLE LASER SPECTROSCOPY 2. FROM CLASSICAL SPECTROSCOPY TO LASER EXCITATION SPECTROSCOPY The traditional spectroscopist is a rather passive participant in recording and assigning an absorption or emission spectrum. Although there are elegant sample- preparation tricks at his disposal (e.g., flash photolysis, resonant A* + BC -+ A + BC* energy transfer), it is almost impossible to prepare the molecule of interest in a single, selectable, rovibronic level.Interrogation by absorption is generally such a weak probe that saturation effects are inapplicable. Interrogation by emission is completely passive ; excited molecules in many states fluoresce spontaneously, over 4n steradians, and into many accessible lower levels. Detection, resolution and frequency measurement are inextricably combined and mutually incompatible. Generally, a huge number of spectral lines is obtained and must usually be assigned by tedious trial-and-error searches for consistent combination differences. The rewards for this passivity are that spectra are often easy and quick to record, a large number of resolution elements are sampled simultaneously, and there is no need to await the availability of a suitable laser.Fluorescence spectra excited by an atomic line or a fixed-frequency laser approach perfect prior sample preparation. The resultant spectrum consists entirely of transi- tions from a single e‘, u’, J’ upper level. Such a sparse spectrum need not be recorded at high resolution in order to obtain highly precise term cnergies because lines are symmetric, unblended, and may be recorded at high signal/noise at moderate resolu- tion. Resonance fluorescence spectra consist of nothing but identified combination differences ; thus, in a sense, they are automatically assigned. Detection remains completely passive and the usual intensity-precision incompatibility is unsolved but of diminished importance. In addition, fluorescence must occur at a faster rate than rotation-changing collisions, otherwise the spectral simplicity will be lost.Laser fluorescence excitation spectroscopy (conventionally but inappropriately designated LIF) involves scanning the frequency of a tunable laser and recording the undispersed fluorescence excited by this laser.2 This represents backward steps in the sample preparation and line assignment areas. However, these are insignificant compared to the advantages obtained by permitting detection, resolution and frequency measurement each to be optimized separately. Fluorescence is collected from n steradians (as compared to 71 in a large spectrograph), over an extended volume of sample (much larger than could be imaged onto a 3 x cm2 spectro- graph slit), from transitions into all lower e”, u”, J” levels simultaneously, and, with pulsed excitation, time-resolved detection can provide useful spectrum-sorting functions. Resolution is limited by the molecular Doppler width (Am,,) or the laser bandwidth (Am,), whichever is larger, thus resolution can, in principle, be increased to the Doppler limit at no cost in signal strength.Frequency measurement is accomplished by comparing the laser frequency at line centre (precision limited to Am, divided by the signal/noise ratio) to an external standard such as a digital wave- meter,4 an interferometer, or the 1, atlas measured by Fourier-transform spectro~copy.~ Without sacrificing the advantageous separation of detection from resolution and frequency measurement, the interaction between experimentalist and molecule can be strengthened beyond that of excitation spectroscopy in either of two directions.Sections 3 and 5 are discussions of multiple-resonance schemes which optimize the state-selectivity of sample preparation and/or interrogation. Alternatively, experi- ments with optically pumped lasers, described in Section 4, maximize the coupling of laser-excited molecules to a detector.R. W. FIELD 113 3. DOUBLE RESONANCE The central idea is that signals are detected only from molecules which satisfy two resonance conditions: state-selection (PUMP) and state analysis (PROBE). Multiple- resonance techniques are most useful for sorting out extremely complex spectra,' gaining access into some special class of energy levels,69 or attaining sub-Doppler res~lution.'*~*~ The most serious defects of laser excitation spectroscopy are its Doppler-limited resolution and the difficulty of rotationally assigning excitation features.In this section, four techniques which are free of either or both of these defects are discussed: selective fluorescence detected excitation ~ p e c t r ~ ~ c ~ p y , ~ ~ ' ~ sub-Doppler fluorescence spectr~scopy,~ intermodulated fluorescence and optical-optical double resonance (00DR).'96y8 The most effective way of rotationally assigning laser excitation spectra involves examination of the spectrum of fluorescence associated with each excitation feature.I3 In a sense, fluorescence is the " probe " in a double resonance scheme. It provides the combination difference that would otherwise be sought by trial and error, without interfering with the ability to measure accurate frequencies of excitation features.An important point is that high precision is not required for this combination differ- ence.l3'l4 Precision data are obtained from the excitation spectrum ; qualitative supplementary information guides a rapid but tentative line-assigning procedure in which definitive assignments are based only on precisely measured line separations from the excitation spectrum. It is tedious to examine the fluorescence spectrum associated with each excitation feature. Selective fluorescence detected excitation s p e c t r ~ s c o p y ~ ~ ~ ~ enables the recording of an excitation spectrum in which almost every feature is explicitly associ- ated, through a narrow range of known and experimentally adjustable lower state combination differences, with another feature observable on the same or a separately recorded selective excitation spectrum.For example, in a lC-'C system, excitation of a u', J ' level results in a progression of fluorescence doublets u", J" = J' & I . If a monochromator is tuned to the R(J' - 1) line of the strongest u ' 4 ' band, then every excitation feature associated with the u', J' level (regardless of u " ) will be selectively detected. Since the mono- chromator band pass must be set large enough that useful signal is detected, excitations of levels a few J-units on either side of R(J' - 1) and P(J' + 1) lines of the u' level will be detected. In addition, a few extraneous excitation lines which produce fluorescence accidentally within the selected bandpass might be detected, but these can be elimin- ated easily by recording another excitation spectrum with the monochromator tuned to the P(J' = 1) ur-u'' line.Fig. 1 illustrates the spectral simplification achieved by this technique as applied to a 1 cm-' segment of the Au = 0 region of the CaBr B2C+-X2C+ sy~tem.'~ The lower trace, even though the resolution is limited only by the Doppler width of CaBr molecules at 500 K, contains an unassignable jumble of incompletely resolved lines belonging to Ca79Br and CaslBr, 0-0 through 5-5 vibrational bands, and both R, and R2 branches. When a monochromator is used selectively to detect fluorescence from a 2 cm-l region of the P, branch of the 0-0 band, the upper trace simplifies into groups of 4 unblended and trivially assignable lines [0-0 R,(J) and 1-1 R,(J -1- 1) for the two Br isotopes].Selective fluorescence detected excitation spectroscopy combines the best features of excitation and resolved fluorescence spectroscopy at a cost of 102-103 in detected fluorescence intensity. It represents a compromise between incompletely separated114 TUNABLE LASER SPECTROSCOPY assignment and detection functions. In addition, by having the ability to select which small subset of lines in a congested spectral region are to be recorded on a given excitation scan, sub-Doppler line-blending and obscuration of critical weak (low-J, forbidden sub-band, perturbed) lines can be eliminated. The frequency measurement function is serendipitously optimized by this compromise between assignment and detection ; line frequencies in a selectively detected, blend-free excitation spectrum may be measured more accurately than in ordinary excitation or resolved fluorescence spectra.15 Recently, multiphoton ionization spectroscopy (MPI) has received considerable attention l6 because of its > lo2 higher detection sensitivity than fluorescence excitation v/cm-' 16 396.0 16 395.5 I r' I I I I 1 I I 1 C O ' ~ B ~ Ca "8r co 79Br ca *'Br 4 8 .5 (0-0) 48,5(0-0) 49.5(0-0) 49.5 (0-0) 0 FIG. 1 .-Selective fluorescence detected excitation spectroscopy. A 0.8 cm-' segment of the CaBr B2X+-X2Z+ excitation spectrum in the Av = 0 region is shown. No pattern is recognizable in the lower trace for the numerous incompletely resolved lines recorded by ordinary excitation spec- troscopy.Note that the selectively detected spectrum simplifies into groups of 4 lines belonging to similar N-values of the same rotational branch, a situation which guarantees that rotational analysis of an ordinary absorption spectrum by trial-and-error would be impossible. spectroscopy. It is important to point out that the cost of this optimization of signal detection will be a return to trial-and-error line assignment, congested spectra and limited resolution. The spectral information contained in the fluorescence is often too valuable to be discarded. In certain situations, selectivity detected excitation spectroscopy is capable of inadequate resolution. Sub-Doppler resolution is commonly achieved by one of three schemes : forward or backward scattered resolved fluore~cence,~ various one- laser non-linear spectroscopies". 7-19 and OODR.'~* If fluorescence is viewed co- or counter-propagating with sub-Doppler excitation radiation, then the fluorescing molecules are velocity selected along the detection line- of-sight.The fluorescence will be Doppler-shifted but not Doppler-broadened.' Consider an excitation line, wE, consisting of two strong A F = AJ hyperfine com- ponents separated by 6. Let the corresponding rest-frame separation of these two components be E in the fluorescence line, wF. If the excitation laser is tuned to the rest frequency of the lower frequency component of coE, then LO^-* observed in the forward/backward direction will consist of two components at toF+ = coF and (LO, + E ) 1 - .( 3R. W. FIELD 115 If wF w coE and 6 N" e, then the two h.f.s.-Doppler components coalesce in the forward direction and are split ca. 2e in the backward direction. In order to take advantage of this scheme for elimination of Doppler broadening, a spectrograph with resolving power considerably in excess of 1 in lo6 is required. Again, resolution and detection are in opposition, although sub-Doppler experiments using Fourier transform' and hybrid grating-interferometric spectrometers have been highly successful.20 A difficulty unique to sub-Doppler Fourier-transform fluores- cence spectroscopy is that short- and long-term frequency or amplitude variations of the exciting laser will seriously degrade both resolution and ~ensitivity.~ The resolution and detection functions are partially separated by one-laser non-linear spectroscopies, but large non-resonant background signals, stringent laser intensity requirements, and sampling of a small fraction of available molecules often restrict the applicability of these technique~.'~~'~-'~ Although coherent two- photon19 and polarization spectroscopies '' offer advantages over intermodulated fluorescence spectroscopy," only the latter technique will be discussed here.Intermodulated and sub-Doppler fluorescence spectroscopies are functionally very similar. The pump beam prepares state and velocity selected upper and lower level populations. In fluorescence spectroscopy the upper level populations are passively probed by forward or backward spontaneous fluorescence.The high-resolution information is present on all fluorescence lines. Alternatively, the upper and lower level populations may be actively probed by the counter-propagating probe beam. The high-resolution information for the single probed transition is present on the total undispersed spontaneous side-fluorescence because of a competition between stimu- lated emission and absorption. The pump and probe beams, modulated at two different audio frequencies, col and co2, cause a small fraction (typically 1%) of the side-fluorescence to be modulated at CL)' & co2 when the two beams compete for the same molecular populations. The intermodulation signal occurs either at the rest frequency or at the midpoint between the rest frequencies of two transitions which share a common level and are separated by less than a Doppler width.21 Whenever a process is probed by stimulated absorption/emission rather than spontaneous fluorescence, the experimenter has access to a wide range of schemes for increasing the coupling between molecule and detector.Interrogation can be made more specific without loss of signal strength. This is why the total fluorescence intensity required for observable intermodulation signals is many orders of magnitude lower than that required to record sub-Doppler fluorescence. It could also provide a basis for putting the entire intermodulation signal onto a diffraction limited probe beam. The reason why intermodulated absorption spectroscopy is not widely exploited is that the effect would be overwhelmed by noise associated with fluctuations of the large and inseparable probe backgr~und,'~~'' a problem not shared by the output of a laser-excited, optically pumped laser.22 The flexibility of intermodulation spectroscopy is insufficient to illustrate the power of a stimulated us.spontaneous probe. OODR, by completely separating the sample preparation and interrogation functions (as well as the detection, resolution, measure- ment, and assignment functions), is maximally flexible. The pump laser excites e', u', JI t e", u", J " , the probe laser excites e*, u", J* +- e', u', J ; , and the effect is detected as either an appearance of fluorescence near the sum of the two laser frequencies or a decrease in fluorescence from e', u', Jf. The longitudinally velocity-selected population in e', u', J ; may be probed in a short time compared to the radiative or collisional lifetime of the e', u', JI, level.' By scanning the probe laser further, the velocity distribution of the population transferred from level J ( to J f by a single collision may also be probed.* Fig.2 illustrates OODR.'116 TUNABLE LASER SPECTROSCOPY The resolution of OODR excitation spectroscopy is set by the bandwidth of the two lasers and the pressure- and power-broadened width of the probe t r a n s i t i ~ n . ' * ~ * ~ ~ The S/N of OODR is increased by more than a factor of 10 when two 1 W, 1 cm-' bandwidth lasers are replaced with two 50 mW, cm-' lasers.' The time-scale of an OODR probe is set by the probe or, for pulsed OODR, by the pump-probe time delay,24 not by the radiative rate.All of the molecules excited into e', u', J; are capable of contributing simultaneously to the OODR signal because the P U .I 5 -$ u h -c1 .I rn 5 s k + -c1 .I + W + w u - Second Excitation R (0) v / = I , f : O Resolved P( 1 ) l o i 4 1 + 200 +I00 V O - 100 - 200 laser frequency/ M Hz FIG. 2.-Alternative detection schemes for sub-Doppler optical-optical double resonance. The pump laser excites BaO A-X 1-0 P(1), the probe laser excites C-A 3-1 R(0). The lower trace is detected through a U.V. passing-visible absorbing coloured-glass filter. The instrumentally broadened upper trace is obtained by monitoring a decrease in the resolved fluorescence intensity of the A-X 1-2 P(1) line. The signal-to-noise ratio is inferior for the upper trace because of the < lo3 lower intensity for resolved us.filtered fluorescence. probe laser stimulates a single, homogeneously broadened transition. The OODR signal is carried by undispersed spontaneous fluorescence without any background. Spectra are trivially assigned because they consist of identified c*, u* combination differences. OODR is 10% as sensitive as fluorescence excitation spectroscopy,' but is capableR. W. FIELD 117 of > lo2 higher resolution,* automatic line assignment, freedom from spectral con- gestion and systematic exploration or exploitation of special levels.6*23*25 When comparing OODR to resolved fluorescence spectroscopy, its 3, lo2 advantages in both sensitivity and resolution make it ideally suited to detailed studies of state-to- state collisional p r o ~ e s s e s .~ J ~ J ~ Fig. 3 illustrates an as yet unexploited spectrum-sorting feature of stimulated- x c) .I E E c1 c aJ .3 liJ 0 G P2 (51) I i; P4 (51) I 16 753 BOO 16 753 600 16 753 400 16 753.200 16 7.43 600 16 743.400 16 743 200 16 738600 16 738 400 16 738.200 laser energy/cm- ’ FIG. 3.-Differential power broadening in OODR. The pump laser excites BaO A-X 1-0 R(50); the probe laser is scanned in the P(51) region of C-A 3-1. The main line is power-broadened to 121 MHz. A weak “‘extra-line ” [verified by scanning in the R(49) region to be a J * = 50 f J’ = 51 transition] is broadened to 47 MHz. The area under this weak sub-Doppler line is equal to that of an unassigned, fully Doppler-broadened, collisional satellite line labelled by an arrow.probe schemes. The ability to saturate a transition depends on its transition matrix element. An allowed transition will power-broaden faster than a forbidden one. Thus, in OODR, forbidden transitions can be made to appear considerably narrower than allowed transitions, but with nearly identical peak height. An intermodulation spectrum could be recorded at such high probe intensity that strong lines are so severely broadened that only sharp, weak lines would be detected. Aside from its requirement of two tunable lasers, the only shortcoming of OODR from the strength-of-probe point of view is that it wastes signal strength by failing to obtain a signal from all probed molecules, which could be directed as a diffraction- limited beam onto a detector.22 4.OPTICALLY PUMPED LASERS It is surprising that optically pumped lasers (OPL) have rarely been exploited to obtain spectroscopic or kinetic information. One is tempted to dismiss them as exotic devices with plausible but unlikely practical applications. Yet OPLs are extremely simple in principle and usually in practice. More importantly, they provide an approach to optimal coupling between molecule and experimentalist. OPLs, when used to generate spectroscopic information, are analogous to OODR, with the advantage that the system of interest provides its own probe laser. A pump laser prepares a longitudinally velocity-selected population in a single e’, u’, J’ level. This population is larger than the thermal population of most zl” > 0 levels, thus significant gain exists for transitions into a wide range of lower levels.Since mirrors with high reflectivity at any wavelength to the red of the pump are much more readily available than a probe laser which is tunable throughout this region, there is a con- siderable tactical advantGge in asking the system to probe itself. An OPL is an extremely simple device, consisting of a pump laser, a means of inserting the pump beam into the spatial mode defined by the OPL resonator mirrors,118 TUNABLE LASER SPECTROSCOPY and an optional intracavity fine-tuning element. Coarse tuning of the OPL is obtained simply by selecting mirrors with high reflectivity in the wavelength region of desired oscillation. Thus one combines perfect prior state selection, the possibility of a strong, sub-Doppler probe and ability to probe wavelength regions free of probe- laser availability restrictions.Fig. 4 illustrates the range of levels accessible using a fixed-frequency 2 0000 16000 10000 5000 0 probe laser 2.2 3.0 4 -0 5.0 RIA FIG. 4.-1, BO; - X’Z: optically pumped laser. Each set of resonator mirrors allows observation of lasing transitions from a single upper level into lower levels typically spanning about 10 vibrational quanta. (Ar+ 514.5 and 501.7 nni lines). More than 700 I2 BO:-X’X:,+ transitions are ob- served and assigned, covering the spectral region 570-1340 nm and ranging from u” = 9 to 96.2s*29 With a tunable pump laser, more than lo6 I2 B-X OPL lines in the 500-1350 nm region are accessible. The untuned OPL output is in the form of a diffraction-limited beam, the spectrum of which is similar to a spontaneous fluorescence spectrum.One difference is that the OPL spectrum shows no satellite lines arising from J ; --f J ; collisions in the B 0: state. Molecules in a single e’, u‘, J ’ level can be transferred by self-stimulated emission into a small number of e”, u”, J” levels. This is manifest in the >20% OPL quantum efficiency (laser photons out/pump photons abs~rbed)~’ and the >30% decrease in side-fluorescence intensity when laser oscillation is allowed to occur.31 The most important distinction between OPL and fluorescence spectroscopy is that the output of an OPL is neither divided over many Franck-Condon allowed transi- tions nor dispersed over 4n sr.This means that extremely high-resolution spectral analysis is possible at no cost in detection sensitivity. The most serious limitation is that many interesting transitions will not have sufficient gain to oscillate. Fortunately, many such transitions will be observable by pulsed OODR.24 Since the OPL upper-level population is prepared and probed collinearly, the OPLR. W. FIELD 119 output will be Doppler-shifted but not Doppler-broadened. A method for obtaining sub-Doppler spectral information, based simply on a scan of the OPL-cavity length,22 illustrates both a strength of stimulated probe schemes and a non-trivial obstacle to the easy applicability of OPL techniques to spectroscopic and kinetic problems. The e’, v’, J’ molecules in the OPL resonator are longitudinally velocity selected.In a linear resonator, forward and backward gains are comparable, but not necessarily at the same frequency. (See discussion of sub-Doppler fluorescence spectroscopy in Section 3.) However, in order for laser oscillation to occur, the length of the OPL resonator, L, must be such that 2L = nA, where iz is the lasting wavelength and n is an integer. Fig. 5 shows that the OPL output turns off and on as the resonator length is scanned and that this pattern varies as the pump laser is tuned across the Doppler and hyperfine broadened pump transi- tion. This scheme for obtaining sub-Doppler information is more sensitive and con- ceptually simpler than forward or backward scattered fluorescence spectro~copy.~ However, in order to obtain spectra similar to those shown in fig.5, both the pump laser and the OPL resonator had to be carefully stabilized by standard but elaborate servo techniques.22 The inconvenience associated with the need for servo-stabilized resonators is only the tip of the iceberg. Fig. 5 indirectly illustrates that the gain profile of an OPL is extremely complicated, consisting of numerous sharp features each of which is sensitive in a different way to power broadening by the pump laser and by the various axial OPL modes which are ~ s c i l l a t i n g . ~ ~ Without careful control over experimental conditions, the signal-to-noise ratio, resolution, precision and meaning of relative intensities in OPL experiments will be seriously degraded. Intracavity experiments are inherently nonlinear ; this is both their strength and weakness.The relationship between OPL output intensity, Einstein A and B co- efficients, and upper and lower laser level populations is c ~ m p l e x . ~ ~ - ~ ~ Nevertheless, it appears possible to use OPLs to measure ratios of &coefficients for a series of transitions from a common upper as well as ratios of total collisiorial removal rates from non-fluorescing lower rovibronic levels of lasing transition^.^' Such schemes rest on the plausible but probably invalid assumption that, by making measurements on OPL transitions from a common upper level, at identical pump intensity and frequency, and at constant pressure in the gain cell, all of the complicated gain lineshape effects will cancel out of the ratios of measured radiative or collisional rates.To summarize, the strengths of OPLs are that they combine perfect prior state selection with a strong probe which couples efficiently to a detector. However, the strong probe is highly non-linear because of the restrictions on lasing frequency imposed by the OPL cavity length and the complicated effects of pump and OPL radiation fields on the above-threshold portion of the gain profile at a given OPL-resonator mode. The sample interrogation function is indirectly in conflict with the signal detection, frequency resolution and frequency measurement functions. It is not possible to interrogate a transition that will not oscillate in an OPL. It is difficult to measure line centre of a lasing transition for which the lineshape has a complex dependence on many pump and OPL parameters.Conversely, detected intensities have little quantitative significance One is faced with a choice between development of elaborate OPL experimental schemes and lineshape theories or elimination of the OPL resonator and a return to strong-probe OODR schemes. The width of each feature is less than 2 M H z . ~ ~120 TUNABLE LASER SPECTROSCOPY Av= +I0 MHz h Av = + 20 MHz A v = -10 MHz Av = - 2 0 MZ FIG. 5.--Sub-Doppler spectroscopy with an optically pumped laser. Five spectra are shown, each obtained by scanning the OPLcavity length and detecting total I2 BO+-XlC: 43-83 P(13) OPL output. Each spectrum was recorded at the specified pump laser frequency-offset from the al h.f.s. component of the 43-0 P(13).Lines are labelled according to hyperfine component (a,) and whether the line resulted from gain in the forward (F) or backward (B) direction. 5. MULTIPLE RESONANCE After double resonance comes triple resonance! The idea of a strong interaction between molecule and experimentalist is intoxicating. One is tempted to pose questions that cannot be answered by single or double resonance schemes. Three experimental schemes will be described here : intermodulated fluorescence spectro- scopy combined with selective fluorescence detection, modulated gain spectroscopy and stimulated-emission pumping. Intermodulated fluorescence provides high resolution but little spectrum sorting. Excitation spectroscopy with selective fluorescence detection allows poorer resolution but optimal spectrum-sorting. If one wishes to measure the hyperfine structure of a rare isotope in the presence of an abundant isotope, one is faced with the problem of locating weak lines in the presence of numerous strong and weak lines associated with the unwanted species.Since the rare and abundant isotopic molecules will have identical transition strengths, differential power broadeningz3 or time-resolved3 schemes will be of no avail. An in-series combination of intermodulated and selective fluorescence would exhibit a poor signal-to-noise ratio because of the respective lo-* and However, an in-parallel combination would be effective and sensitive. Sub-Doppler intermodul- ated broad-band detected fluorescence and Doppler-limited selectively detected signal detection factors associated with the two techniques.R.W . FIELD 121 fluorescence excitation spectra would be recorded simultaneously. The former would provide the resolution, the latter partial spectrum sorting. Small Franck-Condon factors usually prevent the examination of vibrational levels near their electronic dissociation asymptote. OODR schemes can access vibrational levels with turning points spanning a wider range of internuclear distance because the second transition can originate from either inner or outer turning point regions of the intermediate e', u', J' level. A three-step excitation process would access an even wider turning point range. An Ar+ laser pumped Na,B'II,-X'C,+ laser can be tuned to oscillate on transitions into high vibrational levels (u" > 40) of X'C,+, thus providing a steady-state source of significant population in a single, thermally unpopulated, rovibronic level with its outer turning point at very large internuclear distance ( R > 6 A).33 A C.W.dye laser is then used to excite the OPL-prepared population into levels of the A'&,+ and B'n, states near their respective dissociation limits. Since the Ar+ laser, the Na, laser, and the dye laser all excite Na, fluorescence (some of which is reabsorbed, thus exciting secondary fluorescence), the Na, spontaneous fluorescence is useless for detecting the desired dye laser excitations out of u" > 0 levels. However, when the probe dye laser pumps population out of the lower level of the OPL transition, the population inversion density and the OPL output intensity are increased.The OPL performs both sample preparation and signal detection functions. Sub-Doppler transitions (ca. 200 MHz) into Na, AIC,+ levels up to u' = 62 have been The two above " triple resonance " schemes require only one continuously tunable laser. Stimulated Emission Pumping (SEP) requires as many as three tunable lasers. The purpose of SEP is to prepare large populations in single, highly excited (>2 eV) vibrational levels of small polyatomic molecules in their electronic ground state. The sample preparation function in SEP is accomplished using two sequenti- ally fired, pulsed dye lasers. The molecule is first pumped up to an electronically excited level and then stimulated down into the desired level of the electronic ground state.The reason for this circuitous path is to take advantage of molecular geometry differences between electronic states, thereby facilitating preparation of highly and specifically distorted vibrational levels. Typically loo//: of the initial level population is transferred, through the intermediate level, into the desired final level.*" SEP is detected as a resonant decrease in fluorescence intensity from the intermediate level. SEP will provide detailed and unambiguous information about the structural and dynamical properties of individual rovibronic levels, free of complications arising from interactions between sparse and dense manifolds of vibronic levels.34 Once structur- ally interesting levels are identified through two-laser SEP spectroscopy, their uni- molecular and bimolecular dynamical properties may be examined using a third, time-delayed, probe laser.SEP is a folded, pulsed variant of OODR which has the capability of combining the most important features of OODR and OPL techniques. All of the standard functions of a spectroscopic experiment may be independently performed. In addition, the most critical component of dynainical experiments, temporal control, may be optimized without restriction by the relatively long radiative lifetime of the target level. The reader will have to judge whether the end justifies the means. 6. COMPUTER CONTROL Modern tunable lasers are readily adapted to computer control.35 Huge, highly It is important to con- precise, spectroscopic data sets may be rapidly generated.122 TUNABLE LASER SPECTROSCOPY sider computer strategies which will provide optimal frequency calibration, rapid line- assignment, and minimal operator-handling of the spectra, lists of spectral lines and preliminary or effective molecular constants.Two approaches have been suggested: digital wave meter^^ and Fourier Transform spectral atlases.’ It remains to be demon- strated which technique is capable of the most reliable and rapid application at the limiting precision < Frequency calibration is a critical area. cm-’ typical of C.W. dye-laser spectroscopy. 7. SUMMARY A variety of spectroscopic techniques has been discussed from a tactical point of view which stresses the importance of separating preparation, detection, resolution, frequency measurement, and assignment functions.Contrary to naive expectations, many of the usual incompatibilities between spectroscopic functions are avoidable. As schemes to eliminate such incompatibilities are developed, the interaction between molecule and experimentalist is strengthened. As a result, experiments may be designed to answer definitively and rapidly virtually any problem in the electronic spectroscopy of small molecules. A sobering thought is that it now becomes im- perative for molecular spectroscopists to ask the question: Is this the most interesting spectroscopic problem capable of solution with available technology ? I am grateful to many colleagues for reducing to practice many of the techniques described here. Special thanks go to Rick Gottscho (OODR), Brooke Koffend (OPL), Carter Kittrell (SEP), and Roger Back (FTS, OODR, OPL).This research has been supported by grants from the National Science Foundation (CHE-78-18427 and 10178) and the Air Force Office of Scientific Research (AFOSR-80-0254). R. A, Gottscho, P. S. Weiss, R. W. Field and J. G. Pruett, J. Mol. Spectt-osc., 1980, 82, 283. A. Schultz, H. W. Cruse and R. N. Zare, J. Chetii. Phys., 1972, 57, 1354; R. N. Zare and P. J . Dagdigian, Science, 1974, 185, 739. J. G. Pruett and R. N. Zare, J. Chetn. Phys., 1975, 62, 2050. F. V. Kowalski, R. T. Hawkins and A. L. Schawlow, J . Opt. SOC. An?., 1976,66,965; J. L. Hall and S. A. Lee, Appf. Phys. Lett., 1976, 29, 367. S. Gerstenkorn and P. Luc, Atlas dii Spectre d’Absurptioti de fa Mole‘cufe d’lode (CNRS, Paris, 1978); S.Gerstenkorn and P. Luc, Rev. Phys. Appf., 1979, 14, 791. R. A. Gottscho, J. Chetii. Phys., 1979, 70, 3554. R. Bacis, S. Churassy, R. W. Field, J. B. Koffend and J. Vergits, J. Chetii. Phys., 1980, 72, 34. R. A. Gottscho, R. W. Field, R. Bacis and S. J. Silvers, J. Cheriz. Phys., 1980, 73, 599. W. Demtroder, Case Stirdies iri Atotiric Physics, ed. M. R. C . McDowell and E. W. Daniels (North-Holland, Amsterdam, 1976) vol. 6; C. Linton, J. Mof. Spectrosc., 1978, 69, 351. lo M. Dulick, P. F. Bernath and R. W. Field, Cart. J. Phys., 1980, 58, 703. l1 M. S. Sorem and A. L. Schawlow, Opt. Curiimti., 1972, 5, 148; C. Freed and A. Javan, Appf. l2 P. F. Bernath, P. G. Cummins and R. W. Field, Chetii. Phys. Lett., 1980, 70, 618. l3 C. G. Stevens, M.W. Swagel, R. Wallace and R. N. Zare, Cherii. Phys. Lett., 1973, 18, 465. l4 R. W. Field, T. Tanaka and D. 0. Harris, J. Mof. Spectrosc., 1975, 57, 107. l5 P. F. Bernath, R. W. Field, B. Pinchemel, Y. Lefebvre and J. Schamps, J . Mof. Spectrusc., l6 P. M. Johnson, M. R. Berman and D. Zakheim, J. Chetn. Phys., 1975, 62, 2500. l7 A. L. Schawlow, Scietzce, 1978, 202, 141. Phys. Lett., 1970, 17, 53. 198 1 , in press. C. Wieman and T. W. Hansch, Phys. Rev. Lett., 1976, 36, 1170; R. Teets, R. Feinberg, T. W. Hansch and A. L. Schawlow, Phys. Rev. Lett., 1976, 37, 683. l9 L. S. Vasilenki, V. P. Chebotaev, and A. V. Shishaev, JETP Lett., 1970, 12, 1 1 3 ; F. Biraben, B. Cagnac, and G. Grynberg, Phys. Rev. Lett., 1974, 32, 643; M. D. Levenson and N. Bloem- bergen, Phys. Rev. Lett., 1974, 32, 645. 2o H. G. Berry and R. Bacis, Phys. Rev. A, 1973, 8, 36; R. Bacis, R. Collomb and N. Bessis, Phys. Rev. A, 1973, 8, 2255.R. W . FIELD 123 21 E. E. Uzgiris, J. L. Hall and R. L. Barger, Phys. Reu. Lett., 1971, 26, 289; T. W. Hansch, 22 J. B. Koffend, S. Goldstein, R. Bacis, R. W. Field and S. Ezekiel, Phys. Rev. Lett., 1978, 41, 23 R. A. Gottscho and R. W. Field, Chem. Phys. Lett., 1978, 60, 65. 24 C. Kittrell, E. Abrarnson, S. McDonald, D. E. Reisner, R. W. Field, J. L. Kinsey and D. Kata- 25 R. A. Gottscho, J. B. Koffend and R. W. Field, J . Mol. Spectrosc., 1980, 82, 310. 26 S. J. Silvers, R. W. Field and R. A. Gottscho, J . Chem. Phys., 1981, in press. 27 P. C. F. Ip, P. F. Bernath and R. W. Field, J . Mol. Spectrosc., 1981, in press. 2 8 J. B. Koffend, R. Bacis and R. W. Field, J . Mol. Spectrosc., 1979, 77, 202. 29 J. B. Koffend and R. W. Field, J. Appl. Phys., 1977, 48, 4468. 30 J. B. Koffend, P h D . Thesis (MIT, 1978). 31 J. B. Koffend, F. J. Wodarczyk, R. Bacis and R. W. Field, J . Cheiv. Phys., 1980,72,478. 32 J. B. Koffend, R. Bacis and R. W. Field, J . Chein. Phys., 1979, 70, 2366. 33 H. S. Schweda, G. K. Chawla and R. W. Field, unpublished. 34 M. Bixon and J. Jortner, J . Chem. Phys., 1968, 48, 71 5. 35 C. R. Pollock, J. Kasper, G. K . Ernst, W. E. Ernst, S. Blit and F. K. Tittel, Appl. Opt., 1979, I. S. Shahin and A. L. Schawlow, Phys. Rev. Lett., 1971, 27, 707. 1040. yama, Phys. Rev. Lerr., 1981, in press. 18, 1907.

ISSN:0301-7249    

DOI:10.1039/DC9817100111

出版商:RSC    

年代:1981

数据来源: RSC

摘要:

Magnetic-field and Time-resolved Studies of the Electronic Spectrum of HNO BY RICHARD N. DIXON, MARCUS NOBLE AND CAROLINE A. TAYLOR School of Chemistry, University of Bristol, Cantock’s Close, Bristol BS8 ITS AND MICHEL DELHOUME Paris Observatory, 92190 Meudon, France Receired 29th December, 1980 High-resolution laser-induced fluorescence excitation spectra of HNO have been recorded using both C.W. and pulsed dye lasers. These spectra reveal numerous frequency perturbations, often asso- ciated with marked intensity anomalies. All the branches studied were found to be perturbed to some degree with little regularity to the perturbations. The widths of some of the perturbed lines, and of many apparently unperturbed lines, have been found to be sensitive to a magnetic field. Time- resolved measurements yield a mean zero-pressure excited-state lifetime of 23 ps, and collisional quenching rates varying from (1.1 to 15) x These studies reveal that the intensity anomalies arise through selective higher quenching of perturbed levels, coupled with fast rotational-energy transfer.These aspects are discussed in terms of interactions between the levels of the excited 2, ’A” state and high levels of the ground 2, ‘A’ state, with further perturbations from the d , 3A” state. cm3 molecule-’ s-’. The 2, lA’’-T, ‘A’ band system of HNO in the red and near-infrared regions of the spectrum was first observed by Dalby in absorption using the technique of flash photolysis.’ The bands have the simple appearance of type-c bands of a near prolate asymmetric top, although in his analysis Dalby noted a number of weak perturbations of the rotational structure.Bancroft et a/.’ extended this analysis to a total of 10 bands of HNO and 9 bands of DNO, and noted further perturbations. This band system is also known in eniission from chemiluininescent The emission spectrum exhibits breaking-off due to predissociation to H + NO. In a recent study using laser-induced fluorescence of HNO we have shown’ that at threshold this pre- dissociation occurs through Coriolis coupling of the levels of the A, lA” state to high levels of the ground f, ‘A’ state. Analytical potentials for the 2, 2 and a”, 3A’’ states were derived from spectroscopic, thermochemical and quantum-theoretic data as part of that study. It is now accepted that interstate perturbations play a critical role in the time evolution of the excited states of polyatoinic molecules through the processes of inter- nal conversion, intersystem crossing and predissociation.Much has been learned about these decay routes in small molecules through very many studies of excited- state lifetimes8 and fluorescence quantum yields’ of H,CO (and D,CO), which is the lightest stable polyatomic molecule with a well-characterised sharp spectrum. Even so these processes are not yet capable of quantitative interpretation. HNO is iso- electronic with H,CO, but possesses fewer vibrational degrees of freedom, and there- fore has a much lower density of vibrational states for a given energy. Its dynamic126 ELECTRONIC SPECTRUM OF HNO behaviour should therefore be intermediate between that of formaldehyde and that typical of diatomic molecules. This paper presents improved higher-resolution analyses of those bands of HNO between 641 and 578 nm that can be excited to laser-induced fluorescence. Selected portions of this spectrum have been studied in magnetic fields up t o 10.5 kG.Time- resolved measurements over a range of pressures have then been made for a number of normal levels, perturbed levels and magnetically active perturbed levels. These new observations are discussed in terms of the properties of the perturbing states. EXPERIMENTAL In view of the known quenching behaviour of the x, 'A" state of HN05g6 it is important that a study of excited-state dynamics should include measurements in a collison-free regime if at all possible.Kirby et a/.'' have pioneered the generation of HNO by pyrolysis of its adduct with 9,1O-dimethylanthracene, and have used this in a study of its microwave spec- trum." We have used this method to produce HNO at pressures of 0.5 to ca. 200 mTorr, the products of the 70-80 "C pyrolysis being slowly pumped through the fluorescence cells. A parallel study using photoelectron spectroscopy showed that the gas mixture prepared in this way contained considerable fractions of the disproportionation products N20 and H20. Pressure measurements were made using a capacitance manometer, or at the lowest pressures a Penning gauge calibrated at higher pressure against the capacitance manometer. A Coherent CR 490 C.W. dye laser pumped by an argon-ion laser was used to survey the whole spectrum over the ranges of the dyes Rhodamine 6G and Sulforhodamine B at a resolution of 0.5 cm-'.Selected regions were then recorded at a resolution of 0.1 cm-I by inserting a 0.5 mm fused quartz etalon in the cavity and driving this under microprocessor control in synchronisation with the coarser Lyot birefringent filter. Calibration was achieved using the fluorescence of I2 as a secondary standard,'* and the transmission fringes of a 4 mm fused quartz Fabry-Perot etalon (mean spacing 0.847 cm-') as a means of interpolation between iodine lines. The absolute accuracy of measurement of unblended lines is considered to be $0.01 cm-'. A search was made using this laser for magnetic-field activity in fields up to 10.5 kG, attention being concentrated on those regions where perturbations were apparent.A search was also made for electric-field effects using fields up to 3.5 kV cm- ' : no significant electric-field effects on line profiles or intensities were observed at this resolution, although we have been able to measure the excited-state dipole moment of HNO by? DoppleL-free te~hnique.'~ (The dipole moment changes very little on excitation from the X to the A state.) Lines showing magnetic activity were then examined at the Doppler-limited resolution of ca, 1 GHz (0.035 cm-') using a single-mode Coherent CR 599-21 C.W. dye laser pumped by a krypton-ion laser. The dispersions of the scans with this laser were calibrated by simul- taneously recording the transmission fringes of a 16 cm confocal interferometer, with fringe separation 468 MHz.A particular study has been made of the K' = 4 - K" = 3 sub-band of the 100-000 band, which is the last sub-band with extended J-structure below the predissocia- tion threshold. The time-resolved studies were carried out using a home-built dye laser of Hansch design,I4 with oscillator and amplifier stages, pumped by an Oxford Lasers excimer laser operating on XeCl. With a 5 mm air-gap etalon and pressure tuning this exhibited a line width of ca. 2 GHz (0.07 cm-') and very little background superadiance. This width is only about twice the Doppler width, so that it was possible to make separate studies of the time-evolution of the fluorescence with excitation in various closely resolved components of perturbed lines.The time resolution of these measurements was limited to ca. 0.5 ,us by a combination of the photomultiplier and pre-amplifier responses and by the transient recorder (Datalab DL 920). Signal-averaging was accomplished using a PET microcomputer which also controlled the operation of the laser system. Three tunable laser systems have been used in this work.R . N. DIXON, M. NOBLE, c. A. TAYLOR AND M. DELHOUME 127 RESULTS ASSIGNMENTS Fig. 1 compares a low-resolution LIF excitation spectrum of the 4-3 sub-band of the 100-000 band of HNO with a microdensitometer tracing of the corresponding band photographed in absorption by Bancroft, Hollas and Ramsay (2, hereafter denoted by B.H.R.) Two features are apparent from this figure: (i) the excited-state structure breaks off above J' = 11, and (ii) there is a very clear intensity perturbation in the la) 16 140 16 190 v/cm-' FIG.1.-The 100-000 K = 4-3 sub-band of the 2-2 band system of HNO. (a) Microdensitometer trace of a photograph of the absorption spectrum (after Baxroft et Q L ) . ~ (b) Laser-induced fluores- cence excitation spectrum. The weak RQ4 and RR4 branches in (a) are of the 020-000 K = 5-4 sub- band. excitation spectrum for J' = 7 and 8 that is not apparent in the absorption spectrum. Higher-resolution excitation spectra show that these lines have multiple components arising from rotational perturbations [fig. 2(c) and (41. Even so, the sums of the intensities of these close components are less than expected by interpolation between high and low J values.The extent of this loss of intensity in the excitation spectrum, and the range of J values affected, both increase with increasing pressure. A similar behaviour was noted in a number of other sub-bands. We will return to this feature of the excitation spectra in the discussion of the time-resolved measurements. The spectra recorded included, in order of increasing frequency, the vibronic tran- sitions 011-000, 020-000, 100-000, 101-000 and 030-000. The first three show ex- tended K-structure which could be assigned in part using the tabulations of B.H.R. However, the higher resolving power and sensitivity showed up numerous perturba-128 1 ELECTRONIC SPECTRUM OF HNO 1 I I f I 16 100 5 10 15 v1crn-l n _. I I I I I 16120 25 30 35 LO vlcm-' I I I____ I 1 I 16 150 55 60 16 170 75 83 v/cm-l v/cm-' FIG.2.-Assignments of branches in the HNO fluorescence excitation spectrum close to the breaking- off limit at 16 182 cm-'.R . N . DIXON, M . NOBLE, c. A . TAYLOR AND M. DELHOUME 129 tions of up to 2 cm-l, and extra lines, such that it was necessary to re-analyse much of the spectrum. This analysis was accomplished using ground-state combination dif- ferences, computed using the most recent molecular constants of Johns and McKel- lar15 in assignments of the "R, "Q and "P branches with K' 3 2. The P-type sub- bands and low K' transitions are heavily overlapped, and for these it was not possible to make new unambiguous assignments. A section of the assigned spectrum is pre- sented in fig. 2. The assignments are summarised in the Appendix as a tabulation of upper-state term values, most of which are derived from three branches.Extra lines have only been assigned where they are consistently present in at least two branches. There remains a large number of weak unassigned lines throughout much of the spectrum. These are particularly evident near the breaking off limit where there are fewer main lines. Here they constitute about two-thirds of the total number of lines, but are typically only 1-5x of the intensity of the strong lines. The K-type doubling of the ground-state levels is completely negligible for K" =3 at Doppler-limited resolution. For K" = 2 it reaches the Doppler width by J" = 6, and is very large for K" = 1. Thus for levels with K' == 4 it has not been possible to deduce the upper-state parity even where there is an observed K doubling.However, for K' = 2 and 3 the ground-state combination differences are sufficiently different for pairs of asymmetric top components that the upper state parity could be assigned tak- ing into account the type-c selection rules. Many excited-state levels show asym- metry splittings which are much greater than those predicted from the rotational con- stants and not necessarily of the correct sign: some of these are visible in fig. 2 in the "R2 and "Q2 branches of the 100-000 band. Takagi et al.'6917 have recently observed similar unexpected asymmetry splittings in other sub-bands of the A",'A'' state in micro- wave optical double-resonance spectra. The frequencies of single sharp lines agree with those tabulated by B.H.R.to within h0.05 cm-l, but very many lines were found to be multiple at the higher resolution of our spectra. Upper-state term values of levels not greatly perturbed were found to lie generally within ca. 0.25 cm-I of those calculated with their molecular para- meters. In fig. 3 we present the energy shifts between the term values and those given using the symmetric-top approximation with the B.H.R. constants: AT = T(v, J , K ) - [To + G(u) + ( A , - BL,)K2 + BJ(J + 1) - D,,,K4 - DJK,"K2J(J + 01. (1) Some of the perturbations involve the systematic displacement of a considerable number of consecutive energy levels, some are very localised affecting only one or two levels and yet others consist of multiple perturbations of one J value.The strongest and most easily characterised perturbation occurs at low J in the K r = 4 manifold of the 011 vibronic state [fig. 3(b)]. Two lines are observed for J' = 5,6,7 and 8, with a minimum separation of 2 cm-', but there is no satellite line for J' = 4. The intensity distribution among these pairs, together with the pattern of perturbed energies, clearly indicates that the coupling matrix element increases with increasing J. For a singlet-singlet Coriolis interaction the angular momentum de- pendence of this element is proportional to K for AK = 0, and to [J(J + 1) - K(K & 1)]+ for AK = -& 1. The intensity mixing is closest to that for AK = - 1, and the best fit to the data with this assumption leads to : 01 1, K = 4 ATc,K = $3.26 0.25 cm-I ARff.= -0.074 0.005 cm-': Jeff. = 1.188 cm-I (2) = (0.197 & O.O07)[J(J 4- 1) - 1213 cm-I.130 ELECTRONIC SPECTRUM OF HNO / ox I @I I 0 0 0 0 I ATlcm - m lx E1 o ] 4 -- k 6 - 1" x o 0 I -- m I 0 X @I I I 1, i / I 0 L n 0 1 0 0 3 I L " 0 m 0 0 I 0 0 m 0 0 N n M + 2 's 0 0 0 0 s? 0 0 c-4 n M + 'r: 's 0 0 c 0 IG. 3.-Deviations between the term values for vibronic states of HNO A", 'A" and the model calcu- (b) Parity is not known for K = lations of eqn (1). (a) For K = 2 or 3 : 0, e levels; x , flevels. 4. (--- ) Least-squares fits to the perturbations analysed in eqn (2)-(5).R . N. DIXON, M . NOBLE, C . A . TAYLOR AND M . DELHOUME 131 There is a very similar perturbation of the f levels of the K = 3 manifold. again best fitted as a AK = - 1 perturbation, leading to : 011, K = 3(flevels) This is AT,,K = +4.01 -+_ 0.36 cm-I ABeffe = -0.031 & 0.003 cm-': Befp. = 1.235 cm-I (3) The e levels of this manifold are also perturbed, but less strongly and less systematic- ally, and there is a second region of perturbation for J > 14 which precludes unambi- guous assignments since only the "Q branch is not overlapped.These two values of Jeff* for the perturbing levels are substantially lower than the value of 1.263 cm-' for the upper-state levels, and are significantly different from each other. The overall nature of these two regions of perturbation is inconsistent with interactions with a single upper vibronic level of the g, 'A' state. Alternatively, the perturbing levels could belong to the a", 3A" state.In a molecule of the point group C, the direct spin-orbit interaction between a singlet and a triplet state of the same configuration is not forbidden by symmetry, unlike the case of H,CO of point group Czv. In addition the indirect spin-orbit-orbital-rotation interaction active in H2C0 18*19 may also couple the 2 and a" states, so that the allowed perturbations are given by AK = 0, & l , &2 and AN = 0, & l . Several of these matrix elements have a J-dependence similar to that used in eqn (2) and (3). Thus it is not possible with our present assignments to distinguish between a singlet-singlet and a singlet-triplet interaction. Whereas the perturbations discussed above involve crossings in fig. 3 by lines of negative slope (lower B if AN = 0) two further systematic perturbations in K' = 4 of 020 and K' = 4 of 100 involve crossings by lines of positive slope.Since these are more localised the J-dependence of the interaction matrix element cannot be deduced, and has been assumed constant. In these cases we have again assumed a J(J + 1) energy dependence appropriate to a singlet-singlet interaction, giving : 020, K = 4 H1,2 = (0.068 & O.O04)[J(J + 1) - 6]* cm-'. AT,,K = -2.1 -+ 0.9 cm-' Hl,2 = 0.42 & 0.04 cm-l. ABeff. = +0.055 & 0.020 cm-': Befra = 1.32 cm-' (4) and : 100, K = 4 ATy,& = -1.3 & 0.2 cm-' ABeff. = +0.018 & 0.004 cm-': Jeff. = 1.295 cm-' (5) Hl,2 = 0.088 & 0.016 cm-'. The remaining perturbations are too localised or erratic to permit any deduction con- cerning the properties of the perturbing levels.We note that both Dalby ' and B.H.R.2 encountered the same difficulty with respect to other HNO perturbations. MAGNETIC BEHAVIOUR The majority of the HNO lines show no significant activity at 0.1 cm-' resolution and 10.5 kG, as was to be expected for a non-degenerate singlet-singlet transition. This is also the case for many of the perturbed lines, but a few of these show marked line-broadening or changes in intensity. This underlines that the perturbations are not all of one type. Upon further investigation using the single-mode laser it was found that very many lines showed a small increase in linewidth from the Doppler limit of 1.1 GHz (f.w.h.m.) to 1.5-2.0 GHz, equivalent to a magnetic moment in the field direction of up to ca.0 . 0 3 ~ ~ . The perturbed lines in the K' = 4 manifolds of132 ELECTRONIC SPECTRUM OF HNO the 01 1 and 020 states, analysed above in eqn (4) and (9, are of this type. Since these show no more magnetic activity than many apparently unperturbed lines this supports the postulate that these perturbations arise from singlet-singlet interactions. A few of the numerous weak lines underlying the main bands show significant shifts, or much greater broadening such that some have disappeared at 10.5 kG. These are consistent with levels of a triplet state. Three lines which show considerable magnetic activity are those with J’ = 7, 8 and 11 in the K’ = 4 manifold of the 100 state, all of which are perturbed. Fig. 4 and 5 0 GHz 10 0 GHz 10 FIG. 4.-Magnetic perturbation of the 100 K = 4 J = 7 level of HNO 2, ‘A!’, recorded in parallel polarisation. (a) Excitation in RQ3(7) at 16 158 cm-’.(b) Excitation in RR3(6) at 16 177 cm-’. show excitation spectra in various magnetic fields for the J’ = 7 and 8 groups of lines run in parallel polarisation in both “Q and RR branches. At zero field the J’ = 7 group consists of one sharp line and one broad line sepa- rated by 5.55 GHz. The splitting between these is one of the perturbations fitted by eqn (5). In “Q excitation, which accentuates the components with high M, the broad line splits into a triplet, the separation between the outer components being 3.25 GHz in 10.5 kG. In “R excitation, which accentuates M = 0, the broad line splits completely into a doublet of spacing 3.20 GHz in 10.5 kG, with a third weaker under- lying component.Thus both the high M and low M components are split by the magnetic field, with splittings which are approximately first order in the field strength. This can only be brought about if both AJ = 0 and AJ = & 1 Zeeman interactions con- tribute to the splitting. This behaviour is qualitatively just that to be expected for interaction of a singlet level with a Hund’s case (b) triplet level in which N and S un- coup!e at high field.” However, the observed splitting is only 0.06 times that pre- dicted for a case (b) triplet level. The weaker sharp line only broadens by 0.6 GHz in 10 kG, so the magnetic activity does not arise from an off-diagonal Zeeman addition to the zero-field perturbation matrix element. We therefore conclude that the main splitting into a doublet, which was analysed in eqn (5), arises from a singlet-singlet perturbation, but that the stronger component is also coincident with a triplet state level which does not conform to Hund’s case (6).At zero field there are clearly five com- ponents spread over 18 GHz (fig. 5), with a sixth more remote line in both the RR and “Q branches. Two of the weak components are strongly active magnetically, one The J’ = 8 group is even more complex.R. N. D I X O N , M. NOBLE, c. A . TAYLOR AND M . DELHOUME 133 broadening, and one shifting by 8 GHz in 10 kG. The remote level is also active but becomes overlapped by stronger lines when the field is applied. The remaining three components, including the most intense are little affected.The number of com- ponents and their magnetic activity requires that, even when K-doubling is taken into -4.99 kG 0 20 GHz FIG. 5.-Magnetic perturbation of the 100 K = 4 J = 8 level of HNO x, ‘A”, recorded in parallel polarisation. Excitation in KR3(7) at 16 178 cm-’. account, there must be two perturbations to this level, probably one singlet-singlet and one singlet-triplet. A further consequence of a magnetic field is to cause a preferential decrease in intensity of a number o f lines in the spectrum. We show below that at the pressures of most of the experiments (a few mTorr) the excited state is partially collisionally quenched, and that this quenching is linked to perturbations. The selection rule for134 ELECTRONIC SPECTRUM OF HNO magnetic field interactions can result in mixing in a field between levels with AJ = &l which cannot perturb one another in zero field.TIME-RESOLVED MEASUREMENTS A study of quenching rate constants and extrapolated fluorescence lifetimes for the A-8 system of HNO has recently been described by Yamada et aL20 The HNO was generated in a discharge flow system through the radical reaction HCO + NO --f HNO + CO in the total pressure range 0.05-0.5 Torr, and excited by a flash-lamp-pumped dye laser. This had a bandwidth which resulted in simultaneous excitation of many J levels within a sub-band. The lifetimes obtained were in the range 6-10 ,us, and the quenching rate constants in the range 4 x 10-Il-2 x lo-'* cm3 molecule-' s-'. Our initial time-resolved measurements, but with single rotational-level excitation, were made over the pressure range 0.5-200 mTorr.Although measurements over the range 10-200 mTorr resulted in similar extrapolated lifetimes and quenching constants to those reported,20 the Stern-Volmer plots showed considerable curvature. The decay curves at higher pressures were also markedly non-exponential. It is now clear from 100 - 4- .3 s G .- 3 10 W rn Q, +- .3 8 u, Q, a: 1 0 25 50 75 1 t ime/ps FIG. 6.--Serni-logarithmic decay curves for excitation in the RR3(4) line of the 100-000 band of HNO, (a) 0.5 mTorr, (6) 5.4 mTorr, (c) 10.3 mTorr total pressure. our new measurements that energy transfer plays an important role in the decay of HNO fluorescence at pressures of tens of mTorr or higher. We have therefore con- centrated on the pressure range 0.5-1 5 mTorr, and have studied the decay from ca.30 different pumped levels. These have been chosen to include apparently unperturbed levels, levels within systematically perturbed series and randomly perturbed levels. Decay curves were accumulated in the store of the PET microprocessor for 50 to 100 laser pulses, and then output on a recorder in both linear and logarithmic form.R . N . DIXON, M. NOBLE, c . A . TAYLOR AND M . DELHOUME 135 At the lowest pressures of 0.5 to 1 mTorr the logarithmic decay curves were found to be close to linear for at least one decade in all cases where excitation was to a single level. At ca. 10 mTorr the linear portion of many logarithmic decay curves was limited to about half a decade (see fig.6). The initial logarithmic decay constants were determined graphically from the linear portions of the curves if these were suffi- ciently long, or by fitting the decays to a double exponential when necessary and com- puting the initial slopes. The initial slopes were then used in constructing Stern- Volmer plots. The results are summarised in table 1. TABLE 1 .-FLUORESCENCE DECAY RATES AND QUENCHING RATE CONSTANTS FOR SINGLE ROVI- BRONIC EXCITATION OF HNO U' K branch v/cm-' J' kf/104 s-1 k,/10-10 crn3 S-1 01 1 020 1 00 100 15 750.17 751.66 752.60 754.75 755.97 756.98 16 116.06 115.18 113.51 113.75 114.36 110.93 109.15 109.33 16 113.20 112.55 110.77 103.59 097.58 16 172.01 173.87 175.62 177.06 177.25 " 178.28 178.56 178.86" 179.82 180.91" 181.79 181.92 4 5 s 6 m 6 m 7 s 8 m 4 5 6 w 6 m 6 w 8 9 m 9 m 3 4 6 11 14 4 5 6 7 m 7 s 8 w 8 s 8 w 9 s 10 s 11 w 11 m 4.9 i 1.9 4.4 f 0.6 3.8 f 0.4 4.2 i 0.7 6.1 i 1.6 2.7 i 1.8 4.7 i 0.3 5.3 i 0.4 4.1 f 0.5 3.9 i 0.2 3.6 i 0.2 4.2 f 0.2 5.3 f 0.3 4.1 i 0.7 4.4 f 0.6 4.5 i 0.2 4.4 f 0.2 4.1 f 0.5 3.8 f 0.5 4.56 f 0.13 4.48 f 0.33 4.17 f 0.55 4.48 0.42 4.14 f 0.95 4.31 i 0.48 4.54 f 0.42 3.50 f 0.40 4.87 0.48 4.66 f 0.72 w 295 183 rt 16 5.5 & 1.2b 2.6 f 0.7 2.94 f 0.22 2.86 f 0.13 2.92 f 0.36 2.86 i- 0.6 1.64 f 0.16 1.81 f 0.17 1.88 f 0.32 2.83 f 0.16 2.04 f 0.09 2.46 f 0.10 2.28 f 0.17 2.95 f 0.45 1.79 f 0.26 1.46 f 0.09 1.73 f 0.10 2.52 f 0.31 3.65 f 0.34 1.10 f 0.11 1.32 f 0.20 1.76 f 0.30 3.03 & 0.26 2.44 & 0.72 3.56 f 0.28 3.10 & 0.22 2.24 f 0.22 2.59 i 0.24 2.58 f 0.44 5 * 3 1.1 f 0.6b 2.67 f 0.09 1 - " Magnetically sensitive transition.s = Our extrapolated zero-pressure lifetimes are all longer than 20 ,us and are thus several times greater than those reported by Yamada et a1." The quenching rate con- stants range from I x lo-'' to 5 x lo-'' cm3 molecule-' s-', and correspond to half- quenching pressures of 10 to 2 mTorr. The dominant quenching partner in the measurements of ref. (20) was N,, whereas we have a mixture of HNO, N20 and Slow component of strongly double exponential. strong, m = medium, w = weak component of perturbed level.136 ELECTRONIC SPECTRUM OF HNO H20. Since our high-pressure measurements were similar to those of ref. (20), the higher quenching rates measured at low pressure indicate a difference of decay regime, rather than the difference in collision partner.The results in table 1 show that in a number of cases there is a marked discrimina- tion in measured quenching rates with excitation in close-lying levels. In each of these cases one or more of the levels are perturbed. Thus energy transfer cannot be so fast that the decaying excited-state population distribution lacks a memory of the initially populated level. In order to assess the importance of energy transfer we have sought to make a separate measurement of its rate. One excited level makes this possible. We have shown that the J' = 0 level of the 101 vibronic state lies 520 cm-I above the dissocia- tion limit to H + NO, but is stable and emits fluorescence. The excitation spectrum of the 101-000 band at low pressure consists of the one line pP,(l) that leads to J' = 0.All other rotational levels in this state are sufficiently strongly coupled to high levels of the ground state through orbital-rotational coupling that they are predissociated. This one line is too weak to study at the lowest possible pressure, but we have been able to measure a decay rate of (31 j: 3) x lo4 s-I at 5.2 mTorr by summing 350 pulses. If we assume that the fluorescence rate for this level has the typical value of ca. 5 x lo4 s-' then the sum of the quenching and energy-transfer rates from this level is ca. 1.5 x cm3 molecule-' s-'. This rate constant is about three times higher than the highest measured quenching rate constant for a level below the dissociation limit.Thus the fluorescence decay at pressures of tens of mTorr or higher must be dominated by the more persistent levels in a relaxed population. In a short paper on the predissociation of the 101-000 absorption band of HNO Freedman2' noted that the linewidths in this band were pressure dependent, showing an increased broadening at 80 Torr total pressure compared with 10 Torr. At 80 Torr a rate constant of 1.5 x low9 cm3 molecule-1 s-l would lead to a lifetime broadening increase of the linewidth (f.w.h.m.) of 0.02 cm-l, which is just at the limit of Freedman's measurements. Note, however, that collisionally induced line- widths depend on the rate of phase-changing collisions, which is usually higher than the rate of population decay. With energy transfer taken into account the evolution of the population n(i) of an excited level will be given by: where kf is the first-order fluorescence rate constant and k , and k, are second-order quenching and energy-transfer rate constants, respectively, N being the total molecular number density.At this microscopic level we define quenching as irreversible loss from the excited state over the timescale of the measurements. The effective first- order rate constant for the initial fluorescence decay with excitation in level i is then: The slope of the Stern-Volmer plot is therefore given not just by kq(i), but includes contributions from energy transfer. For transfer between levels of equal fluorescence rate this second contribution vanishes, but it can be important where perturbed levels are involved, for which the k,(i) will not all be equal.Thus excitation to a regular level, followed by transfer to a perturbed level with low kf(j), increases the effective quenching constant to above k,(i). Conversely, excitation to a perturbed level withR . N . DIXON, M . NOBLE, c . A . TAYLOR AND M . DELHOUME 137 low k,(i), followed by transfer to a regular level, tends to reduce the effective quenching constant to below the appropriate kq(i). The true microscopic quenching constants of the perturbed levels must therefore be even higher than those in table 1 derived from Stern-Volmer plots of the initial decay rates. We can now assess the measured decay parameters. 011, K = 4 The J-dependent perturbation noted above for this manifold of levels affects all the levels studied with the possible exception of J = 4.The quenching constants for this series are consistently higher than those of unperturbed levels in the 020 and 100 vi- bronic states, 020, K = 4 The two lowest levels ( J = 4,5) are the only clearly unperturbed levels in this series, and have the lowest quenching constants. The higher members are mainly split into doublets with somewhat erratic splittings. Since the lines in each pair have approxi- mately equal intensity we presume that this is an anomalous K-doubling, but this can- not be proved from combination differences. Such perturbations must arise from interactions with more remote levels with large asymmetry splittings. For J = 6 we have one unperturbed level and one almost symmetrically split doublet (see fig.2) but, surprisingly, the highest quenching is for the central level. In general there is a corre- lation between the extent of perturbation and increased quenching. 100, K = 3 The three low J lines studied from this manifold are all clearly unperturbed, and have fairly low quenching constants. The J = 11 level has an anomalously large K- doubling, as also does the level with J = 14. This series of levels breaks off above J = 16. For J = 14 energy transfer upwards in energy can therefore lead to collision- ally induced predissociation. This process may contribute to the high quenching rate. 100, K = 4 The levels in this manifold show a systematic increase in quenching rate up to the predissociation limit at J = 11. Superimposed on this trend there is a maximum quenching for J = 7 and 8, which marks the centre of the local perturbation analysed above.We may comment here that for both these J values the quenching of the mag- netically active levels (highest energy components in both cases) is less than that of the perturbed levels with little magnetic activity, although the differences are not large. The decay with excitation in J’ = 11 is strikingly different from all other cases (fig. 7). The logarithmic decay curve has two linear portions with one very fast decay constant and a slower “ normal ” decay. This fast decay persists at 0.7 mTorr, and arises from intramolecular predissociation. The smaller proportional increase in this decay with increasing pressure gives a similar quenching rate to that for J = 10.We have also found that the intensity in the C.W. excitation spectra of the weak lines lead- ing to J ’ = 11 increases relative to other J values with increasing pressure, as does the intensity of the slow component of the decay curve. This indicates the importance of collisional stabilisation of unstable levels to the intensity distribution in excitation spectra. One general feature of the results in table 1 is the small range of zero-pressure fluorescence rates k,. We had expected that the rates for the perturbed levels would have been reduced rather more strongly below the mean value of 4.35 x lo4 s-’138 ELECTRONIC SPECTRUM OF HNO too 10 1 0 5 10 15 20 25 timelps FIG. 7.-Semi-logarithmic decay curve for excitation in the RR3(10) line (high-frequency component of the 100-000 band of HNO.Pressure 2.6 mTorr. (z = 23 ps). at 0.5 mTorr. Energy transfer may still be causing some rotational scrambling even FLUORESCENCE YIELD The explanation for the intensity anomaly noted in fig. 1 is now quite clear. At the centre of the perturbation ( J = 7 and 8) the increased quenching compared with low and high J levels leads to a lower quantum yield for fluorescence. As the pressure is increased excitation in neighbouring lines is also subject to this loss through energy transfer into J = 7 and 8. This loss is particularly evident in C.W. laser excitation since the fluorescence is then integrated over the complete excited-state decay. We have been able to simulate this behaviour by numerically integrating the Master equa- tions [eqn (6)] for a simple model approximating to the full set of levels in the 100 K = 4 manifold.This model also reproduces the observed departure from single exponential decay at high pressures. A similar striking intensity anomaly is evident in the fluorescence excitation spec- trum of the ~,1A”(050)-~,1A’(000) band of HCCl reported by Kakimoto et aZ.22 Lines from four J values centred on K’ = 0, J’ = 9 and 10 are weak. In the presence of a magnetic field of 10 kG the perturbed lines decrease further in intensity and some in- crease in width by 1-2 GHz. Lifetime studies on this system would confirm whether this is another example of preferential collisional quenching. DISCUSSION The levels of the 2, lA” state of HNO considered in this paper lie ca.16000 cm” above the lowest level of the 2, ‘A’ state, ca. 9000 cm-’ above the presumed origin5 of the a“, 3A” state, and within 700 cm-l of the ground-state dissociation limit at 16 450 cm-l. We have computed the average density of vibrational levels at this energy from our analytical potentials using the Thomas-Fermi statistical approxima- tion, giving one level per 20 cm-’ for the 8 state and one level per 50 cm-’ for the a” state. An important feature of the potentials is that, whereas the 8 state is the groundR. N. DIXON, M. NOBLE, C. A . TAYLOR AND M. DELHOUME 139 state near the equilibrium geometries of the 8 and 2 states, the 8 state potential rises above that of the a” state for modest displacements of the H atom. Consequently the 8 and a” potential surfaces intersect over a wide range of energies.The weak magnetic character of many apparently unperturbed levels of the 2 state is far too large to be consistent with a pure singlet state, but no strong perturbed line has a magnetic moment consistent with a pure triplet state interaction. We therefore propose that most, if not all, of the perturbations arise from the following mechanism. The levels of the 2, ’A” state are mixed through orbital-rotational interaction with nearby levels of the 8, ‘A’ state, and these are in turn extensively mixed through spin- orbit coupling to nearby levels of the a“, 3A“ state. The perturbations then arise when there are AJ = 0 degeneracies between levels of the 2 and 2 states, but with some triplet state contamination of the wavefunctions.In addition there may be isolated direct 3-a” perturbations, but we have found that the only levels which are strongly magnetic are members of multiply perturbed groups. We have already concluded that predissociation of the 2 state proceeds through this coupling to high levels of the 2 state, with no selection rule on K, because of the very large vibrational amplitudes of these 8 state levels. Further evidence is that the greatest predissociation line- widths are for the 101 and the strongest perturbations occur in the 001 and 01 1 levels.2 In all three cases one quantum of the bending vibration v, is excited. Since the x a n d 8 states become degenerate in linear HNO, at energies not much higher than these states, excitation of v3 will enhance the Franck-Condon factors for the 2-8 per- turbations.Gelbart and Freed23 have put forward a model for fluorescence quenching which specifically attributes the rates of collision-induced electronic quenching to the mixed character of the wavefunctions of perturbed levels. In a first-order treatment this rate is given by (p’) times the rate of rotational relaxation, where (p’j is the sum of the squares of the coefficients of pure final-state wavefunctions in the mixed-state wave- function. The quantitative validity of this model has been questioned for intersystem crossing in CO and an alternative formulation For apparently unperturbed levels our quenching rates are 8-15% of the rate of rotational relaxation, rising to ca. 20-300/, for perturbed levels.The rates are in quali- tative agreement with the Gelbart and Freed model but do not correlate with (p2) where this can be estimated. Perhaps this difficulty arises because no level of the 2 state with J > 0 is truly unperturbed. Furthermore, the random nature of the pertur- bations is probably a reflection of the random pattern of levels of the A? state close to dissociation. The pattern of perturbing levels may be more regular at the energy of the 000 level of the x state, permitting a more quantitative interpretation of time- resolved measurements. In H2C0 the lifetimes of many of the levels of the 2,”’’ excited state are greatly shortened by collision-free internal conversion to high levels of the 8, ‘ A , state, which form a “ lumpy continuum ” at this energy.Weisshaar and Moore have found that these lifetimes can be modified by Stark tuning of the interacting levels into or out of resonance.* Such a process cannot occur for HNO since, even if J and parity are the only good quantum numbers for the high ground-state levels, the level spacing is still far coarser than the level widths. This behaviour of HNO can be contrasted with that of formaldehyde. We are indebted to the S.R.C. for funds to purchase the three laser systems, and Keith Rosser’s technical We also for financial support for two of us (M. N. and C. A. T.). assistance was invaluable in the construction of much of the equipment. thank Dr. D. A. Ramsay for the original of fig. l(a).140 ELECTRONIC SPECTRUM OF HNO APPENDIX Table 2 presents the mean upper-state term values. Where K-doublings have been re- solved, and the parities assigned, these are denoted by the e, f notation of Brown et For perturbed levels only the relative intensities of the various lines in the excitation spectra are indicated by s-strong, m-medium, w-weak and vw-very weak.TABLE 2.-EXCITED-STATE TERM VALUES J parity T/cm - J parity T/cm - (a) 011, K = 3 3 4 5 5 5 6 6 7 7 8 8 9 9 10 10 10 11 4 5 5 6 6 7 7 8 8 8 9 10 11 3 4 5 6 7 7 8 8 9 9 10 10 11 11 (b) 011, K = 4 (c) 020, K = 3 e f e f e f e f e f e f e f f f e f e f e e f f e 15 758.24 768.35 780.69 w 780.94 m 781.13 m 796.14 796.27 81 3.81 81 3.99 834.02 834.29 856.68 m 857.03 s 881.62 m 882.31 s 883.33 w 909.20 w 15 920.17 932.53 s 934.51 vw 947.04 m 949.19 m 963.93 mw 966.70 s 983.10 vw 986.75 m 986.99 m 16 009.39 034.65 062.20 16 150.05 160.13 172.73 187.85 205.43 205.50 225.62 225.67 248.32 248.39 273.50 273.51 301.24 301.25 11 11 12 12 12 13 13 13 14 14 14 15 15 15 16 17 18 12 12 13 14 14 15 15 16 16 17 17 17 12 12 13 13 14 14 15 15 16 16 17 17 18 18 e f f e f f e f e f e e e f e e e f e e f e f e f e f f e f e 910.11 s 910.76 w 939.05 m 940.42 s 940.88 w 971.23 w 973.24 s 973.67 m 16 008.55 w 009.12 m 009.28 m 045.52 w 046.78 w 16047.11 m 087.31 m 130.38 m 175.93 m 092.58 m 092.84 m 125.43 160.73 m 161.77 ms 198.44 ms 199.50 m 236.86 ms 239.89 ms 281.77 m 282.29 w 282.89 m 33 1.44 331.46 364.17 364.18 399.43 399.45 436.98 437.17 477.35 477.40 520.15 520.22 565.43 565.55R .N. DIXON, M. NOBLE, c. A . TAYLOR AND M. DELHOUME 141 TABLE 2.-continued J parity T/cm-' J parity T1cm-l 2(d) 020, K = 4 4 5 6 6 6 7 7 8 2 e 2 f 3 e 3 f 4 e 4 f 5 e 5 f 5 e 6 f 6 e 6 f 7 e 7 f 8 e 8 f 9 e 9 f 9 e 10 e 2(e) 100, K = 2 2 ( f ) 100, K = 3 3 4 5 6 e 6 f 7 e 7 f 8 e 8 f 9 e 9 f 10 e 10 f 11 e 2(g) 100, K = 4 4 5 6 7 7 8 8 8 16 296.92 309.62 324.25 w 324.49 m 325.10 w 342.11 m 342.62 w 362.41 16 093.78 093.81 101.52 101.53 1 1 1.77 1 1 1.95 124.67 w 124.68 s 124.90 m 139.81 m 140.02 s 140.03 m 157.93 157.96 178.37 178.42 201.42 m 201.51 s 201.74 m 226.87 m 16 197.89 208.12 220.92 236.22 236.26 254.09 254.13 274.56 274.64 297.43 297.65 323.18 323.18 351.25 16 342.01 352.74 370.07 387.81 mw 388.00 m 408.02 w 408.22 w 408.31 m 9 9 10 10 11 11 12 13 10 10 11 1 1 12 12 13 13 14 14 15 15 16 16 17 17 18 18 19 19 11 12 12 13 13 14 14 15 15 16 16 17 17 8 8 9 9 10 10 11 11 f f f f e f e f e e f e f f f e e e e e f f f f f e e f e f e e e 385.08 m 385.26 m 410.31 m 410.53 m 438.11 m 438.35 m 468.42 501.25 227.20 s 227.27 w 255.29 255.42 286.23 286.88 319.55 320.28 355.58 356.27 394.08 394.38 435.26 435.21 478.66 478.88 524.62 525.1 1 573.22 573.82 351.54 382.10 382.12 415.26 41 5.37 450.96 451.10 489.52 489.67 530.45 530.53 573.31 573.79 408.42 w 408.61 w 430.71 vw 431.31 s 455.98 vw 456.84 s 484.89 w 485.02 mw142 ELECTRONIC SPECTRUM OF HNO F.W. Dalby, Can. J. Phys., 1958, 36, 1336. J. L. Bancroft, J. M. Hollas and D. A. Ramsay, Can. J. Phys., 1962, 40, 322. J. K. Cashion and J. C. Polanyi, J. Chem. Phys., 1959, 30, 317. M. J. Y. Clement and D. A. Ramsay, Can. J. Phys., 1961, 39, 205. T. Ishiwata, I. Tanaka and H. Akimoto, J. Phys. Chem., 1978, 82, 1336. M. A. A. Clyne and B. A. Thrush, Discuss. Faraday Soc., 1962, 33, 139. J. C. Weisshaar and C. B. Moore, J. Chem. Phys., 1980, 72, 5415. K. Shibuya and E. K. C. Lee, J . Chem. Phys., 1978, 69, 5558. ' R. N. Dixon, K. B. Jones, M. Noble and S. Carter, Mol. Phys., 1981, 42, 455. lo G. W. Kirby and J. G. Sweeney, J. Chem. Soc., Chem. Commun., 1973, 704. l1 J. E. T. Corrie, G. W. Kirby, A. E. Laird, L. W. MacKinnon and J. K. Tyler, J. Chem. SOC., l2 S. Gerstenkorn and P. Lucy Atlas du Spectre d'absorption de la molecule d'iode (C.N.R.S., 1978). l3 R. N. Dixon and M. Noble, Chem. Phys., 1980,50, 331. l4 T. W. Hansch, Appl. Optics, 1975, 11, 895. l5 J. W. C. Johns and A. R. W. McKellar, J . Chem. Phys., 1977, 66, 1217. l6 K. Takagi, S. Saito, M. Kakimoto and E. Hirota, Proc. 14th Int. Symposium on Free Radicals Chem. Commun., 1978, 275. (Sanda, Japan, 1979), p. 254. K. Takagi, S. Saito, M. Kakimoto and E. Hirota, J . Chem. Phys., 1980,73, 2570. la C. G . Stevens and J. C. D. Brand, J. Chem. Phys., 1973,58, 3324. l9 J. C. D. Brand and C. G. Stevens, J . Chem. Phys., 1973, 58, 3331. 2o F. Yamada, T. Ishiwata, M. Kawasaki, K. Tsukiyama, K. Obi and I. Tanaka, Proc. Proc 14th Int. Symposium on Free Radicals (Sanda, Japan, 1979), p. 275. 21 P. A. Freedman, Chem. Phys. Lett., 1976, 44, 605. 22 M. Kakimoto, S. Saito and E. Hirota, Annual Review of the Institute for Molecular Science 23 W. M. Gelbart and K. F. Freed, Chem. Phys. Lett., 1973, 18, 470. 24 D. Grimbert, M. LavollCe, A. Nitzan and A. Tramer, Chem. Phys. Lett., 1978, 57, 45. 25 J. M. Brown, J. T. Hougen, K.-P. Huber, J. W. C . Johns, I. Kopp, H. Lefebvre-Brion, A. J. Merer, D. A. Ramsay, J. Rostas and R. N. Zare, J . Mol. Spectrosc., 1975, 55, 500. (Okazaki, Japan, 1979), p. 54.

ISSN:0301-7249    

DOI:10.1039/DC9817100125

出版商:RSC    

年代:1981

数据来源: RSC

摘要:

-Pulsed Dye-laser Studies in the B32, State of Se, Land6 Factors and Perturbations BY GERARD GOUEDARD AND JEAN CLAUDE LEHMANN Laboratoire de Spectroscopie Hertzienne de l'E.N.S. (AssociC au CNRS L.A. no 18), 24 Rue Lhomond, 75231 Paris Cedex 05, France Received 29th December, 1980 Measurements by the Zeeman quantum beats method of Lande factors in the B3C; state of Se2 have been performed for a large number of rotational levels. A theoretical analysis of the g varia- tions near local perturbations shows the great sensitivity of this parameter. The results throw some light upon the interaction responsible for these perturbations and the properties of the " P1, " state which interacts with B3Z:,. 1. INTRODUCTION The study of the magnetic properties of molecular excited states has long been a useful tool in molecular spectroscopy.As an example, one of the most established techniques is the study of magnetic rotation spectra, or the Faraday effect. However, in general, due to the weakness of the magnetic interaction, the direct study of the Zeeman splitting of excited states has been restricted to highly paramagnetic species. To overcome this difficulty, one has to use doppler-free spectroscopic techniques.' The development of pulsed tunable lasers has permitted the use of the quantum-beats technique. These methods have been used on molecules such as I2 and OH 2, leading to a precise determination of excited-state parameters. We will present here a study of the 'OSe, molecule: systematic measurement of Land6 factors in the B3C;(0,+) state has permitted us to demonstrate that these parameters are very sensitive to small perturbations of the molecular state.The Se, molecule has been chosen since it seemed to be a good example of a paramagnetic ( S = 1) molecule without nuclear spins. Moreover the B3Z;, +-+ X3C; band system is fairly ~ e l l - k n o w n ~ - ~ and lies in a wavelength rmge easily accessible to dye lasers. 2. EXPERIMENTAL The optical excitation source is a N2 laser-pumped dye laser built following the Hansch design, with particular attention to the reduction of the linewidth. Our system gives 5 ns pulses of ca. 2 GHz spectral width. This allows in most cases excitation of a single vibration- rotation level of the *%e2 molecule. The laser wavelength is measured with a 1.5 m spectrometer; wavenumbers are then de- termined to within ca.0.3 cm-'. It is then possible to identify previously measured or cal- culated lines of the absorption spectrum of Se2. However, for some lines, especially those belonging to the B l , t X l , system, or near local perturbations, the identification can be difficult. By using several organic dye solutions, we have covered almost the whole spectrum be- tween 3600 and 4200 A. We have therefore been able to study levels belonging to u = 0-6 in the BO,f state. Mainly for intensity reasons the B1, t+ X1, system is more difficult to study; we have been limited to B l , , u < 3.1 44 B3C; STATE OF Se, The selenium is contained in a carefully outgassed fused silica cell. The temperatures of the cell and side arm are separately regulated to maintain the vapour pressure of Sez below 1 Torr.The fluorescence light is viewed at right angles to the laser beam and dis- persed by a small monochromator. For detection two separate systems are used alternatively. (i) First a slow gated photomultiplier merely gives the time-integrated intensity of the fluores- cence. (ii) Otherwise, the fluorescence light is detected by a fast photomultiplier system (2 ns risetime) The decay signal is fed into a Tektronix R 7912 transient digitizer, which catches the whole time evolution of the fluorescent pulse. Adding about a thousand pulses in an averager then is enough to give a good signal-to-noise ratio. The whole detection system is carefully time- scaled against a fixed-frequency generator.When the laser and fluorescence lights are polarized perpendicular to an externally applied magnetic field, Zeeman quantum beats are easily obtained. The modulation depth of the beats superimposed on the exponential decay of fluorescence is generally between 10 and 20%. Substracting the signal in zero field one obtains pure oscillations; the beat frequency is measured and the Lande factor value deduced through the well-known expression The monochromator then scans the main part of the fluorescence spectrum. AV = 2 g p ~ H / h , where Av is the beat frequency, g the excited-state Lande factor, pB the Bohr magneton and H the magnetic field strength. In our experiments, the magnetic field has been set according to two contradictory re- quirements. It must be high enough to have a maximum number of Zeeman periods during the excited-state lifetime z: 2gp,H > h/z.However, care must be taken to limit the mag- netic broadening of the absorption line: the Zeeman width 2J(gB - gX)flBH should not ex- ceed the order of the Doppler width hAvD. The magnetic field was therefore fixed according to l / ~ < 2 g p ~ H / h < AvD/J. A beat frequency of ca. 80 MHz was found to be a good compromise for J values below 100. In these conditions, the gJ measurements have a precision of ca. 2%. This has proved to be at least as sensitive to perturbations as the measurements of spectral line-shifts in the absorp- tion spectrum and in some cases more sensitive. 3. LANDE-FACTOR THEORY IN 3C STATES Very detailed calculations have been performed on 3E Land6 factors, for molecular ground states.' In particular, they show that the influence of far-away perturbing states is important.For the excited B3C; state of *OSe, which is subject to ZocaZ per- turbations, one may therefore expect large variations of g around these perturbations. Even for what appears to be (from spectroscopic measurements) non-perturbed rota- tional levels, the Land6 factor may have an " abnormal " value. For a pure Hund's case (a) or (6) 3C state the Land6 factor is very readily calculated. In Hund's case (a), the electronic spin of the molecule is coupled to the internuclear axis and therefore gJ" = 2CQ/J(J + I ) . This would give for the case (a) BO; state gJ = 0 (account is taken here only of the electronic spins because A = 0).This value in fact represents the maximum Land6 factor value for a S = 1 molecular state (the Zeeman energy of the M = f J sublevels is then AE, = gJpB HM = &2pBH). In fact measurements of the gJ values in the BO;, v = 2 state ( f i g . I), which is not perturbed below J = 90, show that they are almost independent of J and equal to gJ = 9.5 x This is in contradiction to the results expected for pure case (a) or (b) state, whereas the energy splitting between the B1, and BO; com- ponents 2R 2: 77 cm-' [ref. (6)] would indicate that B3C; in Se, is close to case (a)! The explanation of this discrepancy may be found by' considering that molecular states are subject to intramolecular as well as external (Zeeman) perturbations. If On the other hand, in a case (b) 3C.state g: 21 2/J. (in ,us units).G . GOUEDARD AND J . C. LEHMANN 145 0.5 V, and V, are the corresponding terms, the energy shift of a given 1B) level is given in second-order perturbation theory by AE = ‘fl(il( V, + Ve)IB)12/cc(EB - EJ. If both V, and V, have a non-zero matrix element between IB) and li), there is a crossed term between the Zee- man and intramolecular perturbations; this term is proportional to Ye, therefore to the magnetic field strength H and contributes to the Land6 factor value of state B.’ i The sum runs over all possible perturbing li) states, - 11 21 31 41 51 61 71 81 91 J FIG. 1.-g(J) values in BO;, ZI = 2. The full line is displayed only for guidance. As an example consider the interaction between the B1, and BO,+ substates of Se,; it is mainly due to the non-diagonal part of the rotational hamiltonian H N D = -2B J * S; on the other hand, the Zeeman hamiltonian is Hz = ps(L + 2 s ) H.It has been shown’ that the crossed H N D x Hz term along the path BO,+ + BI, -+ BO,+ gives a contribution to g(B0;) = 8B/AE(I, - 0;) 2 7.5 x independent of J which is only 20% below the value found for BO:, u = 2. The same treatment may be applied to local perturbations. However, in this case, one must diagonalize exactly the molecular hamiltonian, whereas the Zeeman inter- action need only to be considered at first order to get the Land6 factor^.^ The results may be written in closed form for the interaction of only two levels 11) and 12) of non- perturbed energies E,” and E;, the non-diagonal molecular matrix element being W,,.Then the g factors of the two mixed states are: In this expression 0 is the mixing angle which characterizes the molecular perturbation 0 = tan-1[2W,,/(Ey - E 3 ] ; g , and g, are the non-perturbed Land6 factors for states 1 and 2 and g,, is the crossed Zeeman term between them. At this point it is important to make a distinction between homogeneous ( A 0 = 0) and heterogeneous (AR = & 1) perturbations. For an homogeneous perturbation between two states described by Hund’s case (a) Thus g,, varies as IJ(J + 1 ) l - I ; for only moderate J > 20 values, g,, will be very small. Moreover for R = 0 states, g,, = 0.146 B3C; STATE OF Sez For an heterogeneous perturbation An = A1 In this case, g,, decreases like l / J only for high J values, and will be one of the main components of the g, variations. In conclusion Lande factors and their variations near a molecular perturbation are a sensitive test of the kind (Ai2 = 0 or & l ) of the perturbation. A representative example of the influence of the g,, term is shown in fig. 2.It causes a large asymmetry in the g(J) curve, and extends the range of J values for which the perturbation is noticeable. J FIG. 2.-Theoretical g(J) curves: (- a -) unperturbedg values; (-------) full curve corresponding to the perturbation in BO:, u = 4, J _N 54.5; (- - - - -1 8 1 2 arbitrarily set to 0. The vertical arrow shows the contribution of the g12 term. 4. EXPERIMENTAL RESULTS The BO,+ state of 8oSe, is known to be perturbed at a number of (v, J ) value^.^^^*'* However, all these perturbations are rather weak and few extra lines have been ob- served.The influence of the molecular perturbation in BO:, v = 4, J 21 54.5 on the g, values is presented in fig. 3. The g, value suffers an abrupt discontinuity for the J value of maximum perturbation; a very weak perturbation also appears at around J = 41. Following the theory expounded previously, we have studied the g, curve. First it is clear from the “Fano Shaped” variation of g that the non-diagonal Zeeman term between the BO,+ and P (for perturber) states is important; the perturbation is therefore heterogeneous and P is a 1, state. Thus the gpB term may be written g,, = h / d J ( J + 1 ) ; the parameter h = (PI(L, + 2S,)IB) is independent of J but depends on the vibrational overlap integral between (BO;, u’ = 4) and (P, v”): h = h(v’, 0”).The perturbation being heterogeneous, the intramolecular matrix element W,, is proportional to [J(J + l)]*, Information about the perturbing states is therefore limited. _____ W P B = (PIB(L& + S+)IB) x d a r n ) . On the other hand, we assume that Eg and Eg vary linearly with J(J + l), B, and BpG . GOUEDARD AND J . C . LEHMANN 147 FIG. 3.-g(J) values in BO;, u = 4. The full line corresponds to a fit of the experimental points J = 41, 43 points are not included in the fit). (the being the corresponding rotational constants, and cross for J = Jo. mixing angle 0 may be written: Therefore the The parameter A = 2(PIB(L, + S,)IB)/(BB - B,) is thus defined to be independent of J (but it depends on u’ and u” through the matrix element).As we have seen that gB is independent of J far from a local perturbation, we have fitted the g(J) variation in BO:, u’ = 4 with the following parameters: gB, g,, h, A and Jo. The fit, as shown in fig. 3, is excellent. TABLE 1 .-RESULTS OF g-FACTOR PERTURBATION CALCULATIONS 1 68.8 9.69(2) 0.5(1) 0.155(7) 3.02( 10) 3 17.8 9.87(2) 2.3(1) 0.337(1) 5.12(4) 5.23 a 4 54.7 9.89(2) -0.26(1) 0.374(3) 5.32( 5) 5.36 2 97.7 9.57(3) -2.5(24) 0.229(33) 2.47(44) 5 64.8 10.26(8) - 1.9(4) 0.408(23) 5.53(29) 5.43 Calculated from 78Sez. The results of such fits for several of the observed g , variations in BO,+ l1 are shown in table 1 . The following facts appear: (i) The gB values vary slowly from u = 1 to 5 in BO;, following the variation of the (BI,-BO;) ( u ) interval according to the relation (ii) The g , values are poorly determined; in general the fits would allow the g , = 0 value.(iii) Unfortunately Wpe, the perturbation matrix element, and (BB - B,), the difference in the rotational constants, cannot be fitted separately. Only the para- meter A , proportional to their ratio, is determined. However, we have calculated AB3C; STATE OF Se, from spectroscopic information.12 These values are also displayed (Acalc) in table 1 ; the agreement between Greenwood's results and ours is good. (iv) As spectroscopic measurements12 could equally well be fitted by an heterogeneous or homogeneous interaction hamiltonian, the main result of our investigations appears in the demon- stration that Land&-factor variations are due to heterogeneous perturbations.This is verified for the five most intense perturbations in BO; shown in table 1 . This seems to indicate that they are due to a single perturbing state in various vibrational levels. 5. DISCUSSION The recent observation of a previously unknown excited state of Se, in an argon rnatrix13 at low temperature with cue 21 190 cm-I has prompted the reanalysis of the perturbations in 'OSe,. We have made a global fit of the positions of the previously mentioned perturba- tions and of some fluorescent 1, rotational levels;12 the derived vibrational and rota- tional constants for state " P " are given in table 2. Taking into account the siiiall TABLE 2.-" Pl,, " STATE SPECTROSCOPIC CONSTANTS (ALL UNITS ARE Cm-I)" T, = 24929.6(9) W, = 191.21(19) Be = 0.065 66(6) wexe = 2.23(1) = 6.3(1) x 10-4 standard deviation of the fit = 0.34 cm-' a Values in parentheses are one standard deviation in units of the last digit quoted.number of experimental points (9), these constants are quite preliminary. The abso- lute u numbering is only tentative but gives the best agreement with isotopic-shift results.12 When they are known, rotational constants derived from deperturbation l2 agree well with the values obtained here. Moreover, any crossing point between this new " P " state and BO; and Bl,, corresponds to a previously observed perturbation. The existence of such a state with cue w 190 cm-', R = 1, is thus firmly established; this state is responsible for the majority of local perturbations in the B3C; (v = 0-6) state.Some indications may be derived from the fluorescence spectra originating in perturbed B3C; levels. From our observations, it appears that perturbations in BO; induce transitions towards the X1, sub-level (which are very weak for non-perturbed BO; levels). Perturbations by state P in Bl,, do not induce transitions towards XO;. Also " P1, " levels observed in fluorescence12 with J = 21, 69, 120, 149 radiate only It might be the 1, com- ponent of a B"31J, state, analogous to the ones advocated in the interpretation of per- turbations in B3C; of S214 and Te,.15 However, this interpretation seems to violate the selection rule AX = 0 for optical transitions between case (a) states, which would only allow transitions between 311 (R = 0 or 2) and 3C (R = 1).It seems to us that this question remains entirely open and that the detailed mecha- nism of the B3C; state perturbations should receive further attention. The remaining problems are due to the possible nature of such a state. to XI,. This P1, state is thus optically connected mainly to Xl,.G . GOUEDARD A N D J . C . LEHMANN 149 M. Broyer, G . Gouedard, J. C . Lehmann and J. Vigue, Ado. Atom. Mol. Phys., 1976, 12, 165. R. Wallenstein, J. A. Paisner and A. L. Schawlow, Phys. Rev. Left., 1974,32,1333; P. Lebow, F. Raab and H. Metcalf, Phys. Rev. Lett., 1979, 42, 85. R. F. Barrow, G. G. Chandler and C. B. Meyer, Philos. Trans. R. SOC. London, Sect. A , 1966, 260, 395. R. F. Barrow, 1. R. Beattie, W. G. Burton and T. Gilson, Trans. Faraday SOC., 1971, 67, 583. H. Christensen and L. Veseth, J . Mol. Spectrosc., 1978, 72, 438. G. Gouedard and J. C. Lehmann, J. Phys. B, 1976, 9, 2113; D. J. Greenwood and R. F. Barrow, J . Phys. B, 1976, 9, 2123. ’ M. Broyer, J. C . Lehmann and J. Vigue, J. Phys. (Paris), 1975,36, 235. * G. Gouedard and J. C. Lehmann, J. Phys. (Paris) Lett., 1977, 38, L 85. lo K. K. Yee and R. F. Barrow, J . Chem. SOC., Faraday Trans. 2, 1972, 68, 1181. l 2 D. J. Greenwood, Ph.D. Thesis (Oxford University, 1980). J3 V. E. Bondybey and J. H. English, J . Chem. Phys., 1980, 72, 6479. l4 R. F. Barrow and R. P. du Parq, in Elementary Sulfur, ed. B. Meyer (Interscience, New York, 1965), p. 251 ; J. M. Ricks and R. F. Barrow, Can. J . Phys., 1969, 47, 2423. ’’ R. F. Barrow and R. P. du Parq, Proc. R. SOC. London, Ser. A , 1972, 327, 279. W. Lichten, Phys. Rev. A, 1971, 3, 594. G. Gouedard and J. C. Lehmann, J. Phys. (Paris) Lett., 1979, 40, L 119.

ISSN:0301-7249    

DOI:10.1039/DC9817100143

出版商:RSC    

年代:1981

数据来源: RSC

摘要:

Studies of the Optical Spectra of CaCl and SrF at Sub-Doppler Resolution BY JOHN M. BROWN, DAVID J. MILTON* AND TIMOTHY C. STEIMLE The Department of Chemistry, The University of Southampton, Southampton SO9 5NH Received 18th December, 1 980 Lines in the (0,O) bands of theA211-X2Z+ and B2C+-XZC+ systemsof CaCl and the B2C+-X%+ system of SrF have been studied at sub-Doppler resolution by intermodulation spectroscopy. The linewidth was typically 40 MHz (f.w.h.m.). The "F hyperfine splitting was resolved in the SrF spectrum and was measured for several rotational lines. The measurements were used in conjunction with the hyperfine parameters for the molecule in the ground state (measured by e.s.r.) todetermine the following parameters for the BZC+ state: . b = 12.9 & 1.6 MHz c = -51 f 28 MHz.These results suggest that the molecule is ionic, Sr+F-, in the B state. It has not proved possible to resolve 35Cl or 37Cl hyperfine structure in any of the lines studied in either the A-Xor the B-X transi- tions of CaCI. There are pronounced crossover signals in the A-X system which have a negative sign ( i e . , opposite to the normal Lamb dip signal). The dependence of these signals on various ex- perimental parameters has been investigated. A density matrix description of the three-level system is presented and solved to 4th order by an iterative method. To this order, the model predicts that the Lamb dip and crossover signals should have the same sign. 1. INTRODUCTION The narrow line, tunable dye laser has provided spectroscopists with the means of recording electronic spectra with unprecedented resolution.Many of the features that were previously masked by Doppler broadening in a conventional grating spec- trum can now be studied in detail, yielding additional information about molecular structure. In particular, nuclear hyperfine effects can be resolved, even in quite light molecules. These interactions tell us how the electrons are distributed within the molecule, in a region near the nucleus. Magnetic hyperfine effects provide specific information about the open-shell electrons whereas electric quadrupole effects depend on the variation of charge density at the nucleus. A variety of methods has been used to record spectra with sub-Doppler line- widths. One of the first was the technique of saturation absorption spectroscopy, in which molecules with zero Doppler shift are selected from a bulk sample by subjecting them simultaneously to two laser beams, travelling in opposite directions.Important observations on the visible spectra of H' and Na2 atoms were made in this way as well as the first of many studies of hyperfine structure in the electronic spectrum of i ~ d i n e . ~ The method worked well for strong transitions of stable species that could be generated in high concentration. However, many problems of interest did not have these com- pliant characteristics and it was soon realized that greater sensitivity could be achieved by detecting the fluorescence excited by the laser rather than the absorption of the laser beam itself. Warwick, Coventry CV4 7AL.* Present address: The Department of Chemistry and Molecular Sciences, The University of152 SPECTRA OF CaCl AND SrF There have been three main areas of development in sub-Doppler spectroscopy based on the more sensitive method. First it has proved possible to detect the fluor- escence excited in the transverse irradiation of a molecular beam.4 If the beam is formed in a supersonic nozzle there is the additional advantage that the molecules are rotationally " cold " which simplifies the electronic spectrum ~onsiderably.~ Nuclear hyperfine structure on optical transitions has even been resolved by longtitudinal irra- diation of a molecular ion beam with a laserY6 although strictly speaking such experi- ments are not sub-Doppler. The second experiment which exploits the increased sen- sitivity of fluorescence detection is known as intermodulated fluorescence. The sample is subjected to two counterpropagating laser beams, chopped at different frequencies and a signal is sought at the sum of the two chopping frequencies.The non-linear response of the molecules to the radiation serves to mix the two frequencies and is restricted to those molecules with essentially zero Doppler shift. Following the initial studies of the spectrum of iodine with this technique,'^^ the method has been used to resolve fine and hyperfine structure in the electronic spectra of several interest- ing free radicals (NH2,9 B02,10911 CaCl," CaF,13 V0,14 PH2,15 CaBr and CaI 16). The third general method of obtaining sub-Doppler information with fluorescence detec- tion is two-photon absorption.The Doppler shift for the first photon can be com- pensated by that for the second photon to produce very high resolution spectra. Perhaps because the readily available wavelength region covered by dye lasers does not match the appropriate disposition of electronic states for molecules of interest, the high-resolution aspect of two-photon absorption has not yet been greatly exp10ited.l~ In this paper, we describe studies of the optical spectra of CaCl anci SrF at sub- Doppler resolution by the technique of intermodulated fluorescence. The aim of the experiments was to resolve magnetic hyperfine structure for the halogen nucleus (the dominant alkaline earth isotopes, 40Ca and "Sr, both have zero nuclear spin).We were successfu1 in our object in the case of SrF and have used our results to determine hyperfine parameters for the molecule in the upper electronic state of the B2Z+-X2Z+ transition, However, we could not find evidence of 35Cl or 37Cl hyperfine structure in any lines of either the B2C+-X2Z+ or the Azll-X2E+ systems of CaC1. Neverthe- less, the study of the A-X system has allowed us to investigate the anomalous cross- over signals in this spectrum'' in greater detail. In an attempt to understand the observations, we have set up the density matrix equations for the levels involved and solved them by an iteration treatment through to 4th order. The results are presented in section 4. 2. EXPERIMENTAL The experimental arrangement used to record the intermodulated fluorescence spectra was essentially the same as that described in the earlier study of CaC1.I2 The free radicals were generated in a high-temperature oven.CaCl was produced in two ways, first by heating a mixture of calcium metal and CaCl, to ca. 900 "C and second by reacting calcium, entrained in argon, with chlorine gas in a Broida-type flow system." SrF was also formed using the Broida furnace, from the reaction between strontium metal and SFs. The operating pres- sure was less than 0.030 Torr for the Ca/CaC12 system and typically 0.30 Torr using the Broida furnace. The radiation source was a Coherent Radiation 599-21 dye laser, pumped with 3 W of all lines from a Spectra Physics 170 argon-ion laser. The studies reported here were made in the wavelength region 580-620 nm, the whole of which could be covered with a mixture of Rhodamine 6G and B in the dye laser.Single-mode output powers were in the range 30-100 mW. The laser output was divided into two beams of about equal intensity that were chop- ped at different frequencies (356 and 499 Hz). These two beams were focused and passed collinearly through the sample from opposite directions. Feedback to the dye laser wasJ . M. BROWN, D . J . MILTON AND T . C . STEIMLE 153 eliminated by rotating the plane of polarization of one beam with a A/2 plate and blocking its return path with an appropriately placed polarizer." Modulation of the fluorescence at the sum frequency was detected with a photomultiplier and tuned lock-in amplifier system. The photomultiplier was shielded by the use of collimating slits but we did not use band-pass interference filters to discriminate against chemiluminescence and other background radia- tion.Absolute frequency calibration was provided by a simultaneous recording of the excita- tion spectrum of iodine.19 Frequency markers for relative measurements were generated at 76.9 MHz intervals by monitoring the transmission of a 1 m, thermally stabilized confocal etalon with a finesse of ca. 65. The signals from the lock-in amplifier, iodine excitation spec- trum and frequency markers could be fed to a multi-channel y-r chart recorder and also, after appropriate digitisation, stored on a floppy disc for subsequent processing. Rotational assignments were made with the help of previous analyses of the spectra of CaCl 20~21 and SrF.22 3.STUDIES O F NUCLEAR HYPERFINE STRUCTURE (i) CALCIUM MONOCHLORIDE For CaCl, we have studied portions of the (0, 0) bands of both the A211-X2E+ system at 620 nm and the B2E+-X2E+ system a t 594 nm in considerable detail with the intermodulated fluorescence technique in an attempt to resolve 35Cl or 37CI hyperfine structure. Although there were a few lines that appeared to show the anticipated quartet structure (particularly for low-N transitions in the B-X system), it was not possible to achieve a consistent analysis and it became obvious that the extra lines are I FIG. 1.-A recording of a " single " line, P,(23), in the (0, 0) band of the BzX+-XzX+ transition of Ca3W. The spacing between the frequency markers from the confocal etalon is 76.9 MHz.The linewidth is measured to be 48 MHz (f.w.h.m.). The laser power in each beam applied to the sample in the intermodulation experiment was ca. 15 mW.154 SPECTRA OF CaCl AND SrF either satellites or arise from hot bands. The majority of lines are clean, single fea- tures (fig. 1). We are forced to conclude that the hyperfine effects on the lines in both band systems are smaller than the best linewidth that we could achieve (45 MHz f.w.h.m., see fig. 1). The corresponding 19F hyperfine structure in the A211-X2Z+ system of CaF has been resolved and measured in an intermodulated fluorescence study by Bernath, Cummins and Field.I3 The explanation of our lack of observations for CaCl appears to be the small magnetic moment of the C1 nucleus.The nuclear spin g-factor for 35Cl is ten times smaller than that for 19F and the isotropic hyperfine parameter is conse- quently much smaller (ca. 28 MHz for CaCl in the X2Z+ compared with 108.5 MHz for CaF).24 The previous work on CaF13 and the present work on SrF suggest that the hyperfine interactions for the A211 and B2Z+ states are likely to be much smaller. In the B-X system, the limiting hyperfine splitting for high N transitions is therefore 4b where b is the Frosch and Foley parameter2’ for the molecule in the X state. Thus the separation between adjacent hyperfine components is expected to be ca. 14 MHz, with a linewidth four times greater. This is consistent with our observa- tions. (ii) s TR o N TI u M M o N OF L u o R I D E Each line of the (0,O) band of the B2E+-X2Z+ system of SrF shows a small doub- ling (ca.50 MHz) when recorded by the intermodulated fluorescence technique. The splitting is a manifestation of the 19F hyperfine interaction. We have measured the splitting for several selected lines in order to determine the hyperfine parameters for the B state. In no case was the splitting fully resolved (a typical example is shown in fig. 2) so it was important to keep the laser power as low as possible, sacrificing signal- I FIG. 2.-The P2(22) line in the (0, 0) band of the BZC+-X2Z+ system of SrF, recorded by the inter- modulated fluorescence technique. The spacing between the frequency markers from the confocal etalon is 76.9 MHz. The splitting arises from I9F hyperfine interactions.J .M . BROWN, D . J . MILTON AND T . C . STEIMLE 155 to-noise for reduction in linewidth. The splittings were measured by scanning slowly over the line in question, at the same time monitoring the transmission of the 75 MHz FSR etalon. Several independent measurements of the splittings for 21 transitions were obtained in this way, including 2 from the (1, 1) band. The results are given in table 1 with the estimated uncertainty of each datum. In selecting the transitions, TABLE l.-I9F HYPERFINE SPLITTINGS IN MHz FOR B2C+-X2Z+ SrF assignment observed estimated obs - calca error 49.8 53.6 53.9 45.2 49.0 50.6 54.0 54.0 52.6 50.5 54.9 54.1 54.2 44.3 54.2 50.6 51.7 47.7 45.0 57.8 43.8 2.0 3.8 3.5 3.6 4.8 7.0 5.4 2.2 1.7 3.5 4.7 10.2 7.9 2.9 7.2 4.2 7.9 2.9 7.2 4.2 4.0 -0.2 3.3 3.5 - 5.3 - 1.5 0.0 2.7 1.5 2.1 -0.5 4.0 3.2 3.8 - 6.2 3.6 0.0 1.1 - 3.2 - 5.8 7.0 - 7.0 a Splitting calculated with b = 95.0, c = 31.0 MHz for the X2C+ state and b = 12.9, c = -51.1 Rotational transition in the (1, 1) MHz for the B2C+ state, and a linewidth of 30 MHz (f.w.h.m.).band. All other splittings are for lines in the (0, 0) band. we have tried to cover both high and low values of the rotational quantum number N and transitions involving both F, and F2 spin components. This increases the prob- ability of being able to determine both hyperfine parameters for the B2C+ state. The analysis of the splittings is based on the hyperfine Hamiltonian of Frosch and F01ey.~~ Details of the requisite matrix elements for a molecule in a 2C state have been given in many places in the literature [see, for example, ref.(16)]. The splittings depend on two hyperfine parameters, b and c and the spin-rotation parameter, y. The two hyperfine parameters for SrF in the ground state have been measured pre- viously in an e.s.r. study of a matrix-isolated sample.26 The corresponding quantities for CaF have been measured both by e.s.r.26 and in a molecular-beam e~periment.~~ A comparison of the two sets of values suggests that those determined by e.s.r. are very reliable. We have therefore fixed the ground-state parameters for SrF at the e.s.r. values in our fit (b = 95.0 MHz, c = 31.0 MHz) and used the data to determine the upper-state parameters. The energy levels were obtained by matrix diagonaliza- tion, neglecting the effects of elements off-diagonal in N .The values for the spin- rotation parameter in the B and X states (-4.058 GHz and 74.6 MHz, respectively), were taken from the analysis of Steimle et al.27 We have also allowed for the effects156 SPECTRA OF CaCl AND SrF of partial overlap of the two hyperfine components by simulating the calculated spec- trum for comparison with the observed splitting. The lineshape was assumed to be Lorentzian with a f.w.h.m. of 30 MHz. The data were fitted with a non-linear, weighted least-squares program, the weights being taken as the inverse squares of the estimated experimental uncertainties. The quality of the fit can be judged from table 1 ; two measurements for lines in the ( I , 1) band were also included since the mole- cular-beam work on CaF in the X2C+ state suggests that the vibrational dependence of the hyperfine parameters is very weak.24 The values for the hyperfine parameters for SrF in the B2C+ state determined by the fit are: b = 12.9 5 1.6 MHz, c 2 -51 & 28 MHz where the quoted errors are one standard deviation of the least-squares fit.Consideration of the matrix representation of the hyperfine and spin-rotation Hamiltonians for a 2C state'' shows that at high N values the 19F splitting becomes constant, equal to 3b and independent of the dipolar parameter c. It is therefore important to include some low-N transitions in the data set, even though they are intrinsically weaker, if the parameter c is to be determined. If b and c have the same sign, the splitting becomes larger than +b at low N in the Fl spin component and smaller in the F2 spin component.Examination of table 1 shows that such trends at low N are not really evident and consequently the dipolar parameter c is very poorly deter- mined. Indeed the quality of fit is almost as good if c is constrained to zero, in which case b is determined to be 12.4 t 1.7 MHz. It follows also from the remarks above that it is preferable to fit the data to the Frosch and Foley parameters, b and c rather than to the Fermi contact parameter (b, = b + c/3) and c, even though the latter set is physically more meaningful. It is also of interest to note that the e.s.r. study of SrF in the X2C state could only establish the magnitude of the 19F hyperfine parameters. In our work we have been able to determine that the assumed positive sign is correct, from the relative intensities of the two hyperfine components in low N transitions. The value obtained for b for SrF in the B2C+ state is of comparable magnitude to the values for CaBr (7 MHz) and CaI (18 MHz) in their B2C+ states.16 A single con- figuration description of the B state of SrF treats the molecule as ionic and places the unpaired electron predominantly in a 5po orbital on the Sr+ atom, with a closed-shell configuration at the F- atom.26 Unfortunately, the Fermi contact interaction para- meter b, is not well determined in our work because of the uncertainty in the dipolar parameter c.Nevertheless, it is clearly very small (for comparison, the Fermi contact parameter for an unpaired electron on a neutral F atom is ca.47.9 G H Z ) ~ ~ and provides strong support for the ionic model. Bernath et a l l 6 have explained their ob- served values of b for CaBr and CaI in the B2C+ states in terms of a model in which the residual charge on the M+ ion distorts the spherical symmetry of the electron distribu- tion around the X- ion and produces a non-zero hyperfine interaction by spin polari- zation. It is clear that a similar explanation would account for the small I9F hyperfine parameter for SrF in the B state. 4. CROSSOVER SIGNALS I N SATURATION SPECTROSCOPY (i) OBSERVATIONS A previous study of the A213-X2C+ system of CaCl by intermodulated fluorescence l 2 has remarked on the dramatic crossover signals observed in this spectrum. The tran- sitions involved are from the two spin-rotation components of a ground-state rotational level to a single level of the A state; the scheme is depicted in fig.3. Laser radiation at a frequency midway between the two transition frequencies can interact with a par-J . M. BROWN, D . J . MILTON AND T . C . STEIMLE 157 ticular velocity class of molecules, since it experiences a Doppler shift up in frequency for the approaching beam and a corresponding shift down for the receding beam. The remarkable feature of the crossover transitions observed in the A-X system of CaCl is that the signals are negative, that is they are of opposite sign to the normal, two-level saturation signals. Preliminary attempts to explain the sign of these signals were unsuccessful and so we re-examined them in order to establish first that they were genuine crossover signals and then to explore their experimental characteristics fur- ther.Portions of the (0, 0 ) band of the A-X transition of CaCl recorded by the inter- modulated fluorescence technique are shown in fig. 4 and 5. These are fairly broad, single scans of the dye laser output covering ca. 30 GHz. Nevertheless the linewidth 2 L 1 FIG. 3.-The three-level system involved in the formation of the crossover signals observed in the AZII-X2C+ system of CaCl. States 1 and 2 are the two-spin components of a given rotational level of the Xstate and state 3 is a single rotational level in the A state. col is the circular frequency that corresponds to the separation between states 3 and 1, hul = W3 - W,, and so on.For the crossover signal to be observed, the separation between states 2 and 1, hms, must be less than the Doppler width of the optical transitions. is narrower than in the previous work12 and there is no doubt that the “ negative ” signal falls halfway between the two spin doublets, particularly in the region of the QRl, and Q1 branches. We have investigated four characteristics of the crossover signals : For low J values, the crossover signal is much stronger than the normal two-level saturation signals (Lamb dips). The first members of the P, and ‘Q12 branches can be seen in fig. 4; for these lines, the Lamb dips are swamped by the crossover signals. As the J value increases, the Lamb dips become relatively stronger until at high J values, when the ground-state spin splitting exceeds the Doppler width, the crossover signal vanishes altogether, as expected. Fig.4 also shows some high J lines which do not show associated crossover signals. The power dependence was investigated by inserting a range of neutral density filters in the primary laser beam. For a power variation over two orders of magnitude, we could not detect a change in the ratio of the Lamb dip and crossover intensities; both signals became weaker as the power was reduced. The Lamb dip linewidth was reduced as the power was lowered (80 to 55 MHz) where- as the crossover signal appeared to have a narrower linewidth (40 MHz) that was in- dependent of power. I t was difficult to vary the pressure in the fluorescence cell used for the Ca/CaCI, system described in the earlier work.” The pressure de- pendence observations were therefore made with the Broida cell by varying the argon (a) rotational quantum number dependence.(bj laser power dependence. ( c ) pressure dependence.I58 SPECTRA OF CaCl AND SrF I I 1 1 2 ~ l - - 1- - 16 097 00 16 096.50 frequency /cm- FIG. 4.-A portion of the (0, 0) band of the A211-X2C+ system of CaCI, recorded by intermodulation fluorescence, covering the region near the origin of the Pl and ‘Q12 branches. The first few members of the P1 and ‘Ql2 branches for Ca3T1 are indicated at the bottom of the figure by the appropriate value of the quantum number N; it can be seen that the crossover signals for these transitions are negative and much stronger than the normal (two-level) Lamb dips.The corresponding signals for Ca3’C1, indicated by asterisks, show similar behaviour. Note that there is an error in the numbering of the high-N lines of the PI and ‘Qlz branches in the diagram of ref. (12). The correct numbering, given here, is higher by one than in previous work. flow rate. The crossover signals were strongest at the lowest pressures employed (ca. 0.030 Torr) and their intensity decreased roughly linearly with pressure until they became unobservable at ca. 0.25 Torr, or above. Most of our observations were made with the polar- ( d ) polarization dependence. 16 10000 16099 50 frequency jcm- FIG. 5.-A portion of the (0, 0) band of the AzII-XZI;+ system of CaCI, showing the Ql and QR,2 branches. The series of negative crossover signals is well-developed with intensity decreasing as N , and hence the ground-state spin-rotation splitting, increases.J .M. BROWN, D . J . MILTON A N D T . C . STEIMLE 159 izations of the two counterpropagating laser beams crossed as described in the ex- perimental section. We also reverted to the more conventional arrangement of parallel plane polarization by crossing the laser beams at an oblique angle in the sample region, as described before.I2 No significant change in the spectrum was observed. Before presenting the results of our density matrix treatment of the three-level system, it is worth reviewing the known examples of crossover signals recorded by saturation spectroscopy to ascertain whether the present results are unusual. We are aware of four other examples besides CaCl.For two of them (I: and NH:), the crossover signals are positive while for Na atoms2 they are negative. Some isolated negative crossover signals were observed by Curl and coworkers in the A2II,--X2l-I, system of B02,29 although there is the hint of positive crossover signals in the same spectrum, shown in ref. (1 1). In addition, we have observed indications of negative crossover signals in the B-X systems of CaCl and SrF in the present work, although we did not try to develop them more fully by working at lower pressures. From the evidence available, therefore, it would appear that a crossover signal is as likely to be negative as positive. (ii) DENSITY-MATRIX TREATMENT OF A THREE-LEVEL SYSTEM We have attempted to explain the sign of the crossover signals in the A-X system of CaCl with a density matrix treatment of a three-level system.Our treatment fol- lows the procedure described by Oka,30 which in turn is based on a paper by S h i m i ~ u . ~ ~ The energy-level scheme is shown in fig. 3. Levels 3 and 2 and levels 3 and 1 are con- nected by electric-dipole matrix elements, whereas levels 2 and 1 are not. The equa- tion for the density matrix of molecules with a particular axial velocity u is set up and solved by a perturbation procedure. The fluorescence intensity is proportional to the population of the upper level 3, which is obtained by integrating the appropriate ele- ment of the density matrix over the molecular velocity distribution. The molecules are interacting with two laser radiation fields travelling in opposite directions, + [,I74 e- i ( d + k z ) + EL.e-i(wt-kz) + c.c.1, where w is the angular frequency of the laser radiation, k is the propagation vector, z is the axial coordinate and C.C. stands for complex conjugate. The field experienced by the molecules can be rewritten + C.C.] (1) E = +[E e-i(u+ku)t + E- e - i ( ~ - k v ) t where E, = E,’ exp [ &ik(uto + z,)]. The molecules moving with axial velocity u pass the point zo at time to. The time evolution of the density matrix [p’] is given by dp’ldt = -i/h[H, p ’ ] - [r(p’ - p’(O))]. The Hamiltonian for the system is H = Ho - E and for the three-level system it is represented by the matrix -f132E -P3lE “w, 1. H = - / i S Z E W2 [_“1,,E 0 (3) I‘ is a relaxation matrix whose elements describe the decay rates between the various levels involved and p’(O) is the initial density matrix.It is anticipated that the elements160 SPECTRA OF CaCl AND SrF y33, 731 and y32 are much larger than the others because of the radiative decay from the upper level. D34 = (iE/h) b32(P123 - p‘32) + p31(P113 - P ’ ~ ~ ) ] - ~ ~ ~ ( p ’ ~ ~ - P‘~~‘’)) (4a) (4b) ( 4 4 Substitution of eqn (3) in (2) gives p2; = (iE/h) p32(P132 - p’23) - Y22 (PI22 - P’22’’’) pi1 = (iE/h)p31(P’31 - PI131 - Yll(P’11 - P‘ll‘’)) Pi1 Z= -icc)1P’31 + (iE/h)[p31(P’11 - P’33) + P~zP’z~I - ?31(p’31 - P’31‘’)) 6 4 2 = -im2p‘32 + (iE/h)[p32(p’22 - PI331 + p31p112] - y32(P132 - P‘32’’)) (4d) (4e) dil = -icosP‘21 + (iE/h)(p32P131 - p31P‘23) - y12(P’21 - p’21(o)), (4f 1 where cul = (W, - Wl)/h, etc.(see fig. 3). beams, eqn (I), is substituted in these equations. separate off the high-frequency terms, we make the replacements : The radiation field for the two laser In addition, in order to be able to pj, = p31 e-’wtl pi2 = pj2 e - W ( 5 ) pi1 = p21 e-iw,t. After some manipulation and neglecting the high frequency terms (the rotating-wave approximation), 32 we obtain d[eXP(Y33t)(P133 - P133(’))l/dt = 4 , 9 3 2 eXP(Y33f)[X+3; exp(iW t ) + x 2 exp(i% t)l -ipjl e ~ p ( y ~ ~ t ) [ x < f exp(iC2Tt) + x;? exp(iC2;t)l + C.C. ( 6 4 + C.C. (6b) + C.C. (W ( 6 4 d[eXP(Y22t1(P’22 - p’22(0))]/dt = ip32 exp(722t)[x<z* exp(iQ: t , + x;? exp(iR.5 t)] d[exp(yiit)(~’ii - ~’ii‘’’)l/dt exp(yiit)[x2 ~ x P ( ~ O T t ) + x;? exp(iQi t)] d [ e W ( ~ ~ ~ - pjl(0))]/dt = -i[x& exp(-iRit) + XG exp(-iQit)] e~31t(p’~~) - pill) + i[x& exp(-XZ,ft) + x~ exp(-iQit)] e ~ 3 1 ~ p ’ ~ ~ d [ e W ( ~ ~ ~ - p32(’))]/dt = -i[& exp(-iQZtt) + x~ exp(-iR;t)] e~3zt(p’~~ - p’J + i[x& exp(-iRit) + x~ exp(-iQ~t)] ey3~tp’~~ (6e) d [ e ~ z l ~ ( P ~ ~ - ~ ~ ~ ( ‘ ) ) ] / d t = i[x+3? exp(il2f t ) + xg? exp(iRit)] ~ ( Y z I + ~ ~ s ) ~ P ~ ~ - i[x& exp(-iR;t) + XG exp(-iR:t)] ~ ( Y z I + ~ ~ s ) ‘ P ~ ~ (6f) where xlj+ = p i j E+/2h, Q1* = (cc) - col u), etc.and u = ku, following the nota- tion used by Oka. We now solve the density-matrix eqn (6) by an iteration or perturbation procedure. The phenomenon that we are trying to describe is an optical-optical double resonance and the use of a perturbation treatment in such a case is questionable because of slow or non-convergence, as Oka30 has pointed out.For this reason we first worked out the steady-state solution to eqn (6), along the lines given by Brewer and The result is very complicated and has not provided us with much insight into the problem. The merit of the perturbation approach is that it produces terms with ex- plicit and simple dependences on the laser radiation fields. Indeed if one is looking only for terms of a particular form, one can avoid much of the algebra. For theJ . M. BROWN, D. J . MILTON AND T . C . STEIMLE 161 intermodulated fluorescence experiment, the terms required depend linearly on the intensities of both laser beams, that is they are proportional to IE+ l2 IE- l2 from eqn (1). The signal is detected in the fluorescence from the upper level 3 and so we look for terms of the required form in the diagonal density-matrix element, pi3.We adopt the following initial (zero'th order) conditions : p;l'o' 1 a 2: 5, P ; ~ ( O ) 1 1 - a, ~$20) = 0, ~ ~ ~ ( 0 ) = pS2(O) = 0. (7) These conditions are substituted in eqn ( 6 4 and (6e) to give the first-order solutions for ~ 3 1 and ~32, which in turn are substituted in eqn (6a) and (6f), and so on. The various routes through to our 4th-order solution are summarized in the diagram P32 P3 1 p 3 3 I I I I I I I I I I I I I I I I 2nd 3rd 4th zeroth ,st o r d e r FIG. 6.-A diagram to show the various paths through to the expression for p;3(4) in the density-matrix treatment of the three-level system. The orders of the stages in the iterative process are shown at the bottom.The symbols beside each arrow indicate the transition dipole responsible for connecting the two elements of the density matrix, in accordance with eqn (6) in the text. The two paths which give non-vanishing contributions to the crossover signal are p;l'o' -+p31(1) -fpg3(2) 'p32'3' +pi3'4) p;2'o' -+ p3z'1' --+pj3'2' -+p31'3' +pj3'4'. and Note that the two elements pZl and pi2 in 2nd-order are connected because p l z = pzl*. in fig. 6 . I x ~ ~ + 121x32-12 can be derived and the result is There are only four ways in which terms of the required dependence pj3(4) = Ixa+ l2 1x32- I2 Y33-I {-l/[(-iai+ + Y31)(-in2- -k Y32) 7331 -1/[(iai+ + Y3i)(-ia~2- + Y32)Y331 - l/[(ini+ + Y31)(2iu - i m ~ + Yi2)(-in22- + Y32)] - ./[(-in,- + y32)2(2iu - US + Y12)] - (1 - .)/[(-iai' + Y31)2(-2iu + ims + y12)] + c.c.} + irrelevant terms.(8) This result refers .to molecules with a particular axial velocity u. The final step in the calculation involves the integration over the Maxwellian velocity distribution N(u), where N(u) is expressed in terms of the number of molecules per unit volume N and the average Doppler shift u as N(u) = N n - ) u0-' exp[-(u/uo)']. (9) Since the crossover signal has a linewidth very much less than the Doppler width, N(u) can be assumed to be constant over the region of interest near co = +(a, + m2)162 SPECTRA OF CaCl AND SrF and taken outside the integral. The line integral may then be replaced by a contour integral and evaluated by summation of residues.30 Only the first term and its com- plex conjugate remain after the integration and the final expression for the population of level 3 is and the resonance occurs at w = $(al + w2) as expected.Note that this result is independent of a. We have described the signal arising from molecules moving with velocity u given by ku = $(col - w2). There is also a signal of exactly the same form from molecules with kv = -+(a, - to2), except that it depends on I x ~ ~ - I ~ Ix32+12. For comparison, the corresponding expression for the two-level Lamb dip at w = wl, detected by intermodulated fluorescence, is NJ4) = -4~N(O)lx~~+ 1’ Ix31- 1’ ~ ~ ~ - ~ { l / [ - i ( w - wl) + y31] + c.c.}. (1 1) Eqn (10) and (1 1) show that, to this order of perturbation theory, the crossover and Lamb dip signals are expected to have the same sign.Both signals should have a Lorentzian absorption lineshape with comparable width since 731 21 7 3 2 . Eqn (10) also supports the accepted folklore that the intensity of the crossover signal is the geometric mean of its two component Lamb dips. The question then remains as to where one should look for terms which alter the relative sign of the Lamb dip and crossover signals. We have explored all possible routes through to p;3 in the 4th order and have established that there are no other contributions to the crossover signal apart from that indicated in eqn (10). However, there are other pathways that contribute to the Lamb dip signals that are evident in fig. 6. The contribution in eqn (1 1) results from the iteration p p -+ p p -+ (p;3‘2’ - p p ) + p31(3’ + P;3’4’, but there is also a contribution from the following path, for example: We have investigated such terms but, although they alter the simple relationship be- tween the Lamb dip and crossover signals, they do not change the relative sign.Furthermore there is not a marked change in the Lamb dip intensity in the high N region of the CaCl spectrum where interference between the two transitions is not pos- sible because the spin-rotation splitting is larger than the Doppler width. The weak- ness of the perturbation treatment of the density matrix equations for double reson- ance effects is that it probably converges very slowly and may even not converge at all. It seems very likely that the explanation of the negative crossover signals lies in higher order term(s) of the expansion. The next contributions to the crossover signals arise in 8th order; we have not pursued this rather daunting calculation further at the present time.We are grateful to the S.R.C. for the purchase of equipment and the support of two of us (D. J. M. and T. C. S.). We would also like to thank Dr. Takeshi Oka for advice on the density-matrix calculation, Dr. Jim Watson for some mathematical guidance, Dr. Richard Lowe for the design details of the 1 m confocal etalonj Dr. Alan Carrington for the use of his minicomputer and Chris Brazier for much help in the laboratory. T. W. Hansch, M. H. Nayfeh, S. A. Lee, S. M. Curry and I. S. Shahin, Phys. Rev. Lett., 1974, 32, 1336. ’ T. W. Hansch. I. S. Shahin and A. L. Schawlow, Phys.Rev. Lett., 1971,27, 707.J . M . BROWN, D. J . MILTON AND T . C . STEIMLE 163 T. W. Hansch, M. D, Levenson and A. L. Schawlow, Phys. Rev. Lett., 1971,26, 946. R. Schmiedl, I. R. Bonilla, F. Paech and W. Demtroder, J. Mol. Spectrosc., 1977, 68, 236. R. E. Smalley, L. Wharton, D. H. Levy and D. W. Chandler, J . Mol. Spectrosc., 1977, 66, 375. A. Carrington, D. R. J. Milverton and P. J. Sarre, Mol. Phys., 1978, 35, 1505. M. S. Sorem, T. W. Hansch and A. L. Schawlow, Chem. Phys. Lett., 1972, 17, 300. G. W. Hills, D. L. Philen, R. F. Curl and F. K. Tittel, Chem. Phys., 1976, 12, 107. 24, 208. 70, 42. ’ M. S. Sorem and A. L. Schawlow, Opt. Commun., 1972, 5, 148. lo A. Muirhead, K. V. L. N. Sastry, R. F. Curl, J. Cook and F. K. Tittel, Chem. Phys. Lett., 1974, l1 R. S. Lowe, H. Gerhardt, W. Dillenschneider, R. F. Curl and F. K. Tittel, J. Chem. Phys., 1979, l2 J. M. Brown, H. Martin and F. D. Wayne, Chent. Phys. Lett., 1978, 55, 67. l3 P. F. Bernath, P. G. Cummins and R. W. Field, Chem. Phys. Lett., 1980,70, 618. l4 A. S-C. Cheung, R. C. Hansen, A. M. Lyyra and A. J. Merer, J. Mol. Spectrosc., to be published. E. Hirota, Chemical and Biochemical Applications of Lasers, ed. C. Bradley Moore, (Academic Press, New York, 1980), vol. 5, p. 39. l6 P. F. Bernath, B. Pinchemel and R. W. Field, J. Chem. Phys., 1981, 74, 5508. l7 N. Bloembergen and M. D. Levenson, in High Resolution Laser Spectroscopy, ed. K. Shimoda, l8 J. B. West, R. S. Bradford, J. D. Eversole and C. R. Jones, Rev. Sci. Instr., 1975, 46, 164. l9 S. Gerstenkorn and P. Luc, Atlas du spectre d’absorption de la niolecule d’iode (CNRS, Paris, 2o L-E. Berg, L. Klynning and H. Martin, Physica Scripta, 1980, 21, 173. 21 P. J. Domaille, T. C. Steimle, N. B. Wong and D. 0. Harris, J . Mol. Spectrosc., 1977, 65, 354. 22 T. C. Steimle, P. J. Domaille and D. 0. Harris, J. Mol. Spectrosc., 1977, 68, 134. 23 J. V. M. de Pinillos and W. Weltner, J. Chem. Phys., 1976, 65, 4256. 24 W. J. Childs, G. L. Goodman and L. S. Goodman, J . Mol. Spectrosc., 1981, in press. 25 R. A. Frosch and H. M. Foley, Phys. Rev., 1952, 88, 1337. 26 L. B. Knight, W. C. Easley, W. Weltner and M. Wilson, J. Chem. Phys., 1971, 54, 322. ” T. C. Steimle, P. J. Domaille and D. 0. Harris, J . Mol. Spectrosc., 1977, 68, 146. 28 P. B. Ayscough, Electron Spin Resonance in Cheniistry (Methuen, London, 1967), p. 438. 29 R. F. Curl and R. S. Lowe, unpublished results. 30 T. Oka, in Frontiers in Laser Spectroscopy, ed. Balian et a[., Les Houches, Session XXVII, 1975 31 F. Shimizu, Phys. Rev. A, 1974, 10, 950. 32 M. Sargent, M. 0. Scully and W. E. Lamb, Laser Physics (Addison-Wesley, London, 1974), p. 33 R. G. Brewer and E. L. Hahn, Phys. Rev. A , 1975, 11, 1641. (Springer-Verlag, Berlin, 1976), chap. 8. 1978); see also Rev. Phys. Appl., 1979, 14, 791. (North Holland, Amsterdam, 1977), p. 53 1. 18.

ISSN:0301-7249    

DOI:10.1039/DC9817100151

出版商:RSC    

年代:1981

数据来源: RSC

摘要:

Rydberg Spectra of Triatomic Hydrogen and of the Ammonium Radical BY GERHARD HERZBERG National Research Council of Canada, Ottawa, Ontario KIA OR6, Canada Received 2 1 st Notiember, 1980 In hollow-cathode discharges through H2 and D2 a number of new emission bands have been observed which have been identified as belonging to triatomic hydrogen (H3 and D3) in Rydberg states. The information about the structures of these radicals in the upper and lower states of the observed bands is summarized. A related Rydberg spectrum is expected and observed for the NH4 radical. Spectra first observed by Schuster in 1872 and by Schuler, Michel and Griin in 1955 in discharges through ammonia are shown by isotope studies to be due to the free ammonium radical. The spectra of ND4 are more intense and sharper than those of NH4.A full analysis has not yet been accomplished but there is little doubt that NH4 is tetrahedral in its Rydberg states. The geometry is very similar to that of the H3+ ion. INTRODUCTION The emission spectrum of the He, molecule, unlike those of almost all other dia- tomic molecules, has been known for almost fifty years to be a Rydberg spectrum, i.e., to consist of a number of Rydberg series. The reasons for this peculiarity are the facts that the ground state of He, is unstable, that there are no low-lying valence states and that the lowest stable state is a Rydberg state. More recently a similar diatomic system, ArH, was discovered by Johns' who observed a transition between the two lowest Rydberg states. Still more recently transitions have been observed from excited states of dimers (excimers) to the un- stable ground state.Such transitions have become very important as a new group of lasers (excimer lasers). Here, however, the excited states are not (or not neces- sarily) Rydberg states and certainly the lower state is not a Rydberg state. The first type of system, exemplified by He, and ArH, might be called a Rydberg molecule or radical to distinguish it from the excimer molecules but neither designation is without objection. The first example of a polyatomic system with stable Rydberg states and unstable ground state (Rydberg radical) is triatomic hydrogen whose spectrum was discovered only two years ago. In what follows I shall summarize the results that have been obtained so far on triatomic hydrogen and then proceed to a discussion of new results on a similar system: the NH4 radical.The spectrum of this radical although known for many years has been recognized as such only during the past year; the interpreta- tion of the spectrum is still incomplete. Here also the ground state is unstable.166 TRIATOMIC HYDROGEN A N D THE AMMONIUM RADICAL THE SPECTRA OF H3 A N D D3 (a) DISCOVERY In the spectrum of a hollow-cathode discharge through H2 and D, broad features were observed2 near 5600 and 7100 A. The former of these features, particularly at liquid-nitrogen temperature and in D2, shows a simple P, Q, R branch structure which was recognized as a [I (parallel) band of H3 or D3 since the resulting B value is very close to the predicted value for the ground state of H3+ or D3+.It is surprising that neither these features nor others subsequently observed had been seen previously in the many studies of hydrogen spectra during the last sixty years. (b) PARALLEL BANDS Since an electron in a Rydberg orbital cannot appreciably change the binding in a molecule, it follows that in the Rydberg states of H3 or D3 the geometrical structure must be close to that in the ground state” of the ion H3+ or D3+ which both from theory (Carney and P ~ r t e r ) ~ and experiment (Oka)4 is that of an equilateral triangle (ie., point group D3J. lndependent proof for the D3,, structure of H, and D, comes from the analysis of their spectra. The 0-0 bands of electronic transitions in a 03h molecule are either ]I or 1 (per- pendicular) bands, that is, have AK = 0 or AK = h l , depending on whether the transition is between two non-degenerate states (A2”-A1’, A,’-“’‘, .. .) or between a non-degenerate and a degenerate state (E’-Al’, E”-Al”, . . .). Transitions between two degenerate states give rise to either j j or bands depending on whether they have opposite or the same symmetry with respect to reflection at the plane of the molecule. Two 11 bands have been observed for both H3 and D,, one near 5600 8, with rather broad lines (linewidth ca. 25 cm-’ for H,, ca. 7 cm-’ for DJ, the other near 6025 8, consisting of fairly sharp lines. These bands have been analysed in detail by Dabrow- ski and Herzberg5 who have obtained the rotational constants B,, Co and internuclear distances yo given in table 1.The closeness of these constants to those of H,+ and D3+ as predicted by Carney and Porter3 and given in the last row of table 1 was the first strong indication that these new bands are emitted by neutral H, and D3 in Ryd- berg states. symmetry have a characteristic intensity alternation for the K = 0 sub-band of I/ bands: for a totally symmetric lower state A,’ the even rota- tional lines are absent for Fermi statistics of the identical nuclei and nuclear spin I = 3, ie., in H3, while they have 10 times the intensity of the odd rotational lines for Bose statistics of the identical nuclei and spin I = 1, i.e., in D3. The observation of such an opposite intensity alternation in the / / bands represents final confirmation of the carrier of the spectrum.As an illustration plate 1 [taken from ref. (5)] shows the observed intensity distribution in the 5600 8, bands of H, and D3 compared with that reconstructed from the final constants. Even though here the K structure is not resolved the effect of the intensity alternation for K = 0 is clearly exhibited. Note especially that P(1) is strong in H3 but effectively absent in D,. For the 6025 A band the K-structure is fully resolved, but the intensity alternation is reversed from that in the 5600 A band showing that the lower state is not A,’ but A2” (see fig. 2 and 3 of ref. (5). Molecules of * There are no other low-lying states of the ion.PLATE 1 .-(a) Photometer curves of the 5600 A bands of D3 and H3. (b) Intensity curves of the 5600 A bands of D3 and H3 reconstructed from the molecular constants determined from (a) assuming a linewidth of 7 and 30 cm-', respectively, and a temperature of 300 and 200 K, respectively.[To face page 166PLATE 2.-Schuster bands of ordinary and heavy ammonia here assigned to the ammonium radical. The sharp lines in the upper spectrum are due to NH,: the Schuster band of NH4 is entirely diffuse. There are no ND, lines in the lower spectrum. [[To face page 1 67G . HERZBERG 167 (C) PERPENDICULAR BANDS Up to now two _L bands have been analysed, one in the red near 7100 A (Herzberg and Watson)6 and one in the infrared near 2.8 pm (3600 cm-') (Herzberg, Lew, Sloan and Watson).' The upper state of the former is identical with the lower state of the latter. The lower state of the 7100 A band is the same as that of the 5600 8, [I band while the upper state of the infrared band is the same as that of the 6025 A 11 band.The analysis of the I_ bands proved to be much more difficult than that of the [I bands. For the 7100 8, band there is the additional handicap of the large linewidth. The spacing of sub-bands in a _L band of an oblate symmetric top (quite unlike a [I band) is 2[C(l - &B] if B'-B" and C'-C" are small. Since in the present case ( N" 1 the spacing of sub-bands is ca. -2B and as a result lines with the same N - K of different sub-bands lie close together, that is, form pseudo-branches nearly coinci- dent with the Q branches. Such a structure is not obvious in the 7100 A band since the linewidth is large and B' - R" is not small, but is clearly observed in the infrared band. Even here it is complicated by the occurrence of A-type doubling for the K = 1, G = 0 rotational levels as well as by A-type resonance between levels with A K = 2 and the same G.Even the normal A-type doubling is fairly large for such a light molecule but here it is made several times larger by the effect of Jahn-Teller interaction in the degenerate electronic state (3p 2E') leading to considerable irregu- larities in the branches and pseudo-branches so that in several instances they are difficult to recognize. These difficulties were overcome by the fitting of model bands to the observed band structure. The principal constants obtained in this way are included in table 1. TABLE ROTATIONAL CONSTANTS IN THE RYDBERG STATES OF H3 AND D3 COMPARED WITH THOSE OF H3+ AND D3' ~~ ~~ 3p 'A 2" 45.50 22.75" 0.8575 22.257 10.782 0.8672 3s 'Al' 44.194 22.676 0.8700 21.985 12.406 0.8726 3p 2E' 42.15db 21.505 0.8908 21.1 50' 10.594 0.8896 2p 2A2'' 44.575 22.28aa 0.8663 22.1 12 1 1 .O w 0.8701 ion 'Al' o ~ s . ~ 43.56a 20.708 0.8763 (ca1c.J 43.23 20.57 0.879, 21.61 10.44 0.8801 2s 'Al' 46.82 23.41" 0.8453 22.989 11.495" 0.8533 a From C = B/2. ( = 0.9222, q = 6.829 an-'. [ = 0.8914, q = 2.966* cm-'. ( d ) ELECTRONIC STRUCTURE It is easy to understand the observed electronic states in terms of a molecular- orbital picture. The orbitals in &, symmetry are lsal' 2sa; 2pe' 2pa2" 3sal' 3pe' 3pa2" 3de' 3de" 3dal' . . . . In all states two of the electrons fill the lowest orbital lsa,' as in the ground state of H3+; the third electron can occupy any of the remaining orbitals.The resulting states of H, are given at the left in fig. 1. The lowest state 2p 2E' is unstable (repul- sive) and describes the reaction of H + H2 in their ground states. The two next168 TRIATOMIC HYDROGEN A N D THE AMMONIUM RADICAL lowest states, 2s 2Al’ and 2p 2A2”, are the lower states of the emission bands observed in the visible region. The upper states of these bands are the n = 3 states 3s 2Al’ (6025 8, band), 3p 2E’ (7100 8, band), 3p 2A2’r (5600 8, band), 3d2E’, 2E”, 2Al’ (5900 A bands, not yet fully analysed). The two infrared bands are transitions within the eV H+ t H,(X’Zg+)+e I I I4 2 10 8 6 4 2 0 FIG. 1.-Observed electronic states of H3 at left and of the dissociation products at right.The excited states of H2 are not included; the lowest of these would be at 11.18 eV. n = 3 manifold with 3p 2E’ as lower state and 3s 2Al’ and 3d ’E’, 2E”, 2Al‘ as upper states. The group of states derived from the 3d orbital are complicated by the tendency of the electronic angular momentum vector to uncouple from the symmetry axis with increasing N . This uncoupling is similar in character to the well-known I-uncoupling in 3d C, n, A states of diatomic molecules, e . g . , He,, and is being studied by 5. T. Hougen and J. K. G. Watson. Further transitions, for example n = 4 -+ n = 2, haveG . HERZBERG 169 been looked for but not observed. They are probably hidden by the ordinary con- tinuum of H, or D,. At the right-hand side of fig.1 the energies of the dissociation products H + H, are represented. It is significant that the lowest excited state of H + H,, viz. H(n = 2) + H2(X lC,+) lies above the energy of the series limit of the Rydberg levels of H,, i.e., above the energy of the ground state of H3+. As a result all the Rydberg states of H3 (other than 2p ,E') have substantial dissociation energies since they can only correlate with much higher lying states of H + HZ. Only the lowest state 2p 2E' correlates with the ground state of H + H,. In other words, because of the large dissociation energy of H3+ [D,(H3+ + H+ + H,) x 4.5 eV] all the Rydberg states (except 2p 2E') are stable bound states. For all Rydberg states we must consider the possibility of predissociation into the repulsive 2p 2E' ground state.Because of favourable Franck-Condon factors this predissociation is most likely to occur in the lowest bound Rydberg state 2s ,A1'. It is vibronically allowed if in either the 2s 'A1' or the 2p 'E' state the degenerate vibra- tion is excited and is indeed observed by the strong broadening of all lines in the bands that have 2s 'AI' as lower state. The second low-lying bound state 2p 'A,'' can pre- dissociate into 2p 'E' only by ro-vibronic interaction, which is much weaker than vibronic interaction and vanishes for J = 0, K = 0. In accordance with this predic- tion the lines of the H3 bands with the 2p 2A2" lower state while fairly sharp for lower N show an increasing broadening at higher N . Actually the broadening is propor- tional to N(N $- I ) - K 2 .For the corresponding D, bands the broadening caused by predissociation is vanishingly small but there is a remaining anomalous Doppler broadening [see ref. (5)]. The n = 3 states of H, and D, must clearly have very much smaller Franck- Condon factors for predissociation into the 2p 2E' state and it is therefore not surpris- ing that emission from these states is observed. Whether predissociation is effectively absent in these states can only be decided by measurement of the radiative lifetime and comparison with predicted lifetimes as given by King and Morokuma.8 state shows clear signs of Jahn-Teller interaction: a much larger than normal A doubling constant and a quenching of the electronic angular momentum. The minimum of the Jahn-Teller distorted potential function is found to be 87 cm-' below the undistorted energy.This corresponds to a shift of the equilibrium position by 0.026 A from the undistorted configuration. The zero- point energy level is well above the vertex of the cone corresponding to D,, symmetry. As emphasized earlier the 3p THE SPECTRA OF NH, AND ND, (a) INTRODUCTION The characteristic which makes possible the existence of stable Rydberg states of H3 is the high proton affinity of the parent molecule H,, that is, the high stability of the H3+ ion. There are many other molecules with high proton affinities and one may expect that, if an electron is added to the corresponding ions, stable Rydberg states of the neutral adducts result. For example, the great stability of the ions H,O+, NH,+ and CH5+ is well-known; the proton affinities of H20, NH3, CH, are 7.1, 8.8 and 5.3 eV, respectively. Therefore Rydberg states of H30, NH, and CH5 would be expected to be bound and corresponding Rydberg spectra may be observable.In one case, that of NH,, such a spectrum has in fact been identified as will be discussed presently. The stability of the ground state of NH, has been the subject of much dis- cussion. Bernstein9 was probably the first to suggest some stability of this state, namely D(NH, -+ NH3 + H) % 20 kcal mo1-'.170 TRIATOMIC HYDROGEN A N D THE AMMONIUM RADICAL (b) SCHUSTER BANDS The Schuster band of ammonia, a broad continuous emission feature with two humps at 5672 and 5639 A occurring in discharges through NH,, was first very briefly described by Schuster lo in 1872 at a meeting of the British Association for the Advance- ment of Science. In the last hundred years it has been observed by a number of investigators, most recently by Schuler, Michel and Grun," but none has obtained a definite identification. The last-mentioned authors have found several additional bands, equally broad and of similar intensity, near 5282, 6497 and 7666 A, which appear to belong to the same system.Furthermore, at higher pressure of NH, they have observed a new system of bands, the strongest at 6635 A, which we shall call the Schuler bands and which will be discussed in more detail further below. Schuler et al. have also obtained corresponding spectra, both Schuster and Schuler bands, with ND3. The isotope shifts are considerable, 440 and 250 cm-l, respectively, for the principal bands.In a note added in proof Schuler et al. mention that they have done some experi- ments with the mixed isotopes NH2D, NHD2 in addition to NH, and ND, and that these experiments indicate that " probably four H atoms are present in the carrier of both spectra " (i.e., the Schuster and Schuler bands). On the basis of this state- ment combined with the conclusion of the main part of the paper of Schuler et al. one would be led to believe that these authors considered it likely that the carrier of both spectra was N2H4 although they do not explicitly state this conclusion. We have repeated the experiments with partly and fully deuterated ammonia and have con- firmed the result of Schuler et al., namely that there are three intermediate bands in addition to those produced with NH, and ND, and that therefore there must be four H atoms present in the carrier of the spectrum.In agreement with Schuler we find the spectrum obtained with ND, much stronger (by a factor of 20) than that obtained with NH, while those with intermediate isotopes Lhave intermediate intensities. It is therefore difficult to reproduce a single spectrum that shows all five isotopic species, but this has recently been accomplished. The dispersion and resolving power that we were able to use (2.57 A mm-') was much greater than that used by Schuler, Michel and Griin (not stated, but probably ca. 30 A rnm-l). While the Schuster band of NH, remains unresolved, that is, is genuinely diffuse, the corresponding band of ND, shows a clear fine structure.This is shown in plate 2. None of the intermediate isotopes shows a fine structure. The situation is not unlike that found for the absorption bands of CH3 and CD, near 2150 A and their intermediate isotopes.12 Clearly, in the present case there is a strong predissociation in the lower state. Even for the band observed in ND, [plate 2(b)] the individual fine-structure lines are distinctly broad. The " lines " are relatively widely spaced, a fact that would be difficult if not impossible to account for on the basis of the assumption that N2H4 is the carrier of the spectrum. Actually spectra taken with 15ND3 yield a very similar spectrum very slightly shifted to longer wavelengths compared with that obtained with 14ND3.In a 50:50 mixture of 14ND3 and 15ND, only a superposition of the two bands obtained with pure 14ND3 and pure 15ND, resulted with no trace of an intermediate isotope which should have been obtained if the molecule responsible were N,H4. We conclude therefore that the Schuster band is due to NH,, the free ammonium radical. On the basis of the observed vibration spectrum it is generally agreed that the NH,+ ion has tetrahedral structure (point group Td). If an electron is added in a Rydberg orbital one may expect the same Td structure for the resulting neutral NH, radical in its Rydberg states. As is well-known, the symmetry types of point groupG . HERZBERG 171 Td are A,, AZ, E, F, and F2. Allowed electronic transitions are F2-A,, F,-A,, E-F,, E-F,, Fl-Fl, F1-F2 and F2-F2.In view of the closeness of the Schuster band to the Na D lines it is tempting to interpret this band as a 3p-3s transition in NH,, that is, an F2-A, transition. Vibra- tional transitions of this type are well-known for CH, and other Td molecules. How- ever F2-Al electronic transitions have so far not been analysed. Triply degenerate states are characterized by a first-order Coriolis splitting first discussed by Teller.13 This splitting yields three sets of rotational levels which in a first approximation are given by F+(J) = BJ(J + 1) + 2BC(J + 1) Fo(J) = BJ(J + 1) (1) F - ( J ) = BJ(J + 1) - 2BcJ where [ is here the electronic angular momentum. In this approximation, of the three sets of levels only F - combines with a non-degenerate state in the R branch (AJ = + l), only Fo in the Q branch (AJ = 0) and only F f in the P branch (AJ = - 1).If [ z 1 as is expected for a state derived from a p-electron, it is easily seen that, in the approximation of eqn (I), all three branches (P, Q, R) coincide. The observed Schuster band of ND4 [plate 2(b)] does show a strong central branch which might be interpreted as the superposition of P, Q, R branches. The remaining much weaker and more widely spaced branches would then have to be interpreted as the forbidden branches arising from the combination of F2- with Al by way of AJ = - 1 (P branch) and F2+ with Al by way of AJ = +1 (R branch) yielding branches of 0 and S form, that is, with a spacing of 4B. Although the observed spacing in the ND, spectrum fits roughly with 4B, the intensity of these forbidden branches seems much too high.In addition the head of the combined P, Q, R branch which should be halfway be- tween the extrapolated 0(1) and O(0) lines does not seem to fit this prediction. At this point J. K. G. Watson pointed out to me that a value of c near -1 would lead to allowed P and R branches of 0 and S form (spacing 4B) with a normal Q branch. Such an interpretation also leads to a relation between the head of Q branch and P and R series [uiz. at the extrapolated position of O(l)] which is readily fitted to the observed spectrum. In a higher approximation the B values for F+ and F - in eqn (1) are slightly dif- ferent from that for Fo by an amount of the order of the rotational constant a.As a result the coincidence of the P, Q and R branches for the [ = 1 interpretation is most unlikely to be complete, throwing further doubt on this interpretation of the Schuster band. Indeed the B’-B” obtained from the second difference of the observed central branch is substantially different from that obtained from the P and R branch. It therefore appears very likely that the ’F2 state involved in the Schuster band has c = -1. The B’ and B” values obtained from the best fit are Bk = 2.5,, BiR = 2.65, B” = 2.3 cm-l. These B values cannot claim much accuracy because of the width of the lines (partly caused by the tetrahedral splittings for higher N values) and because of the need to assume 5 = - 1. Depending on how close the actual [ is to this value slightly different B values would result.( c ) SCHULER BANDS The Schiiler system of bands, while weak compared to the Schuster bands at low pressure, predominates at high pressure. Schiiler et al. used electron beam excitation at 1 atm pressure. We used Tesla coil excitation at pressures up to 0.5 atm. In172 TRIATOMIC HYDROGEN A N D THE AMMONIUM RADICAL plate 3 the principal bands at 6635 A for NH3 and 6750 A for ND, are shown at a reci- procal dispersion of 0.28 A mm-I on the enlargement. As for the Schuster band, the Schuler band in ND, is very much stronger than that in NH3. For the latter, in order to obtain a medium resolution spectrum, we had to use an image intensifier. This together with the greater widths of the lines accounts for the poorer quality of the NH, compared to the ND, spectrum. Again experiments were carried out with partially deuterated ammonia which show conclusively that four H atoms are present in the molecule responsible for this spectrum.Furthermore, we have again taken spectra with I5ND3 as well as a 50: 50 mixture of 15ND3 and 14ND3. These are compared with the 14ND3 spectrum in plate 4. There are only two isotopic spectra in the mixture showing that only one N atom is present in the molecule. Thus we conclude that the ammonium radical (NH4 or ND,) is the carrier of the Schuler band as well as of the Schuster band. We have also obtained the ND, Schuler band with our 10 m grating spectrograph. Even at this resolution the individual rotational lines are as sharp as those of the com- parison spectrum, A number of branches are readily identified in the spectrum but the interpretation of these branches is not obvious.There are two very pronounced heads which are separated by 6.24 cm-' in ND, and 6.92 cm-I in NH4. It seems very likely that this splitting is a spin splitting since it does not change appreciably between NH4 and ND4, and since it is of the same order as (but somewhat smaller than) the doublet splitting of the upper state of the Na D lines (17.20 cm-l). The problem of the interaction of spin and rotation in tetrahedral molecules has not yet been discussed in the literature but J. T. Hougen has taken the first steps in this direction and is collaborating in the detailed interpretation of the Schiiler band. ( d ) ELECTRONIC STRUCTURE It is easy to resolve the united atom orbitals into those of a molecule of Td sym- metry.One finds lsa, 2sa, 2pf2 3sa, 3pf2 3de 3df2, . . , . In the ground state of NH4+ the orbitals are filled up to ( 2 ~ f ~ ) ~ . The additional elec- tron in NH, may go into the orbitals 3sal, 3pf2, . . . yielding 2A1, 2F2, 2E, 2F2 electronic states. In spite of preliminary ab initio calculations of the relative positions of these states by King and Havriliak', and Wright15 an unambiguous assignment of the Schuster and Schuler bands has not yet been obtained. It is clear from the enormous difference in linewidth that the two bands cannot have the same lower state and from the very different behaviour as a function of pressure that they cannot have the upper state in common. The large width of the Schuster band of NH4 makes it likely that the lower state is the weakly bound ground state 3s 2A1 of NH,; but the upper state cannot be 3p 2F2 since this state would be expected to have an appreciable spin doubling which the Schuster band clearly does not have and since would have to be near + 1 while the band structure strongly suggests -1.The lowest state that can have [ = -1 is 3d 2F2.16 This state would have only small spin doubling. Therefore we tenta- tively assign the Schuster band to 3d 2F'2 -+ 3s *A1. The Schuler band, in view of the observed sizeable spin doubling, must in all prob- ability involve the 3p 2F2 state since this is the only state expected to have a large spin doubling. This state must be the lower state of the transition since the only state lower than this is the ground state which has already been invoked for the Schuster band.The upper state of the Schuler band could be any of the higher states which cand I 7 PLATE 3.-Schuler bands of ND4 and NH4. The spectrogram of NH4 was taken at lower dispersion (and with an image intensifier) but enlarged to the same scale as the spectrogram of ND4. A very few sharp lines in the NH4 spectrum are due to NH2. [To face page 172m n "_ z M .. M n - n Z z Ir: 'D_ PLATE 4.-Schiiler bands of ND4 obtained with I4ND3, a 50:50 mixture of 14ND3 and 15ND3, and 15ND3. [To facepage 173G . HERZBERG 173 combine strongly with 3p 2F2. On the basis of the similarity with Na we suggest 3d 'E, in other words, the Schiiler band would be 3d 2E + 3p 2F,.The proposed assignment of the Schuster band implies that it is a forbidden transition in the united atom; the proposed assignment of the Schiiler band gives rather poor agreement with the preliminary ab initio calculations. A more definite assignment will have to await the results of more elaborate ab initio calculations. A decision may also be possible on the basis of a detailed analysis of the Schiiler band. The large difference in the intensity of the Schuler and Schuster bands of NH, compared to those of ND4 can only be understood if there is a difference of lifetime of the upper state brought about by a weak predissociation which would be stronger for NH4 than for ND,. A diatomic example of the same effect is the difference of intensity of the 3140 A band of CH uersus that of CD.I7 I am very much indebted to Drs. J. K. G. Watson and J. T. Hougen for a number of clarifying discussions, to Dr. A. E. Douglas for a critical reading of the manuscript and to Mr. B. Hurley for taking all the spectra underlying this discussion. J. W. C. Johns, J. Mol. Spectrosc., 1970, 36, 488. G. Herzberg, f. Chem. Phys., 1979, 70, 4806. G. D. Carney and R. F. Porter, f. Chem. Phys., 1976, 65, 3547. T. Oka, Phys. Rev. Left., 1980, 45, 531. I. Dabrowski and G. Herzberg, Can. J. Phys., 1980, 58, 1238. G. Herzberg and J. K. G. Watson, Can. J. Phys., 1980, 58, 1250. G. Herzberg, H. Lew, J. J. Sloan and J. K. G . Watson, Can. f. Phys., 1981, 59, 428. H. F. King and K. Morokuma, f. Chem. Phys., 1979, 71, 3213. H. J. Bernstein, f. Am. Chem. SOC., 1963, 85, 484. H. Schuler, A. Michel and A. E. Griin, 2. Natut-forsch., T d A , 1955, 10, 1. E. Teller, Hand- und Jahrbuch d. Chem. Phys., 1934, 9, 11, 43. J. Wright, unpublished. lo A. Schuster, Rep. Brit. ASSOC., 1872, p. 38. I t G. Herzberg, Pt-oc. R. SOC. London, Ser. A, 1961, 262, 291. l4 H. King and S. Havriliak, unpublished. I6 J. K. G. Watson, personal discussion. l7 G. Herzberg and J. W. C . Johns, Astropliys. f., 1969, 158, 399.

ISSN:0301-7249    

DOI:10.1039/DC9817100165

出版商:RSC    

年代:1981

数据来源: RSC

摘要:

The Jahn-Teller Distortion in the F2E State of sym-C,Cl,F$ as Determined from Laser-induced Fluorescence Studies BY TREVOR J. SEARS, TERRY A. MILLER AND 17. E. BONDYBEY Bell Laboratories, Murray Hill, New Jersey 07974, U.S.A. Receired 5th December, 1980 Both gas-phase and matrix laser-induced fluorescence studies of sym-C6CL3F3+ provide detailed information on the energy-level structure of the Jahn-Teller perturbed e’ vibrations in the ground electronic state. These experimental results are analysed in terms of a multi-mode Jahn-Teller theory to yield distortion constants, unperturbed oscillator frequencies, stabilization energies and distorted geometries. These results are combined with similar results for the symmetrical ions C6F3HJ+, C6Cl3H? and C6F2 to reach some general conclusions about the quantitative nature of the Jahn- Teller distortion of benzenoid cations.INTRODUCTION During the past five years there has been a great increase in our knowledge and under- standing of the spectra of molecular ions. Of the larger species studied, much interest has been shown in the spectroscopy of halogeno-substituted benzene ions which fluoresce strongly in the visible wavelength region. The spectroscopic studies show a clear distinc- tion between the D3,, or D6h species, which possess doubly degenerate ground states, and the cations of lower symmetry. In the latter group, the vibrational structures both in the ground state and in the fluorescing state are quite regular and harmonic. In contrast, the ground- state level structure in the former, symmetric cations, appears to be highly irregular as a result of Jahn-Teller distortion^.^-^ Very recently we have shown how the spectra of the symmetrically trisubstituted ions, C6CI3H3f, C6F3H:6 and C6F,+’ may be interpreted in terms of a model Jahn-Teller Hamilton- ian which includes mixing between the various Jahn-Teller active modes * and both linear and quadratic coupling between the electronic and nuclear motions.In this paper we des- cribe how a similar treatment enables an understanding of the available spectral data relating to the ground 8‘ E state of the sym-trichlorotrifluorobenzene cation. In the final section we show how these data, along with those from the other symmetrical ions investigated, lead to some fairly conclusive generalizations about the distortion of the benzene ring in the degener- ate ground states of the symmetrical ionic species.Much of our detailed knowledge of the vibronic structure of the ground electronic state of C6C13F,f derives from the laser-induced emission spectrum of the ion in a neon matrix at 4 K;9 however, detailed analysis of the gas-phase laser-induced fluorescence excitation spec- trum lo leads to the assignment of some hot bands which both reinforce the matrix observa- tions and define the positions of some vibronic levels which were not observed in the matrix study. Finally, some wavelength-resolved LIF spectra taken in the gas phase following laser excitation of Jahn-Teller active vibrations in the upper electronic state provide cor- roboration of the results of the previous two experiments but do not add to the data set.Fig. 1 summarizes the known ground-state vibronic energy-level structure of this ion. The figure includes only those levels which are uniquely associated with the Jahn-Teller active vibrations of e’ symmetry; these are labelled Vg-t’lJ in our numbering scheme.12 The totally symmetric modes v2 and v3 also show spectral activity but do not add to our understanding of the Jahn-Teller effect in the ion.176 JAHN-TELLER DISTORTION I N CbC13FZ OBSERVED CALCULATED '"""I 700 p 74 3 2 v ('=.L) - 13J 2 ===---- - - --_ m y - - - - - - - - - 600 c FIG. 1.-Comparison of observed and predicted Jahn-Teller active levels below 1000 cm-'. The calculated levels indicated by dashed lines have intensities (for emission from the vibrationless level for the b state) less than 1 of the strongest observed emission line (vo0) and hence would not be expected to be observed experimentally.A few Jahn -Teller active levels above 1000 cm-' have been identified experimentally, but a reasonable calculation of them would require m. 100 eigenvalues of a 4-mode matrix, which is impractical computationally. 2. THEORETICAL We have recently 6 , 7 described the form of the vibronic coupling Hamiltonian necessary For a for the understanding of the ground-state energy-level structure of this type of ion. single doubly degenerate Jahn-Teller active mode i the Hamiltonian may be written A; = H T + (27~~w')Q + Q - + 2 ~ ( D / i o ) ~ Q - + K ( ~ ~ c ~ c ( > ~ ) Q : 4- H.C. (1) I? this equation, the first two terms represent the usual harmonic oscillator operators with HT the kinetic-energy part and Q+ and Q.- the complex-conjugate combinations of the two components Qla and Qib of the normal coordinate i.o is the harmonic-oscillator frequency. The second two terms in eqn (1) introduce the linear (D) and quadratic ( K ) coupling para- meters, H.C. indicates the Hermitian conjugate of these terms. Explicit definitions and further discussion of the form of this Hamiltonian are contained in ref. (6) and (7). The Hamiltonian in eqn (1) is for a single, isolated vibrational mode and in order accu- rately to describe the vibronic level structure of C6C13F,f, which has more than one mode of the correct symmetry to show Jahn-Teller activity, we must sum the operators for each mode: H = Y l?:.(2) 7T . J . SEARS, T. A . MILLER A N D V . E . BONDYBEY 177 The ionic energy levels are described by the eigenvalues of this Hamiltonian and the problem reduces to the diagonalization of the matrix of the Hamiltonian set-up in a suitable basis set. That chosen’ is one containing products of the electronic wavefunction ]A) and 2-dimensional harmonic oscillator functions luili> for each active vibrational mode, i. The analysis proceeds by adjusting the parameters mi, Di and Ki for each Jahn-Teller active mode until a satisfactory fit is obtained between the observed and predicted energy levels. 3. RESULTS C6CI3F3f has in principle seven Jahn-Teller active modes of e’ symmetry: however, the spectral evidence 9 7 1 0 shows that they do not participate equally towards the Jahn-Teller stabilization of the ion.Modes 8, 12 and 13 appear strongly active whereas 9, 10 and 14 are evidently less so. Mode 11 was not observed in any of the spectra and presumably contri- butes negligibly to the distortion. The basis set required to perform an accurate calculation for six interacting doubly degenerate modes is prohibitively large7 and the available experi- mental data are not extensive enough to allow determination of the number of parameters which would appear in the Hamiltonian. The calculated energy levels shown in fig. 1 derive from a four-mode calculation which included modes 8, 12, 13 and 14, and the final parameters are given in table 1. TABLE 1 .-JAHN-TELLER COUPLING PARAMETERS FOR f 2 E ” C6C13F: - mode mo,/cm-’ Di Ki r,(linear)/cm- 8 1550 0.15 0 233 12 390 0.60 0 234 13 310 0.32 0 99 14 185 0.05 0 9 ci(total) = 575 cm-’ The values in fig.1 correspond to the lowest eigenvalues of a matrix consisting of all vibra- As can be demon- Thus, the matrices tional levels with 0s vlz < 7, 0 d ~ 1 3 d 5, 0 d v g d 3 and 0 < ~ 1 4 d 2. strated a posteriori (see below) the quadratic coupling is quite small. may be factored l 2 into blocks according to the quantum number j given by I The number of basis states in this matrix is 7236 for j = 1/2 and 6967 for j = 3/2. The diagonalizations were accomplished on a CRAY-1 computer in ca. 150 s using the procedures previously de~cribed.~ Besides the eigenvalues of the matrix, one also obtains the corres- ponding eigenvectors.According to the assumptions of Longuet-Higgins, l 2 transitions of the form X,” have an intensity proportional to the square of the coeficient of the Nth vibra- tional level of mode X , mixed into the vibrationless level by the Jahn-Teller distortion. Similarly, transitions X i depend upon the squared eigenvector of the vibrationless level in the N level of mode X . Comparisons between the predicted and observed intensities for the Jahn-Teller active modes are given in table 2. Clearly there is excellent agreement between the predictions and observations. One may notice that the parameters listed in table I are quite different from the Jahn- Teller parameters given in ref. (9). The principal reason for this is that the present analysis was performed using the (correct) multi-mode calculation rather than treating each mode separately.The multi-mode analysis causes the re-assignment of a few lines in the matrix emission spectrum. These re-assignments are found to be consistent with the positions of severalj = 3/2 levels (see fig. 1 ) which have only been observed in the gas phase. The present work, compared with the older work,’ demonstrates two very important points about Jahn-178 JAHN-TELLER DISTORTION IN C,jClsFJ+ Teller analyses. It is quite important to have as large a data set as possible, for interpreta- tions based upon non-redundant data may lead to serious errors in assignments. Secondly, if more than one vibrational mode shows significant Jahn-Teller activity, a multi-mode analy- sis must be performed.Otherwise the final results will have no physical significance. There is no evidence in the present data set for any quadratic coupling effects and all the Kt were set to zero. The related ions C6F3HZ6 and C6F6+’ showed only very small quadratic coupling effects and C6C13H3+ displayed none,6 so the result in this case was not unexpected. We estimate that the errors in the individual linear coupling parameters are of the order of TABLE 2.-cOMPARISON OF EXPERIMENTAL AND PREDICTED INTENSITIES NORMALIZED TO 100 FOR THE ORIGIN BAND (Ne MATRIX RESULTS”) transition observed predicted emission” voo v1: V12” V14” + v12” 2v13w V13“ V14” + V13n V13” f v12” 100 2 12 29 7 3 6 4 100 2 8 36 7 2 6 6 a From the vibrationless level of the excited electronic state. &20%.These errors are greater than those reported in previous analy~es.~.~ The approxi- mations inherent in the neglect of modes 9 and 10 in this molecule are more serious than was the case in the related trisubstituted ions6 where these modes show much weaker activity. The parameters derived for mode eight can be thought of as “ effective ” in that they certainly contain contributions from the two higher-frequency modes omitted from the model. Haller et ~ 1 . ‘ ~ have shown how the effects of two Jahn-Teller active modes can be treated in terms of a single effective one by use of perturbation theory. A similar treatment could be used in the present case; however, the experimental data are not sufficient to warrant such an analysis. A satisfying feature of the present analysis are the values obtained for the unperturbed oscillator frequencies (mi) which agree closely with those of the parent l4 and the B2A{ state of the i~n.~*’O The available evidence suggests that the in-plane force fields for the parent and first two ionic states in this type of molecule are very similar.In table 1 we also give the calculated Jahn-Teller stabilization energies for each mode and the total depth of the poten- tion below the energy of the symmetrical nuclear configuration. For the reasons given above, this is likely an underestimate of the energy; however, this underestimation is likely to be only slight. 4. GENERALIZATION O F RESULTS FOR BENZENOID CATIONS sym-C6C13F$ is the fourth halogenosubstituted benzene with either D3,, or D6a symmetry that we have studied.As noted above, previous work has been performed on C6F6+ and sym- C6F3HZ and C6C13H3+. Unfortunately, the benzene cation itself is not amenable to study by this technique due to its very low quantum yield for fluorescence. Thus, a crucial question to be answered is whether the Jahn-Teller distortions that have been ob- served and analysed are strongly affected by different substituents or are mainly determined by the benzene ring itself. Fortunately, there are now enough data for different ions so that we can determine an answer to this question. In table 3 we have compiled the most significant results for each of the most Jahn-Teller active modes of the four ions, C6F6+, C6F3H3+, C6CI3H3+ and c6cI3Fz. Table 3 gives Di, the linear Jahn-Teller distortion parameter, which it may be remembered isT.J. SEARS, T. A . MILLER A N D V. E. BONDYBEY 179 just the ratio of the Jahn-Teller stabilization energy to the unperturbed oscillator frequency, wf. It also gives the stabilization energy, E ~ , for each mode and the total stabilization energy for the whole ion, i.e., the sum of all modes. Finally, for each mode is given a geometric distortion parameter, dp,. This parameter denotes the shift along the mass-weighted normal coordinate from the symmetrical geometries (dpl = 0) for which the minimum of the Jahn- Teller distorted potential occurs. It is simply related to the previous parameters. 6pi(a.m.u.*8,) = 8.192/Di/wi where wi is expressed in cm-'. The first thing to be noted from table 3 is that the total stabilization energy for all the ioas lies between ca.500-1000 cm- or ca. 1.5-3.0 kcal mol- l. While there is a ca. 10-20% (de- pending upon the species) uncertainty in the total stabilization energy for the ions, it appears TABLE 3 .-JAHN-TELLER PARAMETERS FOR BENZENOID CATIONS Units are as follows: Di, dimensionless; ci in cm-I; 6pi in a.m.u.*8,. C-C stretch (15) 0.23 370 0.098 0.35 550 0.122 0.18 277 0.089 0.15 233 0.081 C-C-C bendb (17) 0.68 289 0.328 0.73 350 0.319 0.62 260 0.315 0.60 234 0.321 C-F bendb(18) 0.38 101 0.310 0.030 10 0.078 ... 0.32 99 0.263 C-CI bendb ... ... 0.03 5 0.091 0.05 9 0.135 X.51 82 1 922 563 515 C-F stretch (16)b 0.05 61 0.053 0.012 12 0.029 0.02 21 0.036 ... Numbering appropriate for C6F2. The angular bends are converted into 8, by multiplying by the appropriate bond length.Results for C6F3H3+ and C6CI3H: are taken from recent 4-mode calculations and are therefore slightly different from those previously published.6 The entry in this row for C6C13H3+ refers to mode 11 which for this compound is predominantly a mixture of C-H bend and C-Cl stretching motions. that the C1 substituted cations have a smaller stabilization energy, although all the ions are comparable. To get a more detailed picture of the Jahn-Teller effects we have to examine the individual modes. The 4 symmetry coordinate modes for a Dbh species (C6F$) species can be labelled as C-C stretch (1 S), C-F stretch (1 6), C-C-C bend (1 7), C-F bend (1 8) or C-Cl bend. We have chosen these labels for the modes in table 3.From our normal-mode analysis for C6F6+ and C6H3F3+, we know these labels are only approximate, as symmetry-coordinate mix- ing does occur in the normal coordinates. Nonetheless, they are convenient appellations- with some physical justification-and we shall use them. It can be seen that for all four ions >SO% of the stabilization energy comes from two modes, the C-C stretch and the C-C-C bend, both motions characteristic of the ring it- self. One can also note that with one or two exceptions involving the large amplitude, low- frequency carbon-halogen bends for the hexahalogenated species where strong mode mixing with the C-C-C bend occurs, the geometric distortions are clearly largest for these same modes, Moreover, these distortions are almost identical in all four ions, 0.10 i 0.02 a.m.u.i 8, for the C-C stretch, and 0.32 & 0.02 a.m.u.+ 8, for the C-C-C bend (multiplied by the bond length for common units).On the basis of the above observations, we make the following suggestions. The Jahn- Teller distortion of benzenoid cations is characteristic of the benzene ring, being governed by the ring n electrons. The stabilization energies are in the 1-4 kcal range. The distortion of the molecule at the minimum of the potential will be a combination of changes in C-C bond lengths and angles. Converting from the above mass-weighted normal coordinates, we find that the C-C bond length changes are 0.01-0.02 A, with for example 4 C-C bonds contract- ing by this amount and 2 expanding by twice this amount.The C-C- C angles will change by 1-2" with again 4 decreasing from 120" by this amount and 2 increasing by twice this amount. [For more precise results on C6F,+ and C6F3H3f see ref. (7) and (6).]180 JAHN-TELLER DISTORTION I N C6C13F2 Small, almost second-order, changes in these characteristic distortions and stabilization energies can be brought about by substituent effects. For example in the C1 substituted benzenes, more n electron delocalization is likely to occur onto the C1, making the interaction of the eN electron with the ring less intense and slightly decreasing the Jahn-Teller effect. In a similar way purely mass effects can play a small role. For a given distortion parameter D t , the lower-frequency CI modes will have a smaller absolute stabilization energy. How- ever, overall these effects appear relatively slight, and it appears reasonable to think of a characteristic Jahn-Teller distortion and stabilization of the ring for benzenoid cations. J. Daintith, R. Dinsdale, J. P. Maier, D. A. Sweigart and D. W. Turner, Molecular Spectro- scopy 1971 (Institute of Petroleum, London, 1972), p. 16. J. P. Maier and 0. Marthalar, Chem. Phys., 1978, 32, 419. C. Cossart-Magos, D. Cossart and S. Leach, Chem. Phys., 1979, 41, 345 and 363. \-. E. Bondybey, T. A. Miller and J. H. English, J , Chem. Phys., 1979, 71, 1088. H. A. Jahn and E. Teller, Proc. R. Soc. London, Ser. A, 1967, 161, 220. T. J. Sears, T. A. Miller and V. E. Bondybey, J. Chem. Phys., 1980,72, 6070. T. .I. Sears, T. A. Miller and V. E. Bondybey, J. Chem. Phys., 1981,74, 3240. C. S. Sloan and R. Silbey, J. Chem. Phys., 1972,56, 6031. V. E. Bondybey, J. Chem. Phys., 1979, 71, 3586. lo T. J. Sears, T. A. Miller and V. E. Bondybey, J. Am. Chem. Soc., 1980, 102, 4864. l1 T. J. Sears, T. A. Miller and V. E. Bondybey, unpublished results. l2 H. C. Lunguet-Higgins, Adv. Spectrosc., 1961, 3, 429. l3 E. Haller, L. S. Cederbaum, W. Domcke and H. Koppel, Chem. Phys. Lett., 1980, 72, 427. i4 J. H. S . Green and D. J. Harrison, J. Mol. Spectrosc., 1976, 62, 228 and J. Chem. Thermodyn., 1976, 8, 529.

ISSN:0301-7249    

DOI:10.1039/DC9817100175

出版商:RSC    

年代:1981

数据来源: RSC

摘要:

Spectroscopic Studies of Open-shell Organic Cations in the Gaseous Phase : Chlorodiacetylene and Dichlorodiacetylene Cations BY JOHN P. MAIER, OSKAR MARTHALER, LIUBOMIR MISEV AND FRITZ THOMMEN Physikalisch-Chemisches Institut der Universitat Basel, Klingelbergstrasse 80, CH-4056 Basel, Switzerland Received 1st December, 1980 The application of the techniques of emission, laser-induced excitation and photoelectron-pho- ton-coincidence spectroscopies to open-shell organic cations which decay radiatively are illu_strated by such studies on chlorodiacetylene (A”%* ++ X’IIfi) and dichlorodiacetylene (A”211,q, ++ X’IT,,,) cations. The emission and the laser-induced excitation spectra yield the vibrational frequencies of most of the totally symmetric fundamentals for these cations in their 2 a n d xstates.Higher-resolu- tion emission spectra reveal further structure. The detection of coincidences between cnergy- selected photoelectrons and emitted photons show that the radiative process depletes Ievels at least up to 3000 cm-l within their xstates and the data give their fluorescence quantum yields and cascade- free lifetimes. It is shown how these methods provide complementary information on the spectro- scopic structure and decay behaviour of open-shell cations. The cations of chlorodiacetylene, Clf CEZC>~H+, and of dichlorodiacetylene, C ~ ~ C E C ~ ~ C I + , belong to the hundred or so open-shell organic cations which have been found to decay radiatively in the gaseous phase.’P2 This has been established by recording the emission spectra of such cations and by identifying the transitions by reference to photoelectron spectroscopic data.In all these cases the band systems are the result of electronic transitions from one of the two lowest excited doublet states (2x, or 28) to the ground state (2y) of the cations. The detection of the radiative decay has opened up the possibilities in the application of well-established and of newly developed spectroscopic techniques to the structure and decay behaviour of such organic open-shell cations. High-resolution emission spectroscopy, which historically has played a vital role in elucidating electronic and geometric structures of gaseous species, belongs to the former ~ategory.~ By this means a variety of triatomic cations has been investigated4 as well as the cations of diacetylene and recently of the halogenated benzenes.6 The newer methods include laser-induced fluorescence which was first applied to the molecular cation N2+ ’ and subsequently to some other organic cations’ which we found earlier to decay radia- tively, and coincidence measurements.The latter depend on monitoring emitted photons at the same time as energy-selected electrons’ or mass-selected ions,l0 after formation of the cations by photoionisation. In this article the results obtained for chlorodiacetylene and dichlorodiacetylene cation using the emission, laser-induced excitation and photoelectron-photon coin- cidence spectroscopic techniques are presented. The emission spectra of these two182 OPEN-SHELL ORGANIC CATIONS open-shell cations, which lie in the 500-700 nm wavelength region, correspond to the following electronic transitions: 1 1 , 1 2 and the states characterized by R = 3/2 lie energetically below R = l/2.I3 The study of these transitions by the above-mentioned methods yields the vibrational frequencies of most of the totally symmetric fundamentals for these cations in their 2 and 2 states as well as their lifetimes and fluorescence quantum yields in selected levels of their 2 states.As the apparative details have already been presented for the three techniques in earlier articles, only the salient features of the measurements are sum- marized below in the sections dealing with each approach in turn. EMISSION SPECTRA The emission spectra of chlorodiacetylene and dichlorodiacetylene were recorded with a crossed electron-sample beam apparat~s.'~ The excited cations were produced by a 20-40 eV electron beam impinging on an effusive jet of the sample and the result- ing photons were dispersed by a 1.26 m monochromator, e.g.The spectra were recorded on-line with an LSI 11/03 micro-computer using single- photon counting electronics. The emission spectra covering the whole band systems were recorded with optical resolutions of 0.16 nm. The spectra were presented and vibrationally analysed pre- viously;l'~l2 they are nevertheless reproduced in fig. 1 and 2 for comparison with the laser-induced excitation spectra (uide infra) and with the high-resolution scans of the 0; bands (insets of fig. 1 and 2). For the latter measurements a resolution of 0.008 nm was achieved using sample pressures in the collision region around Torr and electron currents of 1-2 mA.It was concluded on the basis of the emission and photoelectron spectra that the R = 3/2 t-) 3/2 and R = 1/2 t) 1/2 components of the J2n, ---f X2n* band systems of chlorodiacetylene and dichlorodiacetylene cation lie energetically so close together that each emission band contains both As in the photoelectron spectra the spin-orbit splittings (0.020 & 0.005 eV) are resolved only for the bands corresponding to the ionization to the ~ ' l 3 ~ states, the implication is that the spin- orbit splittings are also of this magnitude for the f 2 h states. The high-resolution recordings of the " 0; band " show that it is composed of several peaks (fig.1 and 2). A possible interpretation is that it shows that the R = 3/2 and R = 1/2 transitions are separated by ca. 1 and 2 cm-I in the case of chlorodi- acetylene and dichlorodiacetylene cations, respectively. The remaining vibrational structure is due to sequence bands. More definite assignment should be possible when some of the other bands are also recorded with a corresponding resolution. The vibrational frequencies of five fundamentals of chlorodiacetylene and of theJ . P . MAIER, 0. MARTHALER, L . MISEV AND F . THOMMEN 183 I - 19720 19730 1 .n 1 I 1 I 1 I I 1 I 1 1 I 1 1 1 I I I 1 I I I I I I I 17- 14000 16000 18000 20000 C /cm- FIG. 1,-The zzII, -+X2II* emission system of chlorodiacetylene cation. The whole spectrum was recorded with an optical resolution of 0.16 nm, whereas the region around the 0: band was scanned with 0.008 nm (f.w.h.m.), A vibrational assignment of some of the prominent bands is indicated.three C 1; modes of dichlorodiacetylene could be inferred from the low-resolution spectrum to an accuracy of 5 10 cm-'. The assignments are indicated in fig. 1 and 2 and the frequencies are given in table 1 where the molecular values 15*16 are also given. In view of the laser-induced excitation spectra to be discussed in the next section, several of the sequence transitions can no,w be identified in both sets of spectra. In addition, the frequencies inferred for the A2n* state from the hot bands apparent to higher energy of the 0; bands (fig. 1 and 2) are consistent with the interpretation of the excitation spectra.I' 19090 19100184 OPEN-SHELL ORGANIC CATIONS TABLE 1 .-VIBRATIONAL FREQUENCIES (cm- I) OF THE TOTALLY SYMMETRIC FUNDAMENTALS OF CHLORODIACETYLENE (c ') AND DICHLORODIACETYLENE (c;) CATIONS IN THEIR GROUND AND FIRST EXCITED ELECTRONIC STATES INFERRED FROM THE EMISSION (em.) AND LASER EXCITATION (exc.) SPECTRA (fig. 1-4). Uncertainty of all values i 10 cm-'. Values for v 8 of z symmetry are deduced from 2 v8. The ground molecular state values are taken from ref. ( I 5) and (1 6), respectively. v1:v, v2:v, (C-H) (C=C) CI-C=C-C=C-H X'T, + 3327 2252 CI-C-C-C=C-H+ f211a em. 2190 X2na em. exc. 2150 v1:vs v2:v, cI-c~c-c=c-cI X'C; 2245 1202 C1-- C-C-C-C-Cl+ em. 2210 1315 A213a,l, em. 1170 exc. 1165 (C=C) (C-C) v3:va v4:v, v5:v, v8:(sym. (C-C) (C-C) (C-X) bend) 2071 1093 525 335 1910 1180 540 305 550 305 1080 530 3 10 1085 520 310 v3 : v, 330 3 90 3 90 365 (C-W LASER-INDUCED EXCITATION SPECTRA Laser-induced excitation spectra of chlorodiacetylene and dichlorodiacetylene are shown in fig.3 and 4. These were obtained by producing the cations in their ground states by Penning ionization using helium or argon met as table^'.^ and by pumping the electronic transition with a tunable, pulsed, dye laser while the undispersed fluorescence was sampled, e.g. t HeW c IfC= c+* c I The signals were accumulated by a transient digitizer interfaced to an LSI 11/03 micro-computer which also steers the 1 a ~ e r . I ~ The spectra shown have been recorded with a laser bandwidth of 0.02 nni and have been corrected for the laser intensity.The spectra were actually put together from three scans using the appropriate dye so- lutions. The wavelengths were calibrated using atomic emission lines produced by excitation of the helium and argon metastables. Vibrational assignments of some of the prominent bands in the xzlln t f213n laser excitation spectra are indicated in fig. 3 and 4. Also apparent are a few bands due to laser excitation of the C2 Swan system. This fragment is produced by the Penning excitation processes. In table 1 are collected the vibrational frequencies inferred from the spectra, which correspond mainly to the totally symmetric, C + , or C,+ , funda- mentals in the J2& state of these cations. The few hot bands evident to lower energy of the 0; bands yield vibrational frequencies for the f211a state, in agreement with the emission spectra. Some of the sequence transitions are also apparent, especially to lower energy of the bands assigned to the progressions and combinations of the fundamentals (fig.3 and 4).J . P. MAIER, 0. MARTHALER, L. MISEV AND F. THOMMEN 185 c2 i 19000 20000 21000 22000 V /cm-' FIG. 3.-Laser excitation spectrum of the X'IIQ t ~ T I Q transition of chlorodiacetylene cation re- corded with 0.02 nm bandwidth. A vibrational assignment of some of the bands is indicated. The maxima of the 0: bands are found at 19 721 & 3 cm-' and 19 092 3 cm-l in the laser excitation spectra of theA2& t ~'IIQ transitions of chlorodiacetylene and dichlorodiacetylene, respectively. This is in good agreement with the recordings of the 0; emission band with the higher resolution where the further structure is resolved (fig.1 and 2). Finally, it is seen that the excitation spectra cover an energy range of 19000 20000 i~ /cm-l FIG. 4.-Laser excitation spectrum of the Al'II,,, -+ X'IIn,, transition of dichlorodiacetylene cation recorded with 0.02 nni bandwidth. A vibrational assignment of some of the bands is indicated. only about a third of that of the emission spectra. could be detected. photoelectron-photon-coincidence measurements are considered. Outside the shown range, no bands The reason for this becomes apparent when the results of the PHOTOELECTRON- P H 0 TON- C 0 1 N C I D EN C E S P E C TROS COPY The photoelectron-photon-coincidence apparatus consists of two parts; one for the detection of energy-selected electrons and one for any emitted photons, following186 OPEN-SHELL ORGANIC CATIONS photoioni~ation.~+’~ The individual events in each channel are then sampled in delayed coincidence.True coincidences are obtained only if the ejected electron and emitted photon originate from the same ion which was produced in the state defined by the kinetic energy of the photoelectron, t?K.E., This technique has been applied to chlorodiacetylene and dichlorodiacetylene, to show that vibrationally excited cations in the .,@IIQ states decay radiatively, and from the quantitative measurement to determine the fluorescence quantum yields, qF(u’), and lifetimes, ~ ( u ’ ) , of the selected levels u’. True coincidences were detected for chlorodiacetylene and dichlorodiacetylene cations at the internal energy within their X2n, states as indicated above the photo- electron bands in fig.5 and 6. These energy locations correspond to slices of 100 meV centred on the vibrational peak maxima, which are due to the progressions of the totally symmetric C-C stretching fundamentals, v4 and v 2 (see table 1) for chloro- and COINCIDENCES ; 0 I H 9.0 10.0 11 0 12.0 13.0 14.0 15.0 IE /eV FIG. 5.-Photoelectron-photon-coincidence curve for the 0’ level of the A”*IT, state of chlorodiacety- lene cation: N,, 290 Hz; N r , 0.68 Hz; accumulation time, 17 h. The He(1a) photoelectron spec- trum, recorded under the coincidence conditions, shows the internal energies selected, dichloro-diacetylene, respectively. The results show that the radiative channel also depletes the 4“ and 2“ IZ = 1-3 vibrational levels of their A2l7* states.Thus in the emission and excitation spectra population of vibrational levels up to ca. 3000 cm-l within the z211n states should be evident. In the emission spectra only a few weak bands are apparent on the high-energy side of the 0; bands (fig. 1, 2) and therefore the intensity of the radiative transitions especi- ally from the more highly excited vibrational levels is concentrated in their sequence transitions. This is as suggested by the resolved structure in the higher-resolution recording of the 0; emission bands and the assignment of some of the stronger bands to sequence transitions in the emission and excitation spectra (cfi fig. 1-4).On the otherJ . P . MAIER, 0. MARTHALER, L . MISEV AND F . THOMMEN 187 0 COINCIDENCES 600 TIME 9.0 10.0 11.0 12.0 130 14.0 15.0 IE lev FIG. 6.-Photoelectron-photon-coincidence curve for the 0’ level of the A”zIIn,u state of dichlorodi- acetylene cation; N,, 280 Hz; NT, 0.39 Hz; accumulation time, 20 h. The He(1cr) photoelectron spectrum, recorded under the coincidence conditions, shows the internal energies studied. hand in the excitation spectra bands lying more than ca. 2000 cm-’ above the 0: bands are too weak to be detected. The reason for this becomes apparent when the fluorescence quantum yields and lifetimes are inspected. The ~ ~ ( 0 ’ ) values were obtained by measurements of the rates of detection of true electrons, N,, and of true coincidences, NT, because it can be shown that NTINe = f n V V F ( v ’ ) and the collection efficiency for photons,fn,, has been absolutely calibrated in the 200- 900 nm wavelength range.” In order to attain the 10% accuracy for the pF(v’) values ca.lo7 true electrons have to be counted. Typical accumulation times and count rates under the coincidence measurements are given in the legends to fig. 5 and 6. In table 2 are presented the determined ~ ~ ( 0 ’ ) values as well as the cascade-free lifetimes which were extracted by a weighted least-squares linear fit to a semi-logarith- mic plot of the decay part of the coincidence curves (cf. fig. 5 and 6). For comparison, TABLE 2.-FLUORESCENCE QUANTUM YIELDS, VF(U’). AND LIFETIMES Z(V’) OF CHLORODIACETY- LENE AND DICHLORODIACETYLENE CATIONS IN THEIR A STATES CORRESPONDING TO THE INDICATED POSITIONS IN FIG.5 AND 6 The lifetimes given in the last column were obtained using electron impact excitation. 11,12 cation state V&’) z/ns r/ns Cl-C=C-C=C-H+ L2n, 0’ @) 0.79 f 0.08 41 f 2 41 j = 2 5l 36 i 2 s 37 k 2 4’ @ 0.37 -f 0.04 33 &- 3 42 @ 0.18 f 0.02 30 1 3 43 @ 0.06 f 0.01 2l @ 0.14 f 0.02 1 7 f 3 19&3 22 @ 0.035 i 0.007 Cl-C=C-C=C-CI+ A”zII*,, 0’ @ 0.47 f 0.05 2 1 1 2 2 1 f 3188 OPEN-SHELL ORGANIC CATIONS the lifetimes measured by means of a pulsed electron beam (ca. 20 eV) excitation are included.11*12 There is good agreement between the two sets of measurements. From the data given in table 2, the radiative and non-radiative rate constants as func- tion of internal energy can be directly obtained.It is seen that the qF(u’) values fall off with increasing internal energy within the A211a states of chloro- and dichloro-diacetylene cations (table 2). In the laser- excitation spectra the intensity of the bands is proportional to the yF(v‘) value of the emitting level U‘ and to the probability of populating that level initially. Relative to the zeroth vibrational level this may be judged by the Franck-Condon profile of the relevant photoelectron band. Thus, for example in the case of chlorodiacetylene cation the intensity of the 4; band in the excitation spectrum is expected to be about an order of magnitude less than that of the 0: band. This is approximately the ob- served ratio (fig. 3). On the other hand, the q+(u’) data for dichlorodiacetylene cation indicate that the 2; band should be even weaker relative to the 0: band and further- more the absolute values are also expected to be a factor of two weaker than for chlorodiacetylene cation (table 2).In fact the 2; band could not be discerned above the noise in the excitation spectrum of dichlorodiacetylene. CONCLUDING REMARKS It has been shown how the application of the emission, laser-induced excitation and photoelectron-photon-coincidence spectroscopic techniques provide detailed, and complementary, information on the structure and decay behaviour of open-shell organic cations in the gas phase. These methods can be applied to all those cations whose radiative decay is manifested1*’ and in this article such studies on chlorodiacety- lene and dichlorodiacetylene cations in their X’n, and Z’IIa states have been described.The emission spectra yield the vibrational frequencies, of mainly the totally sym- metric fundamentals for the ground cationic states, and with the improvements in resolution, at present down to 0.004 nm, further detail and in some of the smaller cations rotational structure becomes apparent.2 In addition the lifetimes of the cations in the lowest vibrational levels of the excited state can be measured. The laser-induced excitation spectra provide the corresponding vibrational frequency data for the excited electronic state and the presently employed resolution of 0.02 rim can be increased by a factor of ten. The photoelectron-photon-coincidence measure- ments can be used first of all to prove that selected levels of an excited state decay radiatively and the wavelength range of the emitted photons can also be established.18 This can provide valuable information in the rationalization and in the investigations of the cations using the emission and laser techniques.Furthermore, the absolute fluorescence quantum yields and cascade-free lifetimes can be determined. These data enable one in turn to discuss the radiationless decay of excited open-shell cations as function of the internal energy/vibrational excitation because the radiative and non-radiative rate constants can be derived. This work has been supported by fhe Schweizerischer Nationalfonds zur Forderung der wissenschaftlichen Forschung (Project No. 2.212.0-79, E 35). Ciba-Geigy SA, Sandoz SA and F.Hoffmann-La Roche & Cie. SA, Base1 are thanked for financial support.J . P . MAIER, 0. MARTHALER, L . MISEV A N D F. THOMMEN 189 ‘See J. P. Maier in Kinetics of Ion-Molecule Reactions, ed. P. Ausloos (Plenum Press, New York, 1979); J. P. Maier, Chimia, 1980, 34, 219 for reviews of this field. J. P. Maier, 0. Marthaler, L. Misev and F. Thommen, in MoIecuIar Ions, ed. J. Berkowitz (Plenum Press, New York, 1981). ’ See G. Herzberg, Molecular Spectra and Molecular Structure (van Nostrand, New York), vol I, 1950 and vol. 111, 1966, and references therein. G. Herzberg, Quart. Rev. Chem. SOC., 1971,25,201; S. Leach in The Spectroscopy of the Excited State (Plenum Press, New York, 1976). J. H. Callomon, Can. J. Phys., 1956, 34, 1046. C. Cossart-Magos, D. Cossart and S. Leach, Mol. Phys., 1979, 37, 793; C. Cossart-Magos, D. Cossart and S. Leach, Chem. Phys., 1979, 41, 345, 363. P. C. Engelking and A. L. Smith, Chenz. Phys. Lett., 1975, 36, 22. T. A. Miller and V. E. Bondybey, J. Chin?. PhyJ,, 1980,77, 695; T. A. Miller, Faraday Discuss. Chem. SOC., 1981, 71, 175 and references therein. M. Bloch and D. W. Turner, Chem. Phys. Lett., 1975,30, 344. lo J. H. D. Eland, M. Devoret and S. Leach, Chem. Phys. Lett., 1976, 43, 97. l1 J. P. Maier, 0. Marthaler and E. Kloster-Jensen, J. Electron Spectrosc., 1980, 18, 251. l2 M. Allan, E. Kloster-Jensen, J. P. Maier and 0. Marthaler, J. Electron Spectrosc., 1978, 14, ’’ E. Heilbronner, V. Hornung, J. P. Maier and E. Kloster-Jensen, J. Am. Chem. SOC., 1974, 96, l4 M. Allan, E. Kloster-Jensen and J. P. Maier, 1. Chem. Soc., Faraday Trans. 2, 1977, 73, 1406; l5 D. H. Christensen, I. Johnsen, P. Klaboe and E. Kloster-Jensen, Spectrochim. Acta, 1964, l6 P. Klaboe, E. Kloster-Jensen, E. Bjarnov, D. H. Christensen and 0. F. Nielsen, Spectrochim. l7 J. P. Maier and L. Misev, Chem. Phys., 1980, 51, 311. 359. 4252. 0. Marthaler, Ph.D. Thesis (University of Basel, 1980). A25, 1569; M. K. Phibbs, Spectrochim. Acta, 1973, 29, 599. Acta, 1975, A31, 931. J. P. Maier and F. Thommen, Chem. Phys., 1980, 51, 319.

ISSN:0301-7249    

DOI:10.1039/DC9817100181

出版商:RSC    

年代:1981

数据来源: RSC

摘要:

Laser-induced Fluorescence of Trapped Molecular Ions : The CH+ATI f- X'c+ System BY FRED J. GRIEMAN,* BRUCE H. MAHAN, ANTHONY O'KEEFE AND JOHN S. WINN Department of Chemistry, University of California and Materials and Molecular Research Division, Lawrence Berkeley Laboratory, Berkeley, California 94720, U.S.A. Receiced 17th November, 1980 The CH+ and CD+ A1n+ X I Z + absorption spectra have been obtained by laser excitation of these fragment ions. The ions are contained in a mass-selective quadrupole ion trap under collision- free conditions. The spectra therefore reflect the nascent internal-energy distributions of the ions, which were produced by electron impact on CH, (CD4) or C2H2 (C2D2). Both parent gases gave vir- tually identical spectra; large rotational and vibrational excitation was observed.The equipment yas also capable of measuring the radiative lifetime of CH+ (CD+) A2n (u 7 0), and the measured value, 815 ns, is found to be in good agreement with theoretical calculations of this quantity. The CH+ radical has been the subject of numerous experimental and theoretical investigations since its spectroscopic identification by Douglas and Herzberg.' This small, reactive radical is of great importance in combustion reactions and atmospheric chemistry and is believed to play a fundamental role in the creation of many small molecules within the interstellar clouds. The first spectroscopic observation of this ion was made in spectra of the interstellar medium,' the observed transitions belonging to the CH+ A l I I t XIC+ system. Klemperer and Solomon3 have made a detailed analysis of the interstellar processes involving CH+, believed to have a significant effect upon the molecular composition of interstellar clouds.This ion is believed to be important in the formation of CH, CO, CN and several other molecular species. The chemistry involved in these pro- cesses is intimately related to the relative concentrations of the species involved. These concentrations must be inferred from an analysis of observed line strengths of stellar spectra and known or calculated oscillator strengths. In the case of CH+, the uncertainty in the radiative lifetime for the A + X transition manifests itself in a large uncertainty in stellar abundance and, in turn, to confusion over the relative importance of various chemical reactions occurring in the interstellar medium.This uncertainty arises not from the lack of experimental and theoretical study, but from the failure of such study to reach a consistent result. Until recently4 the only experimental method useful in the high-resolution study of molecular fragment ions has been emission spectroscopy. While this technique is a powerful one, radiative transition rates obtained in this manner are subject to errors which are often difficult to identify or estimate. The problems arise from the compli- cated nature of the excitation process. One often creates many highly excited states * Present address: Department of Physics, University of Oregon, Eugene, Oregon 97403, U.S.A.192 L I F OF TRAPPED MOLECULAR IONS which may cascade down to the level of interest leading to a distortion in the mea- sured decay rate.The A 2 n -+ X'C+ radiative decay rate has been the subject of five experimental studies '-' each resulting in apparently single exponential decay curves corresponding to radiative lifetimes ranging from 70 to 630 ns for the (0,O) band. All of the experimental values also appeared to contradict the theoretical estimate of Yoshimine et aZ.,'' zo E 800 ns. We have developed a technique with which we obtain the laser-induced fluore- scence spectra of ions confined to a small (1 cm3) spatial region within a three-dimen- sional radiofrequency quadrupole trap. This trap is similar to the arrangement des- cribed by Dawson,l' and it allows us to store large numbers of ions for time periods which are limited only by collisions with background neutral gas molecules.Using the Langevin estimate for the ion-neutral collision rate, we find that for a neutral gas background pressure of Torr, each ion will experience one collision per ms. This is expected to be an upper bound for the actual collision rate and, in practice, we find little difficulty in storing reactive species, such as CH+, for periods of 5-10 ms in this pressure range. Because the trap can store ions for such long periods this ar- rangement provides an attractive possibility for the study of radiative decay rates. Ions which are in the ground (or an optically metastable) electronic state can be ex- cited with a brief laser pulse and the resulting fluorescence monitored as a function of time.The observed decay rate provides an unambiguous measurement of the upper level's radiative lifetime. The experimental system described here is capable of storing ions in a mass selec- tive mode, thus removing any doubt, in most cases, as to the identity of the ion under investigation. The degree of mass differentiation is variable, and in the high-resolul tion mode of operation ions differing by only 1 a.m.u. can be resolved. Such resolv- ing power is important in any study involving hydrogen-containing ions. We present here a detailed description of our experimental apparatus and results on the CH+ and CDf AlII - X'C+ system. EXPERIMENTAL The ion trap used in this study was one of cylindrical geometry which has been described in detail elsewhere.12 The trap consists of a cylindrical centre electrode and two flat end electrodes, positioned at opposite ends of the cylinder.This configuration was selected instead of the more general hyperbolic geometry in an effort to minimize electric-field distor- tion introduced by the presence of laser-beam entrance and exit holes in the centre electrode. The general configuration is illustrated in fig, 1. The details of trap operation will be briefly outlined here. Further details may be obtained in the references already cited. The principles of operation are an extension to three dimensions of those involved in a conventional quadrupole ion lens. Application of an r.f. voltage to the centre electrode creates a pseudo-potential in which ions are trapped regardless of their charge-to-mass ratio.By floating the applied r.f. voltage at some d.c. bias level a selectivity in the charge-to-mass ratios which are confined in the potential well is introduced. The degree of selectivity is determined by the relative magnitudes of the d.c. and r.f. voltages. By varying the r.f. ampli- tude, at a fixed d.c,-to-r.f. ratio, the mass selection window may be shifted to different values, with ions of larger masses being trapped for greater applied voltages. Given the general design of the trap, the depth of the pseudopotential well can be expres- sed as l3 Here D is calculated in V, V is the maximum a.c. voltage between the electrodes, Zo is half the minimum separation of the cap electrodes, M is the ion mass in a.m.u., and f is theF .J . G R I E M A N , 13. H . MAHAN, A . O'KEEFE AND J . s. WINN 193 applied field frequency in MHz. In the present study the values for these parameters are Zo = 1 , V li 200, M = 13 and f = 1, and the corresponding well depth is 19 V. The well depth determines the number of ions which can be trapped, since the trap reaches its maxi- mum capacity when the space-charge potential cancels the trapping potential. At this point the maximum concentration of ions is given byI3 ions 1.66 x 1060 zo2 I?",,, (s) = Thus for a well depth of 19 V we find that 3 x lo7 ions can be trapped. When the instru- ment is operated as a mass spectrometer the total capacity of the trap is smaller by a factor which is difficult to calculate; however, the variation of laser-induced fluorescence intensity Electrode Cone A FIG.1.-Schematic representation depicting a vertical slice of the quadrupole ion trap used in the present study. A radiofrequency voltage is applied to the centre ring electrode while the top and bottom electrodes are maintained at ground potential. Ions, spatially confined to the enclosed region, possess a nearly Gaussian density distribution peaking at the centre of the trap. with increasing mass selectivity for ions such as N2+ and CO+ indicates an ion density drop by a factor between 2 and 5. In the present study, the fragmentation pattern of CH4 neces- sitated the use of bias conditions capable of discriminating between ions differing by only 1 a.m.u. Ions are created within the trap by electron impact ionization of a selected background neutral gas which is introduced into the vacuum chamber through a variable leak valve and which is maintained at a pressure of from to Torr.Fragment CH+ was generated by the electron impact dissociation of CH4 (cu. 2-3% of the total ion products) and CD+ was produced analogously from CD4. The trap has been found to attain maximum ion density after a 2 ms electron pulse from an electron gun producing an average e- beam current of 10 PA. The electron gun consists of a resistively heated 1 cm length 0.25 mni diameter thoriated tungsten wire mounted on a ceramic base and enclosed within a metal shield which is floated at the centre electrode potential. This shielding serves to reduce scattered light emitted from the hot filament. A series of lenses focuses the electrons through a hole in the centre elec- trode and into the trap.One of the focusing lenses also serves as an electron shutter through the application of a high-voltage pulse to the lens element. While the trapping potential operates continuously, the remainder of the experiment is operated in a pulsed mode at a repetition rate of 10-40 Hz. The remainder of the experiment194 LIF OF T R A P P E D MOLECULAR IONS is then most easily described by considering one experimental cycle which consists primarily of three parts: ion creation and confinement, excitation of the ions, and fluorescence signal detection. The experiment begins with the initiation of the electron beam which creates ions for a period of several milliseconds. The electron gun is then gated off and a delay period of several hundred p s ensues.During this time any necessary mass selection of the ions is allowed to stabilize and any excited electronic states which have been created are permitted to relax radiatively. Radiative relaxation of excited vibrational levels within stable electronic states is expected to be slow on the time-scale of an experimental cycle. The pressure range in which we operate allows little or no collisional relaxation. After the delay period, a 10 ns, 1 cm- bandwidth laser pulse from a Molectron DL-200 dye laser pumped by a Molectron UV-1000 nitrogen laser is passed through the ion cloud. Laser-induced fluorescence is then monitored at right angles to the laser beam. A lens and mirror system directs some of the fluorescence through the wire mesh end electrodes to a cooled RCA 8575 photomultiplier tube.In addition to fluorescence, scattered laser light may also reach the detection system. In order to minimize this background radiation, the laser beam is collimated to ca. 0.5 cm dia- meter with two lenses and directed through 0.5 m arms (fig. 2) containing light baffles on the P D P - E f COMPUTER SYSTEM I GATED COUNTER U LENS ION TRAP BAFFLE I I ,g]+ I \ ELECTRON --. A GUN 8 >LECTRON M U L T I P L I E R POWER SUPPLY TO PUMPS I 1 FIG. 2.-Experimental arrangement used in the frequency-scanning experiments. The depicted arrangement and timing sequence are discussed in the text. entry and exit side of the trap. l4 In addition to the bame system the 10 ns laser pulse allows the use of gated detection techniques to reduce the effects of scattered laser light. The initiation of the detection gate can be variably delayed with respect to the laser pulse, and its duration can be varied between 0.1 and 10 p s .The duration of the gate is primarily determined by the radiative lifetime of the species being investigated. The fluorescence detection gate and the signal from the photomultiplier tube are sent to a gated single-photon counting system which consists of an L.R.S. 621-BL discriminator and a 100 MHz counter (Ortec model 770). After the fluorescence detection, the experimental cycle is completed by pulsing the ions out of the trap to an electron multiplier. In order to get a consistent ion signal from cycle to cycle a high-voltage pulse must be synchronized with the r.f.trapping potential and applied to the bottom trap electrode. The resulting ion signal is measured by an L.R.S. 227-sg inte- grator and is used to normalize the fluorescence signal. The fluorescence signal is also nor- malized with respect to laser power which is measured by a photodiode and the gated inte- grator. The final signal gathered by the integrator is used to calibrate the laser wavelength. A signal from a trigger photodiode initiates a fluorescence detection gate.F. J. GRIEMAN, B. H. MAHAN, A . O’KEEFE AND J . S . WINN 195 Calibration is accomplished with the use of the optogalvanic effect ’’ in which the laser beam is directed into a hollow cathode discharge lamp containing neon. The fluorescence excita- tion spectrum is then calibrated with respect to Ne metastable transitions which are known throughout the visible region.The detec- tion gates initiated by this timing circuitry are generated by Tektronix P.G. 501 pulse genera- tors. The timing and gating logic are not shown in fig. 2 in the interest of clarity. An on- line PDP-8f computer is responsible for the overall control of the experiment. At the end of an experimental cycle, the computer gathers the signals from the integrator and initiates a new cycle. After a predetermined number of cycles the computer retrieves the signal from the counter and advances the laser wavelength by a preset increment. Typically the signal is averaged over several hundred laser pulses before advancing the wavelength.The com- puter normalizes the data, stores them on a disc and produces a hard-copy graph of the spectrum. The CH + (CD +) spectra were produced under the following experimental conditions. The ions were created from CH, (CD,) at a background pressure of Torr with an ioniza- tion period of 2.0 ms. The fluorescence detection gate was 1 ps in duration and was initiated 200 ns after the laser pulse. The fluorescence signal was averaged over 500 laser pulses at each wavelength and the wavelength was advanced by 0.1 A increments. In the determination of radiative lifetimes, the experimental arrangement is modified somewhat. Control of the experimental timing is shifted from the computer to an internally controlled pulse cycle. The timing sequence is unchanged when operating in this mode, but the laser wavelength remains fixed and the signal is collected continuously.The output signal from the photomultiplier is fed into a discriminator and then into a Tracor Northern NS 575 digital signal averager with a Biomation time base. This system provides a 10 ns channel width which is suitable for the radiative decay rates encountered in the present studies. The Biomation time base is triggered by the pulse which initiates the detection gate in the normal mode of operation. The total width of the signal averager’s time base in these studies was 1024 channels, i.e. 10.24 ,us. For the collection of frequency scanned spectra, the Molectron UV-1000 nitrogen laser is used as a pump source for the dye laser because such lasers possess relatively high duty cycles.The fluctuations in laser power are corrected for on a shot-to-shot basis as discussed The timing of an experimental cycle is controlled by a series of logic circuits. Nd -YAG LASER FI LASER D l SCRl Mi NATOR v I SIGNAL AVERAGER 0- ___c VACUUM CHAMBER FIG. 3.-Experimental arrangement used to determine radiative lifetimes. is fed to a PDP-8 computer for data analysis. The resulting signal trace earlier. For measurements of fluorescence decay rates, however, it is desirable to maintain a steady laser power level throughout each determination. The frequency-tripled output of a Quanta Ray Nd:YAG laser was found to vary in output power by only a few per cent and was used as the dye laser pump for all decay rate measurements.196 LIF OF TRAPPED MOLECULAR IONS For radiative decay measurements the dye laser is tuned into resonance with a strong transition in the vibronic band under study.The resulting fluorescence signal is then accu- mulated as a function of time in the signal averager for a period of several hundred thousand experimental cycles. The dye laser is then detuned from resonance, typically by several A, and a background signal is subtracted for an equivalent time period. Data from the signal averager are then transferred to a PDP-8 computer for analysis. The experimental arrange- ment is schematically depicted in fig. 3. RESULTS AND DISCUSSION Laser-induced fluorescence spectra were collected for the 0 t 0 and 2 t 1 bands of the A'rI - XIC+ system for both CH+ and CD+. Representative portions of each band are reproduced in fig.4 and 5. Observed line frequencies are given in tables 1-3. Because no new lines belonging to the 2 t 1 band of CH+ were observed, we I l l l l l l l l l l l l [ l l l l l I I I I I 1 1 1 (0-0) Band R Branch I I I 1 I I I-) "i 23419 23697 wavenumber /cm- FIG. 4.-Representative portion of the spectrum obtained for the (0,O) band of the CH+ A'II+- X'Cf system. The Q9 line belonging to the (2, 1) band of the same system is identified with an arrow in the lower portion of the figure. While other members of this band should be apparent in this wavelength region, overlap with members of the (0, 0) band precludes their identification. have omitted this band from the tables. Assignments for lines belonging to the CH+ band system were based upon the rotational constants of Douglas16 which accurately reproduced low-J components of each band but increasingly overestimated transition It is clear that for a diatomic hydride such as CHf, high-J levels of a band system may not fit a Dunham expansion formula unless higher-order correction terms are included (i.e.Hu terms). Using the molecular constants of Douglas we calculated Hu constants for each band system but could not obtain ade- quate agreement between calculated and observed frequencies. Therefore, an un- weighted least-squares fit of all observed line frequencies was used to produce a new set of molecular constants given in table 4. The A doubling constant, qo, was found to be identical to that deduced by Douglas and Morton: qo = 0.038 ern-'.The rotational constants for CD+ obtained in emission studies by AntiC-Jovanovid et aZ.I7 were found to be adequate in reproducing low-J components of each band for . frequencies at higher J.F . J . GRIEMAN, B . H . MAHAN, A . O'KEEFE A N D J . S . W I N N 197 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 ~ 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 ~ (0-0) Band I 1 fl Q Branch I I I I I I I P Branch R Branch 117 116 115 ' Q Branch 114 1 I3 112 I P Branch 112 I I I X 19 (2-1) Band 23456 23824 wavenumber /cm- FIG. 5.-Representative portions of the spectra obtained for the (0, 0) and (2, 1) bands of the CD+ A'rI t X'X+ system. The unobserved Plo line (denoted by x ) is buried in the (0, 0) bandhead. TABLE BAN OBSERVED LINE FREQUENCIES (IN cm-') FOR THE CH+ A'n + X'C+ (0,O) BAND ~~ ~~ J R branch Q branch P branch 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 23 619.85" 23 637.71 23 650.45 23 658.14 23 660.52 23 657.55 23 649.10 23 635.01 23 615.1 23 589.2 23 557.1 23 518.6 23 473.5 23 422.0 23 362.8 23 296.2 23 222.2 23 139.9 23 049.1 22 950.6 22 842.0 22 724.5 - 23 591.89 23 581.77 23 566.56 23 546.23 23 520.70 23 489.92 23 453.76 23 412.16 23 365.0 23 31 1.9 23 253.3 23 188.1 23 117.4 23 039.5 22 955.0 22 863.0 22 763.8 22 658.1 22 544.2 - 23 536.26 23 498.57 23 455.87 23 408.37 23 355.9 23 298.4 23 235.8 23 168.2 23 095.3 23 017.0 22 933.2 22 843.9 22 748.4 22 647.5 22 540.0 " Entries quoted to two decimal places are the more accurate values from Douglas and Herzberg.' The two data sets were used as listed here in the least- Those with one are from this measurement.squares reduction.198 LIF OF TRAPPED MOLECULAR IONS TABLE 2.4BSERVED LINE FREQUENCIES (IN Cm-l) FOR THE CD+ A1n t x'c+ (2, 1) BAND J R branch 0 branch P branch 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 24 105.67" 24 112.90 24 116.01 24 115.80 24 111.49 24 103.30 24 091.05 24 075.02 24 054.90 24 031.0 24 003.6 23 972.1 23 937.5 23 897.1 23 853.0 23 802.2 23 753.0 23 687.9 24 09 1-05 24 083.28 24 071.90 24 056.70 24 037.65 24 014.69 23 987.85 23 957.16 23 922.42 23 883.75 23 840.82 23 793.8 23 741.0 23 681.0 b - 24 039.15 24 013.15 23 983.24 23 949.84 23 912.44 23 871.31 23 825.4 23 724.4 23 667.0 b - " Entries quoted to two decimal places are the more accurate values from AntiC-JovanoviC e l al." Those with one are from this measurement. The two data sets were used as listed here in the least- squares reduction. Missing entries were weak or overlapped and not resolved.TABLE 3.-oBSERVED LINE FREQUENCIES (IN Cm-') FOR THE CD+ A'n +- X ' c + (0, 0 ) BAND J R branch Q branch P branch 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 23 760.03" 23 769.83 23 776.93 23 781.35 23 783.00 23 782.02 23 777.90 23 771.10 23 761.53 23 748.90 23 733.31 23 714.65 23 692.1 23 667.5 23 639.4 23 608.2 23 573.5 23 493.2 23 448.0 23 397.9 23 345.3 23 286.9 23 219.3 b - 23 744.90 23 739.48 23 731.40 23 720.62 23 707.10 23 690.86 23 671.86 23 650.05 23 625.40 23 597.88 23 567.38 23 497.8 23 458.4 23 416.2 23 370.2 23 320.9 23 269.1 23 212.2 b - 23 714.63 23 694.20 23 671.10 23 645.35 23 616.94 23 585.96 23 552.25 23 516.0 23 477.0 23 435.8 23 390.3 23 343.6 23 292.8 23 239.4 " See footnote a to table 2.See footnote b to table 2.F . J . GRIEMAN, B. H . MAHAN, A . O'KEEFE AND J . s. WINN 199 this isotopic species but, again, predicted line frequencies too high for lines of in- creasingJ [e.g. 5.8 cm-' too high for the (0,O) RI5 line]. An unweighted least-squares fit of all observed transition frequencies produced the set of molecular constants given in table 5. The great difficulty encountered in attempting to fit high-l components of each band with molecular constants derived using lower-l terms is most likely due to the TABLE 4.-MOLECULAR CONSTANTS (IN CM-') FOR THE CH+ (0, 0) BAND voo 23 596.81 (01)" BoRP 11.453 2 (37) BoQ 11.416 9 (41) DoRP 0.002 050 (9) D oQ Bo" 13.930 3 (40) D *" 0.002 049 (1 1) 0.001 373 (1 1 ) a Numbers in parentheses represent a one standard deviation uncertainty in the last digits of each constant.unusual nature of the A'H potential curve for both CH+ and CD+. The nature of this potential is best understood if one examines the corresponding potential in the iso-electronic species, BH. This molecule has been the subject of careful emission studies" and has been determined to possess a barrier in the rotationless potential curve of the AIII state.'. Early calculationsz0 indicated that the A'II state arising from B(2P) + H(2S) is initially repulsive in nature, but a strong interaction between this curve and an attractive curve arising from B('D) + H('S) overcomes the repulsion and gives rise to the observed bound state.While it might be expected that the A'll state of CH+ would exhibit similar be- TABLE s.-MOLECULAR CONSTANTS FOR THE (0, 0) AND (2, 1) BANDS OF CD+. VALUES IN C ~ - ' A N D A. V BRP BQ DRP DQ HRP HQ Bd 4 R: 4 B D" H" Be" a e Re" N (0, 0) 23 747.71 (1) 6.285 (20)" 6.280 (21) 0.001 08 (14) 0.001 16 (16) 1.12 (29) x 1.41 (37) x 0.016 7.627 (20) 0.001 06 (15) 1.55 (35) x (27 1) 5.492 (44) 5.479 (49) 0.000 88 (48) 0.00071 (62) 2.5 (1.5) x 24 095.08 (1) 1.2 (2.2) x 6.481 (30) 0.39 (3) 1.228 (5) 0.014 7.416 (46) 0.001 25 (56) 4.5 (2.0) x 7.733 (40) 0.21 (4) 1.125 (4) See footnote a, table 4.200 LIF OF TRAPPED MOLECULAR IONS haviour, the appearance of a potential maximum is by no means assured. The presence of the attractive charge-induced dipole force means that at large R values, A'II may well exhibit attractive behaviour.If the interaction with the attractive curve arising from the C+('D) t H('S) asymptote becomes strong before the repulsive nature of the A'II curve overcomes the chargelinduced dipole attraction, the result is a bound potential curve with unusual R dependence but no potential maximum. The calculations of Green et a1." suggest that, in fact, there is no actual barrier in the rota- tionless curve. As is apparent upon examination of the spectra reproduced here, we have observed no laser-induced fluorescence which is not attributable to the particular ion under study. The present study covered the wavelength region from 4130 to 4400 A. A thorough search throughout this wavelength region was conducted for any possible interfering species such as background neutral fluorescence or transitions originating in metastable levels of CH+ populated in the ion-formation process.No such inter- ference was found. The recent photodissociation studies of Cosby et aZ.22y23 suggest that the 311 state of CH+ (T, 2: 9200 cm-') is, in fact, populated to some extent in the electron-impact ionization of CH,, although the relative amount of population in the singlet and triplet manifolds is difficult to estimate from the available data. If the analysis of Carre24 is correct then transitions belonging to the 3C t 31J system might be expected to appear strongly only below 3700 A, thus leaving the spectral regions under study here free from interference.Examina- tion of the spectra reveals a linewidth noticeably larger than the bandwidth of the dye laser. The source of this broadening is the large Doppler profile of ions stored in a trap such as this. Recently25 a study of the spacial distribution of Li+ in a quadrupole trap similar to our own determined a linewidth consistent with an ion thermal tem- perature of 5000 K. Thus, for an ion of a similar charge-to-mass ratio and for similar operating conditions, it is reasonable to expect such a line broadening. We estimate an ion translational temperature of ca. 4000 K for ions stored under the conditions described here. A result of this large translational energy is that ions produced with a given rota- tional population distribution may be thermalized to a temperature much higher than room temperature. This may be done by introducing some inert gas to the vacuum chamber in addition to the parent gas, but at a higher partial pressure, and allowing the ions to experience many high-energy collisions.Using this approach it should be pos- sible to populate collisionally high J levels in selected ions and to observe, by suitable means, rotational predissociation, thus providing accurate estimates of dissociation energies. Such information is generally lacking for molecular ions. Another interesting feature which is evident in both spectra is the high rotational temperature of the ions even in the absence of such collisional effects. This stands in contrast to the apparently much cooler rotational distribution observed in emission studies.One may infer from published emission photographic plates and tabulations of observed lines a rotational temperature of 350-400 K for ions created in the AlII state, whereas, in the present study, rotational distributions corresponding to temperatures of 3000 K are seen for the X'C+ state. This broad rotational distribu- tion is not the result of high-energy ion-neutral collisions involving CH+(CD+) of the type mentioned above as the experimental conditions under which these spectra were obtained provide a nearly collision-free environment for the ions. Thus, this distribution reflects the initial product state distribution of CH'(CD+) X'C+ created in the electron-impact fragmentation of methane (excepting possible radiative contri- butions from short-lived electronic states).The cooler distributions characteristic of * Several interesting features are apparent in the spectra presented here.F. J. G R I E M A N , B . H . MAHAN, A . O'KEEFE AND J. s . WINN 201 the emission studies are most likely the result of the many thermalizing collisions which may occur at the high pressures employed. A comparision of the relative intensities of transitions originating in U" = 0 and U" = 1 also provides a measure of the degree of internal excitation. By scaling the observed transition intensity by the transition probability an estimate of the relative vibrational populations can be made. For CH+ a convenient line for this comparison is the Q9 line which is clearly separated from other components in both bands.Using calculated oscillator strengths tabulated by Kusunoki and Ottinger26 a vibrational temperature of ca. 5500 K is obtained. This value is nearly a factor of two greater than the corresponding temperature derived earlier from the rotational distribution. Laser-induced fluorescence spectra were also recorded for the 0 t 0 band of this transition using C,H, as a source of CH+. It was found that the apparent rotational and vibrational population distributions are essentially identical to those obtained using CH, as a source. The rotational distribution is quite broad (as is expected for a much smaller Boll) with a characteristic temperature of ca. 3000 K. The presence of transitions originating in U" = 1 is more pronounced in this case (as expected) permitting a more accurate determination of the molecular constants than for CH+.The feasibility of using the previously described experimental system to measure Examination of the spectra obtained for CD+ yields similar conclusions. t inle Ins result, T = 815 3 25 ns, is the average of several such measurements. FIG. 6.-Least-squares fit to the data collected from the decay of the A'II (u' = 0) level of CH+. The radiative lifetimes was tested by recording the radiative lifetimes of excited electronic states in several molecular ions for which reliable experimental data already exist. The systems chosen were the N2+B2C -+ X2C system with a u' = 0 radiative lifetime of 60 ns and the CO+A*ll + X2C system with a u' = 2 radiative lifetime of 3.25 p s . These two limits provide a sensitive test of the reliability of our approach for both short and long lifetimes. The results were in excellent agreement with accepted value^.^^-^^ Direct measurements of theA'I1 radiative lifetime were made for the u' = 0 level in both CH+ and CD+.The line used The results for CHf are presented in fig. 6.202 F. J . GRIEMAN, B. H . MAHAN, A . O’KEEFE AND J . s. WINN to pump this transition was the strong bandhead line consisting of the overlapping R3 and R, members of the 0-0 band (see fig. 4). The radiative lifetime determined in this study, zo = 8 15 ns, is larger than any previous experimental determination. This result is, however, in rather good agreement with the theoretical results of Yoshimine et al.1° which predict a radiative lifetime for the A’Il (u = 0) state of from 660-800 ns.The range of presently accepted experimental values spans 250 ns8 to 630 n ~ . ~ All previous attempts to record the radiative lifetime of this ion have relied upon some variation of the high-frequency electron-deflection technique described by Smith.’ This method is potentially subject to error in several respects. High-energy (several keV) electrons are used to create the excited CH+ from some parent neutral molecule. Such high-energy electrons make any specific state selection in the ion impossible and almost certainly result in population cascading from highly excited electronic states of the ion. A second significant source of error which may manifest itself is the rapid spacial dissipation of ions from the effective viewing region due to the strongly repul- sive ion-ion forces.Erman’ has attempted to minimize the distortions of the mea- sured decay curve due to the second effect by introducing low-energy electrons to the ionization region, thus neutralizing some of the space charge due to the positive ions. In this respect, it is significant that the result he obtained, zo = 630 ns, is that in closest agreement with our own. We measured the radiative lifetime for CD+A1n (u’ = 0) using the overlapping R3 and R5 lines (see fig. 5) as a pumping level. 50 ns, is slightly larger than that found for CH+ but the accompanying error is larger due to the weaker signal. The oscillator strength for the (0, 0) band of CH+ can be estimated by scaling the calculated value” (foo = 6.45 x by the ratio of the calculated to the measured radiative lifetimes.The result is fbo = (5.8 0.2) x The result obtained, 820 S U M MARY These experiments have probed the ro-vibronic distribution of the ground elec- tronic state of CH+ and CD+ fragment ions created by electron-impact ionization of CH, and CD,. These ions are found to have substantial rotational and vibrational excitation, and many new transitions originating from high-J states are observed. Essentially the same distributions are found whether methane or acetylene are used as parent gases. In addition, the radiative lifetime of the u = 0 level of the All3 state was measured and found to be in good agreement with calculated values, but consi- derably longer than previously measured values.The technique described here should prove to have wide application to the study of small reactive ions. We are grateful to Prof. C. B. Moore’s group for the use of and assistance with the signal averaging equipment used in this work. The Nd:YAG laser was supplied by the San Francisco Laser Center supported by the National Science Foundation under Grant CHE79-16250. This research was supported by the U.S. Department of Energy, Office of Basic Energy Sciences Division of Chemical Sciences under Contract No. W-7405-Eng-48. J. S. W. acknowledges support as an Alfred P. Sloan Research Fellow. We thank A. Kung for assistance with this laser. A. E. Douglas and G. Herzberg, Can. J. Res., 1942, 20, 71. W. S. Adams, Astrophys. J . , 1941, 93, 11. W. Klemperer and P.M. Solomon, Astmphys. J., 1972, 178, 389. F. J. Grieman, B. H. Mahan and A. O’Keefe, J . Chem. Phys., 1980, 72, 4246. W. H. Smith, J. Chem. Phys., 1971, 54, 1384.L I F OF TRAPPED MOLECULAR IONS 203 R. Anderson, D. Wilcox and R. Sutherland, Nucl. Inst. Meth., 1973,110, 167. N. H. Brooks and W. H. Smith, Astrophys. J . , 1974,196, 307. P. Erman, Astrophys. J., 1977, 213, L89. lo M. Yoshimine, S. Green and P. Thaddeus, Astrophys. J., 1973, 183, 899. P. H. Dawson and N. K. Whetton, Dyn. Mass Spectrom., 1971, 2, 1; P. H. Dawson, Int. J . Mass Spectrom. ion Phys., 1974, 14, 317; P. H. Dawson and M. Meunier, Int. J. Mass Spec- tram Iotr Phys., 1979, 29, 269; P. H. Dawson and C . Lambert, J . Vac. Sci. Technol., 1975, 12, 941. ’ J . Brzozowski, N. Elander, P. Erman and M. Lyyra, Astrophys. J., 1974, 193, 741. l 2 M. Benilan and C. Audoin, Itit. J. Mass Spectrum. Ion Phys., 1973, 11, 421. l 3 H. G. Dehmelt, A h . Aron7. Mol. Phys., 1967, 3, 53; 1969, 5, 109. l4 J. G. Pruett and R. N. Zare, J . Chem. Phys., 1976, 64, 1774. l6 A. E. Douglas and J . R. Morton, Asrvophys. J., 1960, 131, 1. l8 G. M. Almy and R. B. Horsfall Jr, Phys. Rev., 1937, 51, 491, l9 G. Herzberg and L. G. Mundie, J . Chem. Phys., 1940, 8, 263. ’O A. C. Hurley, Proc. R. Sue. Lotidon, Ser. A , 1961, 261, 237. S. Green, P. S. Bagus, B. Liu, A. D. McLean and M. Yoshimine, Phys. Reu. A, 1972, 5, 1614. 22 P. C. Cosby, H. Helm and J. T. Moseley, Astrophys. J., 1980, 235, 52. 23 P. C. Cosby and H. Helm, presented at Thirty Fifth Symposium on Molecular Spectroscopy, Columbus, Ohio, June, 1980. ’‘ M. Carre, Physica, 1969, 41, 63. 2 5 R. D. Knight and M. H. Prior, J. Appl. Phys., 1979, 50, 3044. 26 I. Kusunoki and Ch. Ottinger, J. Chetir. Phys., 1979, 71, 4227. ’’ L. W. Dotchin, E. L. Chupp and D. J. Pegg, J. Chenz. Phys., 1973, 59, 3960. 28 V. E. Bondybey and T. A. Miller, J . Chein. Phys., 1978, 69, 3597. 29 G. R. Mohlmann and F. J. DeHeer, Chem. Phys. Lett., 1976, 43, 170. D. S. King and P. K. Schenck, Laser Focus, March 1978, p. 60. A. Antid-Jovanovik, V. Bojovik, D. S. PeSie and S. Weniger, J. Mol. Spectrosc., 1979, 75, 197.

ISSN:0301-7249    

DOI:10.1039/DC9817100191

出版商:RSC    

年代:1981

数据来源: RSC

摘要:

The 4114 A Absorption System of the HCCS Radical BY S. L. N. G. KRISHNAMACHARI” AND D. A. RAMSAY Herzberg Institute of Astrophysics, National Research Council of Canada, Ottawa, Ontario, Canada KIA OR6 Received 15th December, 1980 The transient absorption spectrum observed in the region 3770-4170 A during the flash photolysis of thiophene is assigned to the HCCS free radical. Isotope shifts are observed when deuterated thiophene is used and establish the presence of a single H (or D) atom in the carrier. Under high resolution the bands show doublet P and R branches. The bands are tentatively assigned to a 211(b)-211(b) transition of a linear molecule, though the possibility that the molecule is slightly non- linear in one of the combining states is not excluded. Rotational assignments are given for the two strongest bands of HCCS, including the 0-0 band near 4114 A.The rotational constants for this band are Bo” = 0.188 41(6) cm-’ andBo’ - Bo” = -0.014 399(6) cm-l, where the error limits are 30. (See, however, note added in proof.) A new transient absorption spectrum’ has recently been observed in the region 3770-4170 A during the flash photolysis of thiophene (C4H4S). The spectrum of the SH radical was also observed and it was tentatively suggested that the new spectrum might be assigned to the C4H3 radical. In the present work the experiments have been repeated using deuterated thiophene and the isotope shifts show that the carrier of the bands contains one H (or D) atom. The spectrum of the normal compound has also been investigated under high resolu- tion so that the rotational fine structures of the bands are observed.These rotational structures provide strong evidence that the carrier of the bands is a linear molecule in at least one of the two combining states. If the analysis is carried out on the basis of a linear t+ linear transition, the ground-state rotational constant which is obtained is consistent with the assignment of the bands to the HCCS free radical. However, further work needs to be carried out, especially with the deuterated species, to deter- mine if the radical is linear or bent in the ground state. The present paper therefore constitutes a preliminary report on the rotational and vibrational analysis of the spec- trum. EXPERIMENTAL A N D RESULTS The flash photolysis experiments were carried out using an apparatus which has been described earlier.2 Thiophene from Aldrich Chemical Co.was used at a pres- sure of 0.25-0.50 Torr (33-67 N m-2) and was diluted with 100 Torr (13 332 N m-2) of argon. The photolysis flash was produced by discharging two 288 pF capacitor banks charged to A6 kV. The time delay between the photolysis and source flashes was ca. 25 p s . Survey spectra were photographed using the 2nd order of a 6.6 m Eagle spectro- graph and Kodak 103-0 plates. High-resolution spectra were obtained using the 14th and 15th orders of a 7.3 m Ebert spectrograph. Path lengths up to 24 m were used * Visiting Scientist from the Spectroscopy Division, Bhabha Atomic Research Centre, Bombay 400085, India.206 SPECTRUM OF THE HCCS RADICAL and the number of flashes required to give satisfactory exposures for the Ebert spectra varied from 60-1 20 depending on the particular band under investigation. A sample of deuterated thiophene was prepared by the method described by Dawson and G i l l i ~ .~ A mixture of 17 cm3 CF,COOD (Merck, Sharp & Dohme), 4 cm3 D,O and 5 cm3 thiophene was heated to 90 "C and agitated for 17 h. The reaction mixture was cooled, diluted with 20 cm3 H,O and the thiophene separated. The procedure was repeated after which the purity of the sample was checked by examining the infrared spectrum. The weakness of the C-H stretching bands near 3100 cm-' compared with the C-D stretching bands near 2300 cm-I indicated that a high degree of deu- teration had been achieved.The flash photolysis experiments were repeated using the Eagle spectrograph, and isotope shifts were observed for all the bands. The bands of the normal species ap- TABLE 1 .-BAND-HEAD MEASUREMENTS FOR HCCS W A int a v,,,/cm-' Av from 0-0 assignment band/cm- ' 4172.95 4149.32' 41 13.73 4089.15 4082.89 4073.80 4054.33 4053.67 4046.85 4025.99 4005.44 3992.13 3969.05 3967.73 3959.78 3946.57 3936.03 3934,19 3933.2 1 3928.39 3921.49 3910.12 3889.13 3880.61 3877.67 3856.16 3 825.65 3823.70 3819.72 3780.18 3772.82 3768.81 1 2 8 3 1 1 6 3 1 3 2 10 6 1 1 1 8 2 3 1 1 2 5 4 7 2 4 3 1 2 1 1 3762.69 b , c 1 23 957.1 I 24 093.55 301.98 448.09 485.53 540.15 658.01 662.04 703.56 831.60 958.97 25 042.18 187.85 196.23 246.76 331.29 399,15 41 1 .OO 417.34 448.50 493.27 567.42 705.40 761.82 781.36 925.20 26 131.93 145.27 172.50 446.24 497.88 526.07 569.18 - 344.87 - 208.43 0 146.11 183.55 238.17 356.03 360.06 401.58 529.62 656.99 740.20 885.87 894.25 944.78 1029.31 1097.17 1109.02 11 15.36 1146.52 1 191.29 1265.44 1403.42 1459.84 1479.38 1623.22 1829.95 1843.29 1870.52 2144.26 2195.90 2224.09 2267.20 - 208 0-0 356 - 208 = 148 356 740 - 208 == 532 657 740 740 + 356 - 208 = 888 740 + 356 = 1096 2 X 740 - 208 = 1272 1403 or 740 + 657 = 1397 740 + 2 x 356 = 1452 2 X 740 = 1480 2 X 740 + 356 - 208 =1628 2 X 740 + 356 = 1836 740 + 1403 = 2143 or 2 x 740 + 657 =2137 2 X 740 + 2 x 356 = 2192 3 x 740 = 2220 ~~ ~~~~ ~ (I Visual estimates.* These measurements are from Eagle spectrograms while the rest are from Several additional weak bands have been observed as far as 3233 A and will Ebert spectrograms.be discussed in a later publication.S . L . N. G . KRISHNAMACHARI A N D D . A . RAMSAY 207 TABLE 2.-BAND-HEAD MEASUREMENTS FOR DCCS Av from 0-0 & l ~ int a v,,,/cm-l band/cm-' assignment 4197.58 4157.19 4 1 43.49 41 14.91 4105.26 4097.8 8 4097.10 4090.69 4085.90 4085.59 407 1.06 4063.02 4060.84 4060.46 4049.48 4049.12 4035.23 4007.91 3986.56 3979.33 3977.02 3969.83 3966.16 3964.99 3963.52 3960.09 3954.17 3946.46 3946.07 3942.75 3942.51 393 1.90 3930.88 3895.27 3892.79 3881.07 3874.93 3872.27 3868.62 3866.94 3851.38 3 847.74 3836.56 3 8 3 6.50 3834.77 3784.13 3780.1 1 3774.26 3768.83 1 1 2 1 7 3 2 3 2 2 2 9 1 1 2 2 2 2 7 3 2 1 1 1 1 1 2 10 3 1 1 1 2 1 1 1 4 2 2 2 1 1 6 6 2 2 1 1 3 23 823.26 24 054.73 134.21 301.85 358.97 402.84 407.51 445.75 474.39 476.28 563.64 61 2.25 625.45 627.75 694.54 696.74 78 1.72 950.67 25 084.30 129.86 144.48 189.99 213.31 220.73 230.12 25 1.97 289.78 339.16 341.68 362.99 364.59 432.98 439.62 672.15 688.50 766.10 806.94 824.65 849.01 860.25 964.75 989.27 26 064.99 065.45 077.15 426.13 454.25 495.27 533.43 - 535.71 - 304.24 - 224.76 -57.12 0 +43.87 48.54 86.78 115.42 117.31 204.67 253.28 266.48 268.78 335.57 337.77 422.75 591.70 725.33 770.89 785.51 83 1.02 854.34 861.76 871.15 893.00 930.81 980.19 982.71 1004.02 1005.62 1074.01 1080.65 1313.18 1329.53 1407.13 1447.97 1465.68 1490.04 1501.28 1605.78 1630.30 1706.02 1706.48 1 171 8.1 8 2067.16 2095.28 2136.30 2 1 74.46 0 - 0 253 725 725 + 253 = 978 2 X 725 = 1450 2 X 725 + 253 = 1703 3 x 725 = 2175 a Visual estimates.Several additional bands have been observed in the region 3663.7 to 3525.7 A.208 SPECTRUM OF THE HCCS RADICAL TABLE 3.-vACUUM WAVENUMBERS AND ROTATIONAL ASSIGNMENTS FOR LINES IN THE ABSORPTION BANDS OF HCCS (IN cm-') 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 24 301.983 H 01.888 01.788 01.670 01.530 01.357 01.155 00.931 00.669 00.392 00.086 299.73 6 99.371 98.967 98.559 98.509 98.113 98.049 97.621 97.544 97.104 97.018 96.566 96.464 96.01 1 95.849 95.409 95.276 94.792 94.626 94.120 93.974 93.455 93.273 92.739 92.543 91.974 91.787 91.224 91.01 1 90.41 6 90.203 89.597 89.365 88.744 88.500 87.857 87.599 86.947 86.676 85.990 85.729 85.031 84.741 84.040 83.723 24 295.849 95.276 94.626 93.974 93.273 92.543 91.787 91.01 1 90.203 89.365 88.500 87.599 86.676 85.729 84.741 83.723 82.690 81.620 80.519 79.396 78.237 77.05 1 75.843 74.616 74.570 73.357 73.288 72.066 71.996 70.746 70.668 69.387 69.289 68.01 8 67.927 66.609 66.493 65.174 65.053 63.706 63.572 62.204 62.055 60.681 60.537 59.137 58,966 57.555 57.383 55.958 55.763 54.328 54.11 1 52.651 52.435 50.989 50.715 49.247 48.992 47.539 47.229 45.720 45.429 43.924 43.604 42.088 41.759 40.234 39.862 38.357 37.960 36.427 36.022 25 042.180 H 42.039 41.970 41 329 41.680 41.490 41.277 41.03 1 40.761 40.458 40.126 39,759 39.377 38.957 38.515 38.059 38.008 37.562 37.488 37.022 36.947 36.462 36.374 35.870 35.780 35.249 35.134 34.61 1 34.491 33.923 33.803 33.216 33.090 32.483 32.336 31.707 31.562 30.924 30.773 30.099 29.929 29.247 29.061 28.361 27.429 26.529 25.540 24.532 23.531 22.476 21.380 20.274 19.134 28.179 27.23 1 26.299 25,301 24.295 23.260 22.181 21.107 19.947 18.800 25 037.217 36.678 36.092 35.480 34.850 34.170 33.457 32.728 3 1.970 31.182 30.365 29.516 28.641 27.732 26.796 25.834 24.835 23.812 22.762 21.673 20.567 19.421 18.245 17.044 15.814 14.551 13.269 11.971 11.901 10.629 10.545 09.260 09.180 07.860 07.767 06.442 06.323 04.969 04.850 03.482 03.356 01.968 01.812 00.41 3 00.260 24 998.843 24 998.684 97.227 97.070 95.600 95.422 93.928 93.752 92.248 92.041 90.529 90.321 88.777 88.552 87.01 1 86.753 85.208 84.939 83.357 83.085 81.507 81.204 79.617 79.296 77.685 77.362 75.754 75.405PLATE 1.-Rotational fine structure and assignments for the 0-0 band of the 41 14 A absorption system of the HCCS radical.[To face page 209S . L . N. G . KRISHNAMACHARI A N D D . A . RAMSAY TABLE 3.-Continued 209 0 -0 band" N R(N) 56 57 58 59 60 61 34.469 34.070 32.510 32.066 30.529 30.021 28.453 27.988 26.405 24.331 23.781 73.785 73.410 71.771 71.392 69.740 69.334 67.647 67.263 65.582 65.124 The low-frequency components of the R doublets at high N are overlapped by lines of the P branch (see plate 1). peared only weakly in the spectrum, thus providing further evidence for the extensive deuteration of the thiophene sample. The amount of material available, however, was insufficient for high-resolution studies to be carried out with the Ebert spectro- graph. All the spectra were calibrated by using an iron hollow-cathode lamp to provide a reference spectrum.Measurements of the bands were carried out using a comparator equipped with a photoelectric scanning device. Standard wavelengths were taken from the tables of Cro~swhite.~ Band-head measurements for the normal species are given in table 1 and agree well with the measurements given by the earlier workers.' Band-head measurements for the deuterated species are given in table 2. The vacuum wavenumbers for the lines of the two strongest bands in the spectrum of the normal species are given in table 3. The precision of the measurements is considered to be &0.005 em-' for lines which are not overlapped. DISCUSSION (a) ROTATIONAL ANALYSIS The strong band with a head at 41 13.73 A lies close to the long-wavelength end of the spectrum and can be assigned with reasonable certainty to the 0-0 band of the electronic transition.The rotational fine structure of this band is shown in plate 1 . Most of the stronger lines can be accommodated into a P and R branch with small doublings at the higher N values. There is no evidence for a strong Q branch though, in view of the many weak unidentified lines, the possible presence of a weak Q branch cannot be excluded. The band therefore has the appearance of a C-C or JJ-II, A-A, , . . transition of a linear molecule, or a A& = 0 transition of a molecule which is linear in one state and non-linear in the other.' The rotational numbering for a P and R branch cannot be determined unambi- guously unless further information is available; only the relative numberings of the lines can be established.The absolute numbering can be determined, however, if satisfactory combination differences can be obtained from several bands having one state in common. The 0-0 band and the strong band at voo + 740 cm-' can be num- bered (table 3) so that the ground-state combination differences agree, usually to with- in &0.02 cm-'. Doubling of the P and R lines is observed for N values greater than ca. 25 and, in forming combination differences, mean values for the doublets were used. The same combination differences can be obtained from the band at voo + 1479 cm-', though the measurements for this band are less extensive and less accurate.210 SPECTRUM OF THE HCCS RADICAL The ground-state combination differences for the 0-0 and +740 cm-’ bands were fitted by least squares to the equation A,F”’(N) = 4 B,”(N + +) - 8 D,”(N + +)3 and the following values for the rotational constants were obtained: Bo” = 0.188 41(6) cm-’ Do” = 4.3(18) x lo-* cm-I where the error limits in brackets are 30.The standard error for the fit of 72 combina- tion differences was 0.0095 cm-’. Upper-state rotational constants were obtained in the following manner. Ground- state term values were calculated from the above constants and were added to the frequency of each line to obtain the corresponding upper-state term values. These were then fitted by least squares to the equation F’(N) = VO + B’N(N + 1) - D’NZ(N + 1)2 and the values for the constants are given in table 4.The fit for the 0-0 band is not quite as good as for the +740 cm-I band since the P and R branches overlap for the former band but not for the latter. TABLE 4.-BAND ORIGINS AND ROTATIONAL CONSTANTS (IN Cm-’) a VO 24 299.690( 6) 25 039.938(5) B’ - B” - 0.01 4 428(9) -0.014 764(8) 108(D’ - 0”) - 0.95(27) - 2.04(24) B’ 0.173 98(6) 0.173 64(6) 108D’ 3.3( 18) 2.3(18) fitted of fit no. of lines 87 93 standard deviation 0.0073 0.0063 a The error limits are 30. Note that the error limits for the differences of the constants (B’ - B”) and (D’ - D”) are considerably smaller than the error limits for the constants individually. Attempts have been made to fit the bands at + 146, +360, +886 and + 1097 cm-l from the 0-0 band but none of these bands gives combination differences which agree with the present ground-state values.However, the difficulties in obtaining absolute numberings in bands without Q branches is well-known and the possibility that the present numbering may need to be revised cannot be eliminated at this stage, especially in view of the overlapping found in the 0-0 band. (b) ASSIGNMENT OF THE CARRIER From the nature of the experiment it is reasonable to assume that the carrier of the bands does not contain atoms other than C, H and S. From the isotopic shifts ob- served when deuterated thiophene is used it is clear that the carrier contains only a single H (or D) atom. The ground-state rotational constant is too small for the car- rier to be assigned to a diatomic molecule and suggests that a small polyatomic mole- cule is responsible. The value for B,” does not agree with the known B-values for C2H (1.4568 cm-1)6 or C4H (0.1587 cm-l)’ and is too small to be compatible withS .L . N. G. K R I S H N A M A C H A R I A N D D . A . RAMSAY 21 1 C3H or HCS. Although the HS, radical if linear might be compatible with the observed B-value, a spectrum has already been assigned to this radical* and is not observed in the present experiments. It seems reasonable to assume that the carrier contains one or more S atoms since another new spectrum has been reported in a similar region during the flash photolysis of selenophene (C4H4Se).9 The only molecular species which appears to be compatible with all the observations is the H-C=C-S radical. If typical values are taken for the various bond lengths," e.g. r(C-H) = 1.08 A, r(C=C) = 1.34 8, and r(C-S) = 1.56 A, and the radical is assumed to be linear, then the calculated B-value is 0.1916 cm-l close to the experimental value of 0.188 41 cm-'.Since the radical has an odd number of electrons, the electronic transition is expected to be doublet-doublet in character which is consistent with the appearance of doublet splittings in the spectrum. (C) N A T U K E O F T H E E L E C T R O N I C T R A N S I T I O N The HCCS radical is isoelectronic with NCS which has a well-known absorption spectrum" in a similar region of the spectrum. Two electronic transitions have been assigned to the NCS spectrum; the stronger is an A('IIi)-X('IIi) transition with its origin at 26 054 cm-I and the weaker is a B(ZC+)-X(211i) transition with its origin at 26 844 cm-'.There is extensive vibronic interaction between the A('IIi) and B('C+) states. It is tempting to assign the HCCS spectrum to the corresponding 'Il--'FI transition in view of the rotational structure of the bands. For NCS, the ground and excited 'll states both belong to case (a). Two sub-bands are observed, viz. 'n3/2-'n3/2 and 2I-I 1 ,'-Zn 1/2, each consisting of a P and R branch and a very weak Q branch. No doubling of the lines has been observed even at the highest J values. For HCCS it is more probable that the transition is of the type 'KI(b)-'II(b) in view of the doublet splittings which have been observed. Alternatively, the molecule may be slightly non-linear in one of the states.A large departure from linearity is ruled out by Franck-Condon considerations [see (d) below]. (d) VIBRATIONAL ANALYSIS The assignment of the 0-0 band is confirmed by the isotope shifts; the 0-0 band is shifted by 57 cm-' to higher frequency by deuteration. The strong band at \roo -4- 740 cm-' in the spectrum of HCCS corresponds to the strong band at voo 1- 725 cm-I in the spectrum of DCCS. Both frequencies lie close to the C-S stretching frequency (755 cm-') in the 'II excited state of NCS. Further- more the r value for this vibration in NCS ( x = {-0.000 40 cm-') is very similar to that found for HCCS (CL = t0.000 34 cm-', table 4). The assignment of the 740 and 725 cm-I intervals to the C=S stretching frequencies in HCCS and DCCS can there- fore be made with confidence.The +356 cm-I interval for HCCS corresponds to the $253 cm-I interval for DCCS. The isotope ratio p = 0.71 suggests that a hydrogenic motion is involved. Most probably the bands are sequence bands involving the HCC (or DCC) bending vibration. The strong bands at voo + 1097 cm-I for HCCS and voo + 980 cm-' for DCCS are readily explained in terms of combinations of the sequence intervals with the C=S stretching frequencies. Similar assignments can be made with two quanta of the C-S stretching frequencies (tables I and 2). The above assignments account for most of the strong features in the spectra of the two molecules. However, there are numerous weak features which remain to be Both frequencies are observed up to u = 3.212 SPECTRUM OF THE HCCS RADICAL assigned.For HCCS all of the strong bands have weaker bands associated with them at 208 cm-I to lower frequency; for DCCS, a similar interval appears with some of the stronger bands but not with the 0-0 band. The + 1403 cm-’ interval for HCCS may correspond to a fundamental frequency in the excited state or to the combination 740 + 657 = 1397 cm-’. The corresponding interval for DCCS is not immediately apparent. CONCLUDING REMARKS The assignment of the spectrum to the HCCS radical is supported by the method of observation of the spectrum, the isotope shifts when deuterium is substituted, the rotational analysis of the bands and the vibrational structure of the spectrum. The spectrum is tentatively assigned to a 21&21-I transition in which both states are close to Hund’s case (b) though it is possible that the molecule might be slightly non-linear in one of the combining states.Many similarities have been noted between the spectra of HCCS and NCS. Indeed the presence of many unexplained features in the spectrum of HCCS may arise from Renner-Teller interactions involving the bending vibrations and from vibronic interactions with a second excited state. Much work remains to be done on the rota- tional analyses of the bands of HCCS and DCCS and also on the bands at shorter wavelengths which may constitute a second band system. Note added in proof: Further work on the rotational analyses of other bands of HCCS and DCCS suggests that half-integral values of the rotational quantum numbers are probably to be preferred. The rotational quantum numbers of the R lines in table 3 should be increased by + and those of the P lines should be decreased by f; The revised value for the ground-state rotational constant is Bo” = 0.195 68(8) cm- . We are indebted to Dr. G. Herzberg for helpful discussions and for kindly reading the manuscript. We thank Dr. L. C. Leitch for his advice on the preparation of the deuterated thiophene and Dr. D. G. Cameron for obtaining the infrared spectra. We also acknowledge the valuable assistance of Mr. M. Barnett in all phases of the experi- mental work. S . L. N. G. Krishnamachari and T. V. Venkitachalam, Chem. Phys. Lett., 1978, 55, 116. J . W. C. Johns, S. H. Priddle and D. A. Ramsay, Discuss. Furaday Soc., 1963, 35, 90. K. M. Dawson and R. G. Gillis, Aust. J. Chem., 1972,25, 1221. H . M. Crosswhite, J. Res. Nat. Bur. Stand. USA, Sect. A , 1975, 79, 17. G. Herzberg, Electronic Spectra of Polyatornic Molecules (D. Van Nostrand Co., Inc., Princeton, N.J., 1966). K. D. Tucker, M. L. Kutner and P. Thaddeus, Astrophys. J. Lett., 1974, 193, L115. ’ M. Guelin, S . Green and P. Thaddeus, Astrophys. f. Lett., 1978, 224, L27. * G. Porter, Discuss. Furuduy Soc., 1950, 9, 60. lo M. D. Harmony, V. W. Laurie, R. L. Kuczkowski, R. H. Schwendeman, D. A. Ramsay, S. L. N. G. Krishnamachari and T. V. Venkitachalam, Chem. Phys. Lett., 1979, 67, 69. F. J. Lovas, W. J . Lafferty and A. G. Maki, J. Phys. Chem. Ref. Data, 1979, 8, 619. R. N. Dixon and D. A. Ramsay, Cun. f. Phys., 1968,46, 2619.

ISSN:0301-7249    

DOI:10.1039/DC9817100205

出版商:RSC    

年代:1981

数据来源: RSC

摘要:

Doppler-limited Laser Spectroscopy of Electronic Transitions in SnO BY MICHAEL A. A. CLYNE AND MICHAEL C. HEAVEN? Department of Chemistry, Queen Mary College, Mile End Road, London E l 4NS Receiued 1 st December, 1980 Laser-induced fluorescence with narrow-band excitation and time resolution has been used for the first time to investigate the quantum-resolved dynamics of excited states of '20Sn0. Transitions to both the A'lT and 6' "(1) excited states were observed. The 1-0 band of SnO A-X is unperturbed, whilst the u' = 3 and u' =1 4 levels of SnO A'll are perturbed. The u' = 3 level of A'lT and the u' x 13 level of b'3il(l) show a strong mutual perturbation. The u' = 4 level of A'II is somewhat more weakly perturbed by a 'A(1) or jII(2) state. Analysis of the spectra of the A-X and b'-X transi- tions is described, Lifetimes (z,) of selected ro-vibrational states of the ATI and b 'I'I(l) manifolds, and quenching rate constants with O2 (kM), are reported for the first time as follows:- A'n, V' = 1 : To = 160 * 20 nS; k M = (2.2 f 0.6) X lo-'' Cm3 S-'; A'n, V' = 3 : To = 140 10 nS; k M = (6.9 3 0.6) X lo-'' Cm3 S-'; A'n, V' = 4 b' "(I), 0' = 14 : 70 = 580 & 34 ns; k M = (3.6 & 1.0) X lo-'' Cm3 S - ' : To = 130 & 20 nS; k M 1.4 X 10-" Cm3 S - l ; Fundamental information regarding the nature, spectroscopic constants and life- times of excited electronic states is most easily derived from absorption spectra.However, many existing data on free radicals and transient molecules are derived from emission spectra, which are inherently more complex and more difficult to analyse.It is a basic limitation of absorption spectroscopy that a reasonably long optical depth of absorber is required, thus leading to the need for large absorber concentrations. For labile species, this requirement sometimes can be satisfied by flash-photolytic production and absorption spectroscopy in real time. An alternative approach, which eliminates the need for high radical concentrations and is thus potentially very powerful, is to use a narrow-band dye laser to excite fluorescence.2 When undispersed fluorescence intensity is observed, as a function of laser frequency, the resulting laser excitation spectrum resembles the corresponding absorption spectrum, although the line intensities are modified by perturbations or predissociations in the excited state^.^ Additional information, regarding the ground state, is obtainable through analysis of the fluorescence spectrum.The approach is extremely sensitive, and has been used successfully to detect free-radical densities as low as lo6 ~ m - ~ . Either continuous or pulsed formation of radicals may be employed, the former method being experimentally the simpler. In this method, labile species are generated by reactions of atoms in a discharge-flow system. In this paper; we exemplify the utility of laser-induced fluorescence (LIF), as a means of unravelling the nature and lifetime of several excited states of the SnO mole- cule. Over the years, a large amount of effort has been expended on the analysis of electronic spectra of Sn0.4-9 However, many uncertainties have remained, including an absence of lifetime data for the excited states involved.We show that a relatively limited effort using LIF has succeeded in solving outstanding problems in the t Present address: Bell Laboratories Inc., Murray Hill, New Jersey, U S A .214 LASER SPECTROSCOPY OF ELECTRONIC TRANSITIONS quantum-resolved dynamics of excited SnO. Recent interest in SnO has arisen from the possibility of using the low-lying a-X transition as a chemically-pumped laser," since the elementary reaction, Sn + NzO -+ SnO (a) + Nz, has an appreciable photon yie1d.lO-l EXPERIMENTAL Ground-state SnO molecules were generated in a discharge-flow system, using the reac- tion of Sn(CH,), with a flow of 0 3P atoms that were produced by a microwave discharge in 02.Detection of SnO fluorescence was made in a flowing cell, using pulses from a frequency- doubled, narrow-band dye laser. GENERATION O F SnO MOLECULES Ground-state 0 3P atoms in O2 carrier gas were reacted near 0.5 Torr total pressure with Sn(CH3), (Aldrich). Weak visible chemiluminescence was observed to emanate from the reaction zone; however, little or no ultraviolet emission (which would interfere with detection of fluorescence) was detected. As an alternative possible source of SnO, the reaction 0 3P + SnCI,, was examined. However, no SnO LIF was observed in the products of this reaction. Addition to a stream of 0 3P atoms to trace flows of Sn(CH3), (<0.1 mol%) was found to give intense LIF of SnO (A-X) from the reaction products.Natural Sn has several abundant isotopes, of which the principal species are:-1 16, 14.2%; 117,7.60/,; 11 8,24.0%; I19,8.6%; 120, 33.0%; 122, 4.6%; and 124, 6.0%. This diversity of isotopic species leads to consider- able complexity in the spectra of SnO. Therefore, a sample of 'ZoSn(CH,), was prepared from 200 mg of 98% pure lz0Sn, supplied by the Electromagnetic Separation Group (Chemis- try Division, A.E.R.E., Harwell). This sample of '20Sn(CH3)4 was sufficient for several LIF experiments with SnO, and greatly simplified the rotational analysis of perturbed bands. The lZ0Sn(CH3), was synthesized from 120Sn by first oxidizing the metal to SnCl, with elemental CIz, then reacting the SnCl, with lithium methyl: SnCl, + 4LiCH3 -+ Sn(CH3), + 4LiCl.The products were separated by vacuum distillation, and purified from solvent ether by frac- tionation. DETECTION OF SnO BY LIF Excitation of SnO molecules was carried out in a laser fluorescence cell that formed part of a flow system [see fig. 2 of ref. (1 3)]. The laser crossed the flow system ca. 2 cm downstream from the zone of reaction between Sn(CH3)4 and 0 3P atoms. Detection of fluorescence was by means of a fast photomultiplier (2 ns rise-time, EM1 9816B), and a filter (Chance OBlO, 50% transmission range 350-480 nm), to block the excitation wavelength and most of the visible chemiluminescence. It consisted of an oscillator-amplifier configuration, pumped by a Nd-YAG laser, also based on an oscillator-amplifier design. The wavelength of the dye oscillator could be scanned continuously and linearly, using pres- sure-scanning of an air-spaced Fabry-Perot etalon coupled to an echelle grating.Dye fundamental wavelengths between 630 and 670 nm were generated using ethanolic solutions of Rhodamine 640, Cresyl Violet Perchlorate, or Oxazine 720. Typical pulse energy was 2-5 mJ at 10 Hz, which was frequency-doubled in a KDP crystal, to give ca. 100-250 pJ per pulse between 315 and 325 nm, with bandwidth of 0.05 cm- '. Because of the flat angle- tuning curve of KDP near 320 nm, no adjustment of the doubling crystal was needed during each pressure scan of 0.15 nm. Laser-excitation spectra of SnO were obtained by recording undispersed (filtered) fluores- cence intensity, processed with a boxcar integrator (Brookdeal), as a function of laser wave- length.Lifetime measurements were made with a detection system estimated to have a time The dye laser has been described previ0us1y.l~M . A . A . C L Y N E A N D M . C. HEAVEN 21 5 constant < 20 ns. Each fluorescence decay waveform from the photomultiplier was captured with a fast transient digitizer (Biomation 6500), and averaged in a mini-computer (Nicolet LABSO). Data were stored on floppy discs, and were analysed with BASIC programs, using standard statistical procedures to give lifetimes.14* lo Logarithmic decay curves could be obtained using the LABPAC program of the minicomputer. RESULTS SPECTROSCOPIC BACKGROUND The most well-known band system of SnO is the A'II-X'C+ transition, whose bands lie at wavelengths 2300 nm.This is one of the transitions studied in the present work. Some confusion arises in the literature, since certain authorsI6 have labelled the A'II state as D l n ; we follow recent workers9* in adopting A'n. Each band consists of one P, Q and R branch. The combination defect is negligible, since the II-doubling of the upper AlII state is very small.' A complication is the presence of several natural isotopic species of SnO, principally l2OSn0. ''sSnO and 'l6SnO. The The rotational structure of the A'II-X'C+ transition is very simple. 3c 3 I rn 2 b . n L W 2! I I I I I I I 1.6 1.8 2.0 2.2 rlA FIG. 1.-Potential energy curves of SnO. Data shown are Morse functions for XIC+, A'll and b' 311(l); 311(2) and 'l-I(O', 0-) minima are indicated. Note nesting of A W with the components of ' H i , which facilitates mixing of states and overlapping of spectra in certain wavelength regions.216 LASER SPECTROSCOPY OF ELECTRONIC TRANSITIONS isotopic shifts in the A-X system are small, due to the small magnitude of the vibra- tional frequencies, we'' and and the closeness of the reduced masses of the iso- topic species.The A-X transition of SnO is the analogue of the Fourth Positive system (A'II- X'C+) of CO." However, whilst intercombination transitions (singlet + triplet) are weak for the light CO these transitions become more allowed for the hea- vier monoxide radicals of Group IV elements, such as SnO. There is a tendency to- wards Hund's case (c) coupling in the excited states of SnO, thus leading to radiative transitions between the various L2 sub-states of 311 and theground XIC+ state.Some of these transitions are known,6 and are designated Y3II( l)-X'C+ and b311(O+)-X'Cf. The former transition is the analogue of the A311(l)-X1Z+ system of the halogens," and shows single P, Q and R branches. The latter transition is the analogue of the B3n(Of)-X1C+ system of the halogens," and shows single P and R branches only. There should also be "((2) and 3rI(O-) components of the 311 state of SnO, which do not have allowed radiative transitions to the ground state. In addition, there is a low- lying a3C+-X'C+ transition that shows discrete str~cture.~*'O Fig. 1 shows a summary of the relevant energy states, based on the data of ref. (9) and (16). LASER EXCITATION SPECTRUM OF THE 1-0 BAND OF SnO (A-X) The 1-0 band of SnO ( A - X ) , lying between 332 and 337 nm, is the only band that has been rotationally analysed previously, and which also is accessible to laser excita- tion from a Boltzmann vibrational distribution near 300 K (>99% in U" = 0).Fig. 2 shows a portion of the first laser excitation spectrum of the 1-0 band of natural SnO. For J 3 30, the spectrum appears as distinct clumps of lines, which can be assigned to the seven most abundant isotopic species of Sn. Assignments of the two most abundant species, l2OSn0 and "*SnO, are shown in fig. 2. The expected single P, Q and R branches were readily identified for this 'I'I-'Z+ transition. 332.9 laser wavelength/nm 333.0 33 3 .I 0 ;o I Q) c 3 u= R c.l .F. 52 i2 1'8s FIG. 2.-Laser-excitation spectrum of 1-0 band of SnO (A-X).Natural SnO. Note clumps of lines, with fine structure due to the small isotopic shifts.M. A . A . C L Y N E AND M. C. H E A V E N 217 It is noted that the P(J) and R(J + 14) lines are completely overlapped from the band head, up to at least J = 27. At higher J-values, the calculated P and R lines separate, as was reported by Lagerqvist et al.,7 who analysed the emission spectrum from a rotationally hot source of SnO with J < 112. The observed P and R lines in our work ( J < 40) do not clearly separate; but the R lines of the doublets are rather weak due to the low rotational temperature (ca. 300 K). However, the extension to lower J values (J < 40) of the analysis by Lagerqvist et al.7 evidently is erroneous, since satisfactory combination differences cannot be obtained from their assignments below J = 40. Inevitably, the spectrum of the emission, such as that used by Lager- qvist et is more complex than that of the present laser-excitation spectrum.This greater complexity arises from the higher rotational temperature and greater linewidths, compounded by overlapping of bands from different U” progressions. A supposed perturbation of J = 40, which was reported by Lagerqvist et a[.,’ is accounted for by an inadvertent transition by these workers from the correct rotational assignments above J = 40, to erroneous assignments below J = 40. In any event, we were unable to detect any irregularity in the combination differences for either A doublet, in the vicinity of J = 40. ISOTOPIC SHIFT I N T H E 1-0 BAND The isotopic shifts in the 1-0 band could be readily measured.Accurate disper- sion for the SnO spectra was obtained by determining an excitation spectra of BrCl (B-X) under the same conditions. The ground-state combination differences for BrCl spectra were formed. Since the ground-state rotational constants B,,, are known accurately from Coxon’s work,20 the spectral dispersion, which is precisely constant, is reliably defined. As an example, the 120-118 isotopic shift 6v’ was measured for six pairs of Q branch lines between J == 12 and J = 20, and determined to be (0.443 & 0,017) cm-l (20). For low J values, most of the isotopic shift is due to the vibrational shift 6Gi, which may be calculated from eqn (1):- 6G’ = (p - i)C!l,’(U‘ + 3) - (p’ - l)coe’Xe’(V’ + *)z -(P - l)to,”(u” + 5) - t (P’ - l)OJc”Xe”(U” + 3)2.Using data for we and w,x, from ref. ( I 6), the magnitude of 6G’ was calculated to be 0.437 cm-l. There is a small rotational isotope correction of -0.023 cm-l to be made, giving 0.414 cm-’ as the calculated value of 6 ~ ’ for the 1-0 band. Considering the small magnitude of 6v’ (some 4.9 pm in wavelength), the agreement of observed and calculated values is quite satisfactory. The sign and magnitude of the isotopic shift confirms unequivocally the vibrational numbering of the A l I I state given by Lagerqvist et Smith and Meyers suggested a revised numbering that would be inconsistent with the present isotopic shift. Theys based their suggestion on analysis of the rather broad spectral features of matrix- isolated SnO.It seems likely that their analysis* was complicated because of over- lapping between the A-X and the b’-X systems of SnO, as is found in this paper. LASER E X C I T A T I O N SPECTRUM OF THE 4-0 BAND Laser-excitation spectra of the 4-0 and 3-0 bands of SnO ( A - X ) showed heads near the expected wavelengths. These bands have been seen previously only at low resolu- tion, in work carried out many years ago? The present rotationally resolved spectra218 LASER SPECTROSCOPY OF ELECTRONIC TRANSITIONS '12 '13 of natural SnO showed a complex line structure which, unlike that of the 1-0 band, appeared to be intractable. It then became apparent that both the 4-0 and 3-0 bands are perturbed, and also show additional transitions which overlap, Consequently, the intractability of the mixed-isotopic spectra of SnO is not surprising, and the whole analysis was based on the spectrum of '*OSnO.In the 4-0 band a strong perturbation centred at J' = 18 is observed, which affects the P, Q and R branches equivalently. The perturbation extends down to low J values, and the observed band origin v4,0 is shifted. A homogeneous perturbation (IAAI =0) therefore is indicated. The occurrence of the maximum perturbation at similar J values in the Q (e sub-level) and in the P and R [f'sub-level] branches rule out 3C- as the interacting state, since a 111-3C- perturbation occurs at three different J values in the P, Q and R branches.21 Therefore, spectra were obtained using l2OSn0. 0 '19 ,20 ,21 I22 '14 '15 '16 '17 '18 16-01 ,16 ,I7 , , R '22 23 I 24 '25 '26 '27 r-- ~~ 31i.85 31i.90 314.95 laser wavelength/nm FIG.3.-Laser-excitation spectrum of 4-0 band of '20Sn0 (A-X). at J' = 18, affecting all three branches. Note strong perturbation centred Extra lines due to the perturbing state also are shown. Rotational assignments were based on the combination differences of the (un- perturbed) ground XIC+ state. A value of 0.3555 & 0.0006 cm-' was obtained for the constant B,", in fair agreement with B," = 0.354 648 cm-', reported by Tor- ringz6 from microwave data. Our value for B," was based on both PR and RQ com- bination differences ; the combination defect E was zero, indicating that A-doubling of the upper A'lJ state is negligible: R ( J ) - Q(J + 1) = Q ( J ) - P(J + 1 ) + E 2: 2B,.,(J + 1).(2) The centrifugal stretching constant Dr,, had a negligible effect on the rotational term values for J < 40, and could not be estimated in the present work. Several " extra " lines, originating from transitions to the perturbing state, were observed; some of these are shown in fig. 3. The assignments and vacuum wave- numbers are summarized in table 1.M . A . A . C L Y N E A N D M . C . HEAVEN 219 TABLE VACUUM WAVENUMBERS AND ASSIGNMENTS FOR THE 4-0 BAND OF SnO, A-X"*b 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 31 767.01 66.69 66.20 65.53 64.72 63.76 62.56 61.26 59.78 57.73 55.61 53.18 64.43 61.93 59.56 57.27 54.98 52.65 50.29 47.82 45.31 42.70 40.06 37.31 34.42 31.49 28.42 25.29 22.04 18.70 15.28 11.72 7.99 4.04 700.73 696.71 92.63 88.44 84.I8 79.85 75.42 70.84 31 764.93 64.43 63.76 62.99 62.07 61.01 59.78 58.44 56.92 55.23 53.32 51.27 48.98 46.37 43.55 40.41 50.92 47.71 44.64 41.65 38.65 35.55 32.51 29.31 26.13 22.82 19.44 15.95 12.42 8.76 05.03 701.15 697.21 93.12 88.97 84.71 80.41 75.98 71.41 66.21 62.01 57.33 52.34 47.42 42.07 36.84 31 763.79 62.99 61.01 59.60 57.98 56.39 54.63 52.69 50.64 48.42 45.98 43.41 40.62 37.63 34.31 30.75 26.91 36.75 32.76 28.99 25.29 21.55 17.82 14.01 10.13 6.22 702.21 698.12 93.93 89.64 85.31 80.84 76.33 71.69 66.90 62.01 57.05 51.95 46.57 41.83 36.45 30.93 25.45 19.69 13.28 8.06 Vacuum wavenumbers for the 3-0 band of SnO, A-X, and for the two bands of SnO, b'-X, are Extra lines as follows: R(17), 31 767.22; R(18), 31 750.46; available on request to the authors.Q(18), 31 754.35; Q(19), 31 736.92 P(19), 31 740.97; P(20), 31 722.68.220 LASER SPECTROSCOPY OF ELECTRONIC TRANSITIONS A l~'~lI(l)-XlZ+ B A N D OF SnO In the long-wavelength tail of the 4-0 band of SnO A-A', the laser excitation spec- trum showed a band-head near 3 15.7 nm. No previous report of this band could be found in the literature. A short section of the new band is shown in fig. 4, where the single P, Q and R branches are clearly discernible. Analysis of this band was straight- forward. The rotational numbering was fixed by combination differences formed on the lower state. The rotational constant thus obtained, namely B," = 0.3554& 0.0006 cm-', identifies the lower state as X'C+, u" = 0.I I I I 1 I 1 1 I I I 1 1 2 3 4 5 6 7 8 9 10 11 12 A 0 b ' - X 74-0 b a n d I A A U U I I I I 315.7 315.8 laser wavelength/nm Region near the band-head is A few lines due to high J values of the overlapping 4-0 band FIG. Lt.-Laser-excitation spectrum of 14-0 band of '"SnO (b'-X). shown. Note regular PQR structure. of the A-X system also are shown. Combination differences formed on the upper state from the PR branches were regular and showed no evidence for perturbations, up to the highest J' level examined, namely J < 28. The rotational constant B' for the appropriate A-doublet was deter- mined to be 0.2605 -& 0.0008 cm-', i.e., considerably less than &' for the All2 state. The band origin vo was determined to be 31 666.5 & 0.2 cm-'.Combination relations involving the Q branch [see eqn (2)J revealed the presence of significant A-doubling in the upper state. The data are shown in fig. 5 in terms of a plot of E against J(J + l), where E is defined in eqn (2). Within the scatter of the data, a linear correlation is obtained. Least-squares fitting to the equation ~ / 2 == qJ(J + 1) gave q = B,' - Be' = -0.0012 cm-'. The observation of single P, Q and R branches, and the identification of the lower state as X%+, imply that the upper level is a ll state. This conclusion is further sup- ported by the relatively large magnitude of the A-doubling. SnO has only two low- lying ll states, riz. A'II and 311r. Thus, the upper state must be one of the case (c) components of 317i. 311(2) and 'l-In(O-) components are eliminated due to the selection rules governing radiative transitions.'II(O+) is ruled out, since a 311(O+)-X1E+ bandM . A . A . CLYNE A N D M . C . HEAVEN 22 1 J(J + 1) 0 200 400 I - 5 - ..-b r( I 0. FIG. 5.-A-doubling in the b’ 3n(l) state, v’ = 14. Figure shows a plot of the combination defect e against J(J + 1) [see eqn (2)]. shows two branches, and not three, as observed. Therefore, we conclude that the upper state is V3II( 1), in accordance with the development of three branches per band, and the form of the dependence of E upon J.” THE 3-0 BAND OF SnO ( A - X ) ; AND A PROBABLE b’3n(l)-X1Z+ BAND We have recorded for the first time the rotationally-resolved spectrum of the 3-0 band of SnO A-X. The spectrum showed single P, Q and R branches, as expected, but the structure was extensively perturbed. The shift of the band origin is large (ca.tens of cm-l), in accordance with Connelly’ss suggestion of a massive vibrational perturbation in the v‘ = 3 level. The perturbation therefore is definitely a homo- geneous one. Rotational assignments were made, as before, by forming ground- state combination relations. Because of the perturbed upper AlII state, it was not possible to obtain a reliable value for the unperturbed upper state rotational constant B3’. The analysis of the 3-0 band is discussed below. In addition to the 3-0 band of SnO (A-X), a second band y has been observed in the laser-excitation spectrum. The head of this band lies to shorter wavelengths of the 3-0 band head, and had been reported previously by C~nnelly.~ He assigned y as the 7-2 transition of SnO A-A’.However, this assignment is ruled out by the present work, since we were able to show that the present distribution of SnO X’C+ molecules contained a negligible population in the u” = 2 level. As shown in fig. 8 (see following section on perturbation analysis), the A113(v’ = 3) state and the upper state of the y-band mutually perturb each other. The structure of the y-band suggests that it belongs to the same b”rl(l)-XlZ+ transition as the new X-band seen in the tail of the 4-0 and A-X band. However, assignment of the upper state of the y-band to 3A(l) cannot definitely be excluded. 3C- can be ruled out, since only one perturbation is seen. The weight of evidence (see below) favours assign- ment to Y3rI(l).One problem is that the combination defect E for the band y de- creased as a function of increasing J , which would be consistent with an upper ”n(O+) state, rather than with an upper ’I3( I) state. This apparent conflict of evidence on the nature of the upper state can be resolved, if it is recalled that the upper state is per- turbed and thus may not show a normal magnitude for the A-doubling. A value for B’ of 0.256 & 0.002 cm-‘ was estimated from the upper-state combination differences in band y. Because of the presence of perturbation in y , this value for B’ is not highly accurate. We note, however, that the estimated magnitude of 0.256 cm-’ is close to that determined for the band X. The new band y was regular in form, and showed single P, Q and R lines.222 LASER SPECTROSCOPY OF ELECTRONIC TRANSITIONS THE b ’ 3 n ( i ) STATE A further argument can be given, as follows, in favour of the assignment of the band y as belonging to the b’311(l)-X1C+ transition.The parameters of a Morse function for the b’ state are well-defined. Thus, D, = 19 157 cm-I assuming ground- state products of dissociation, re = 1.992 A and co, = 560 cm-1;6 hence, the Morse parameter p = 1.85 in A-1 units. The anharmonicity constant m,x, is not known, but may be estimated from the expression, m,xe = coe2/4D, = 4.1 ern-,. The result- ing potential-energy curve and vibrational-energy ladder is shown in fig. 1. The band X has an observed excitation energy (band origin) of 31 667 cm-I (see above), which is close to the calculated energy of 31 741 cm-’ for the nearest vibrational level, namelyv’ = 14ofthe b’ state.Thus, the band X is very probably the 14-0 band of SnO (b’-X). If the mexe value is adjusted slightly, from 4.1 cm-I to 4.5 ern-,, the resulting Morse function gives 31 655 cm-l for the energy of the 14-0 band, i.e., very close to the observed value of 31 667 cm-’. The same Morse function gives 31 221 cm-, for the energy of the next lower 13-0 band of b‘-X. This energy for the 13-0 band is very close to the energy of 31 208 cm-, determined for the unperturbed band y. Consequently, it seems likely that the bands X and 7 actually are the 14-0 and 13-0 bands of the W 3 l l ( l)-XIC+ transition. Deutsch and Barrow’s value6 of B,’ = 0.3010 cm-l may be combined with our determination BI4’ = 0.2605 cm-’, in order to give Be’ = 0.3057 cm-I and M, = 0.003 1 cm-I in the expression A value for ae may be estimated from the Pekeris relation, applied to a Morse function, namely This relation gives a, = 0.0028, in satisfactory agreement with the experimental value of 0.0031.One should note the limitations of the Morse function, but the data on the W3II(l) state, as yet, are insufficient to form a reliable RKR function. B,’ = B e - ae(U’ + 3). M, = ~[(CO,X,B,~)*-B~~J/’W,. LIFETIME MEASUREMENTS The lifetimes z of the A’II and lf3rI(O+) excited states of SnO were determined in the Torr pressure range. Table 2 summarizes the results. Initial excitation of states of 120Sn0 was carried out as follows:-(i) near the head (low J ‘ ) of the 1-0 band of SnO (A-X); (ii) in the Q(15) line of the 3-0 band of SnO ( A - X ) ; (iii) in the P(36) line of the 4-0 band of SnO ( A - X ) ; (iv) in the R(12) line of the 14-0 band of SnO (b’-X).The results for the v’ = 1 and u‘ = 3 levels of SnO A’n, and for the U’ = 14 level of SnO b’3rI(l), are extensive enough to indicate a trend for T to decrease as a function of increasing O2 pressure. The data were plotted according to the Stern-Volmer for- mulation, where zo is the lifetime at zero-pressure, and k , is the constant for removal with colli- sion partner M. The signal-to-noise ratio was very high in all cases. (2) z-l - - z0-l + kM[MI, The resulting values for z0 and kM were as follows:- U’ = l(A1l-I):- zo = 160 & 20 ns; kM = (2.2 & 0.6) x lo-’’ cm3 s-’; u’ = 3(A1n):- zo = 140 -& 10 ns; k , = (6.9 & 0.6) x lo-’’ cm3 s-’; u’ = 4(A11-I):- z0 = 130 20 ns; kM < 1.4 x lo-’’ cm3 s-’; u’ = 14[b’3n(l)]:-~o = 580 + 34 ns; kM = (3.6 -& 1.0) X lo-’‘ cm3 s-I.andM . A . A . CLYNE AND M. C . HEAVEN 223 TABLE 2.-LIFETIMES OF A 1 n AND u3n(1) STATES OF 120sno correlation correlation run no. O2 pressure/ coefficient O2 pressure/ coefficient mTorr z/ns of fit run no. mTorr z/ns of fit 32 24 25 23 31 20 22 08 09 10 02 17 03 01 11 50 49 61 60 59 56 125 164 175 182 205 233 267 85 104 127 169 197 209 259 298 143 168 144 169 195 23 1 (i) initial excitation:--A’ll, u‘ = 1, J’ 2: 10 153 0.939 19 289 109 142 0.975 21 297 106 137 0.980 30 327 112 134 0.98 1 29 352 118 131 0.975 28 362 127 129 0.975 27 379 121 108 0.971 26 391 115 114 0.964 12 325 67 107 0.975 13 348 65 99 0.968 14 373 63 98 0.984 15 397 60 79 0.988 16 426 54 76 0.980 05 457 63 88 0,996 06 499 56 68 0.994 07 536 53 133 0.998 48 191 121 131 0.996 526 0.968 55 276 482 53 1 0.983 54 300 474 52 1 0.981 58 387 476 476 0.992 (ii) initial excitation:-A’ll, u‘ = 3, J’ = 15 (iii) initial excitation:-A1lI, u’ = 4, J‘ = 35 (iv) initial ex~itation:-b’~ll(l), u’ = 14, J’ = 13 0.988 0.982 0.976 0.984 0.988 0.991 0.982 0.996 0.994 0.995 0.989 0.993 0.993 0.964 0.993 0.996 0.988 0.989 0.998 The only other work on the lifetime of the SnO states of which we are aware is a pre- liminary study of Capelle and Linton,’ who used a nitrogen laser in broad-band excita- tion of fluorescence from the Sn + N,O reaction.They’ reported two decay times, extrapolated to zero pressure:-160 & 20 ns, with a moderate pressure dependence; and 130 These magnitudes for zo are broadly in agreement with the present work, although it is clear that several transitions were excited simultaneously in the earlier work,’ thus making a detailed interpretation impossible. Because of experimental limitations, the lowest pressure (of 0,) used was 85 mTorr. At this pressure, the mean time between collisions of SnO and 0, is calcu- lated to be of the order of 1 ps at 300 K. Therefore, collisional modification of the A’II state at the lower pressures is unlikely, and the extrapolation to zero pressure should give a reliable value of zo, which is essentially equal to zR, the radiative lifetime of A’n.Evidently, mixing of states zlia perturbations in the absence of collisions does not significantly affect zo, since the values of zo for unperturbed levels (u’ = 1) and per- turbed levels (u’ = 3) are the same, within experimental error. The situation is not so clear cut in the case of the longer-lived b’31-I(l) state (zo = 580 ns), collisions with which are more probable at the experimental pressures of 0, used (table 2). Rotational transfer and/or collisional mixing of states may occur in this case, although there is no evidence that such processes affect the measured life- 60 ns, with a strong pressure dependence.224 LASER SPECTROSCOPY OF ELECTRONIC TRANSITIONS time of SnO W31T(l). Therefore, it is very likely that the value of 580 It: 34 ns is a true estimate of the radiative lifetime of SnO Y311(l).As expected, the value of T~ for the A'IT state is considerably shorter (ca. $) than that for the b'311(l) state, in their radiative transitions to the X'C+ ground state. However, the 580 ns lifetime of the l ~ ' ~ I l ( l ) state of SnO is much shorter than the ca. lo2 ,us values found for the lifetimes of the A 3 n ( l ) states of Br222 and ICl.23 This result indicates that the B 3 n ( l ) state of SnO approaches Hund's coupling case (c) much more closely than the similar states of the halogens. The b'-X transition of SnO thus is an almost fully allowed radiative transition. The considerable variation in the k , values for the various initially-formed states is interesting. It is noted that the maximum value of k , was found for the most per- turbed level (u' = 3) of A'H.This result could indicate that collisional, as well as non-collisional curve-crossing has a maximum probability for the u' = 3 state. Col- lisional transfer froin A'II to the "(2) state would result in a decrease in lifetime, since this triplet state cannot radiate in allowed transitions to the ground X'C+ state. DISCUSSION PERTURBATIONS Perturbations are a frequent occurrence in the spectra of diatomic molecules, such as SnO. At certain ro-vibrational levels (v', J ' ) , the frequencies (and intensities) of rotational lines deviate substantially from the regular series given by the usual term value formulae. The deviation reaches a maximum value as a function of increasing J ' , and then suddenly changes sign, showing a decreasing tendency with a further in- crease in J ' .Sometimes lines showing both positive and negative deviations occur in the neighbourhood of the maximum. One of these two lines is called an " extra " line. Laser-excitation spectra can be expected to show perturbation phenomena some- what similar to those of the corresponding absorption spectrum. One difference is that any line broadening in absorption will be reflected as a weak or absent line in the excitation spectrum, in which fluorescence intensity is recorded. In the present studies of SnO, predissociation, which leads to lifetime shortening of the excited state, is absent; thus broadening phenomena are not exhibited. Clearly, spectral perturbations are to be explained in terms of mixing of states, usually involving the upper electronic state.We summarize briefly those aspects of the analysis of perturbations that are directly relevant to the SnO spectra studied here. More complete treatments are given elsewhere, for example in ref. (21) and (24). Perturbation leads to a mutual repulsion of the electronic energy levels concerned. The selection rule AJ = 0 hold rigorously for all perturbations, and the strength of the interaction is inversely dependent upon the energy separation between the pair of rotational levels that can mutually perturb. A perturbation between two singlet states is illustrated in fig. 6, where total energy is given as a function o f J ( J + 1). In fig. 6, and the following discussion, the superscript* is used to denote quantities relating to the hypothetical unperturbed situation.Thus, the straight lines TF and TZ represent the unperturbed rotational term values of the electronic states 1 and 2. Extrapolation of these lines to J = 0 gives the unperturbed band origins v:,~ and v:,~. Curves Tl and T2 represent the observed rotational term values resulting from the interaction between TT and T;. For any given value of J , the displacement of TI from TF is equal and opposite to the displacement of T2 from TT, i.e., *(TI + T2) = $(T* + T?).M . A . A . CLYNE A N D M . C . HEAVEN 225 Clearly, for T, and T2, the values of the apparent rotational constants, B1 and BZ, are dependent on J. For perturbed states, molecular constants obtained directly from the observed term value series cannot be incorporated into the overall term scheme of the molecule.Consequently, the observed data must be analysed by techniques which yield the un- perturbed constants, since these are fundamentally more significant. Kovacs'l and others have developed an analysis for the recognition of perturba- tions and the determination of unperturbed constants. A more powerful approach has been developed, e.g. by Field et for the analysis of SiO spectra. Their method '0 J(J + 1) hypothetical unperturbed states. FIG. 6.-Perturbation between two singlet states. Superscript* denotes quantities relating to the for the deperturbation of the SnO data is used in a further study;25 in the present work, the simple approach of K o v ~ c s , ~ ' which leads to informative results, has been used as a preliminary analysis. In using this technique, we neglect the small centrifugal-stretch correction, as this is insignficant for the range of J values observed in the present study.Essentially, the method consists of plotting the rotational constant B of the term series against J . Perturbations are revealed in this plot as a deviation from a horizon- tal line. For example, if the term series followed the curve TI of fig. 6, a plot of B against J would asymptotically approach Bf at low J , and Bf at high J. Maximum deviation from a horizontal line occurs at Jo. Another possibility, which also is fre- qently encountered, is that the observed term series will cross from TI to T2 in travers- ng Jo. In this case the plot approaches Bf both at low J and at high J , but large oscillations are observed in the region of Jo.226 LASER SPECTROSCOPY OF ELECTRONIC TRANSITIONS The B value for the level J is obtained from eqn (3A) and (3B), In the absence of A-type doubling, eqn (3A) and (3B) give identical results.When transitions belonging to both the perturbed and perturbing systems are observed, it is possible to determine the unperturbed constants of both states, using the expression (4) 9 +[fX,l(J) +f*,z(J)I = w: + B2*) - B" (4) where x = Q or PR. Norm- ally, one of these constants can be determined by extrapolation of the (B, J ) plot to a horizontal region. An important feature illustrated in fig. 6 is that extrapolation of the term series Tl and T2 to J = 0 does not necessarily give the unperturbed band origins.The apparent origin relevant to a particular J value can be obtained from the expressions, Either B,* or Bi must be known in order to use eqn (4). gdJ) = +[(J + 1)Q(J - 1) - (J - l>Q(J>I =- vo(J) (5A) (J - l)[P(J + 1) + R(J - l ) ] } = v,(J) - B". (5B) Eqn (5A) and (5B) are equivalent to extrapolating the tangent at point J on the term value curve to J = 0. Over the regions of J where T1 is approximately parallel to T:, the g,(J) curves are horizontal, giving the value of the perturbed band origin v ~ , ~ . At higher J values the g,(J) curves smoothly change over to the horizontal lines repre- senting Because of the symmetric nature of perturbations, we may derive eqn VO,l(J) + vo,z(J) = V0.1 + V0,Z = vo*,1 + b?!,2. (6) Eqn (6) can be used to calculate the unperturbed band origin of one of the states, if this quantity is known for the other state involved.PERTURBATION A N A L Y S I S OF T H E 4-0 BAND OF SnO (A-A') Fig. 3 shows part of the perturbed structure of the 4-0 band of SnO ( A - X ) , centred near J' = 18. Fig. 7(a) shows a plot against J of the difference between observed calculated term values, based on Bi* = 0.300 cm-' (see below). Clearly, the per- turbed term series is of the type which is represented by a crossing from TI to T2 on traversing Jo in fig. 6. Plots offa(J) and fpR(J) were constructed [fig. 7(b)] in order to determine the un- perturbed rotational constant for the o' = 4 level. Only at high J-values do the limbs of these curves tend to limiting values, from which may be obtained a value of &* = 0.300 cm-I [fig.7(c)]. Note that the data points for the " extra " lines (fig. 3 and table l), which are also plotted in fig. 7(b), are consistent with the main body of the data. Based on the analysis of the 1-0,O-0 and 0-1 bands, Lagerqvist ef al.' obtained the following expression for the variation of B:* with u':- B;" = 0.31455 - 0.0025 (u' + 3).M . A . A . CLYNE AND M . C . HEAVEN 227 - 0.05- -0.06- 3 I $ -0.07- . n 5 c: -0.08- -0.09- 31770 .+ i.' J J (b) p' 1G --I 1 I I I 20 30 LO 50 J r- 10 I t 1 I I 20 30 40 J FIG. 7.-Perturbation analysis of the 4-0 band of SnO (!-A'). ,!a) shows variation with J of A (observed - calculated) term values: 0 , main lines; 0, extra lines. Note that asymptote at high J tends to zero, i.e., the perturbation tends to zero as J + K .(b) showsfQ(J) andfpR (J) plots [eqn (5A) and (5B)l: 0, ~ Q ( J ) ; 0, ~ P R ( J ) . (c) shows g Q ( J ) and gPR(J) plots [eqn (5A) and (5B)I: 0 , gQ(J); 0, &'PR(J).228 LASER SPECTROSCOPY OF ELECTRONIC TRANSITIONS This expression gives &* = 0.3033 cm-’, which is in reasonably good agreement with the measured value of 0.300 cm-’. It is possible that the measured Bi* has been slightly underestimated, since the f x ( J ) curves may not have quite reached a limiting value at the maximum J-values of 48, Application of eqn (4) gives Bp = 0.230 & 0.005 cm-l for the rotational constant of the perturbing state. Examination of fig. 7(a) indicates the nature of the perturbing state. Thef,(J) plots are strongly curved in the region 4 \< J’ < 18, and extrapolation to J’ = 0 does not give the unperturbed rotational constant.These features indicate a strong perturbation extending down to J ’ = 0. The selection rules for perturbation allow interactions characterized by AA = 0 or j, 1 in case (a), or ASt = 0, 1 in case (c). Homogeneous interactions (i.e., AA = 0 or ASt = 0) occur only through rotational coupling of the states; consequently the strength of such an interaction is proportional to J ( J + 1). As shown in fig. 7, this behaviour is not exhibited by the data for the u‘ = 4 level of SnO A’II. The interaction evidently is homogeneous. Consideration of the electronic structure of SnO shows that the only states of suitable energy for the observed interaction are the components of the low-lying 317i state, namely, 31-1(2), 31-1(1), 31-1(0+) and 311(O-).b3n(0)+ and 311(0-) are eliminated, since these states can only perturb one A-doublet component of A’n, whereas both A-doublets in fact are symmetrically perturbed. 6’ ”(1) is improbable, since the b’-X transition radiates strongly, whereas the perturbing state appears only as a few “ extra ” lines. Therefore, 311(2) is the most probable perturb- ing state in case (a) approximation. This assignment is consistent with the observed weakness of the perturbation in the u’ -1 4 level, as compared with the perturbation in the u’ = 3 level. However, since there is a known tendency to case (c) coupling in the excited states of SnO, an equally probable perturbing state for the u’ = 4 level is the 3A(1) state. As the perturbation affects lines of low J , the band origin must be determined from the high J region of the g,(J) curves [cf.eqn (5A) and (5B)l. Fig. 7(c) shows the g&) curve, which tends to a value of v : , ~ = 31 767.5 0.5 cm-’. Utilizing this value, and applying eqn (6) to the data gives the band origin of the perturbing state v&; we obtain I?;S,~ = 31 778.1 0.7 cm-’. VIBRATIONAL CONSTANTS OF THE A’n STATE Evaluation of the unperturbed origin for the 4-0 band makes possible a more accurate determination of the vibrational constants for the A’IT state of SnO. Lager- qvist et u I . , ~ having data for 2,’ = 0 and 1 only, obtained rough estimates for LI),’ of 580 cm-*, and coe’.xe’ of 3 cm-’. We now combine the u‘ = 0, 1 data with our u’ = 4 result, in order to give: we’ = 587.6 cm-’; we’xe’ = 2.5 cm-’.The value of T, - G”(0) is 28 803.8 cm-I, which, combined with G”(0) = 821.5 [ref. (9)] gives T, = 29 625.3 cm-’. PERTURBATION A N A L Y S I S OF THE 3-0 BAND OF SnO (A-X) Whereas the 4-0 band shows essentially only transitions involving the perturbed state, the 3-0 band shows two sets of transitions that mutually perturb each other. Analysis of the g ( J ) and g*(J) curves (fig. 8) shows that the two states interact strongly. A mean value for [g(J) -1- g*(J)]/2 of 31 208.2 & 0.3 cm-I was obtained [eqn (6)], The perturbation in the 3-0 band differs from that in the 4-0 band.M . A . A . CLYNE AND M. C . HEAVEN 229 This result may be used to determine the band origin of the perturbing state, before interaction, vf, if the unperturbed band origin ~ f , ~ of the A'll state is known.A value for v3 ,o = 31 208.9 cm-' is readily obtained by calculation from the vibrational con- stants derived in the previous section. Hence the value of vp" before interaction is 31 207.5 cm-'. It is noted that this value is very close to that for the unperturbed 7 -0.07 h -0.09-/ r 8"- B; --- -0.10 1 I I I 0 10 20 30 40 J r( I i 31210 i+ 31190 t I I I 1 0 10 20 30 40 J FIG. S.-Perturbation analysis of the 3-0 band of SnO ( A - X ) and of the perturbing 13-0 band of SnO (6'-A'). (a) showsfQ(J) plots (m) for the A'n statefQ,A(J), and (@)for the b'W(l) statef+(J). Note symmetrical nature of the plots and limits for rotational constant differences.[BL is rotational constant for A'II(v' = 3); Bp is rotational constant for V3n(l) (u' = 13).] (6) shows gQ(J) plots for the A'II and l~'~n(l) states; notation as in (a). band origin for the 3-0 band, in agreement with the strong interaction that is observed. Measured (perturbed) band origins are v ~ , ~ = 31 233.4 0.2 cm-I and v p = 31 182.7 0.4 cm-'. We may consider the rotational constants of the states involved. Analysis of the B' value for the A'll (v' = 3) state is not reliable, because thef(J) andf*(J)curves do not reaching limiting values. However, the data yield a value for B" - (B'z + B ' 3 / 2 - - -0.074 & 0.001 cm-l, where &* and &* are the rotational constants for the A'II and for the perturbing state. Lagerqvist's constants were used to determine Bj* = Thus, the energy of interaction between the states is ca.25 cm-'.230 LASER SPECTROSCOPY OF ELECTRONIC TRANSITIONS 0.3058 cm-'. Substitution then gives BL* = 0.256 &- 0.002 cm-', a magnitude which is close to that observed for the v' = 14 level of the Y311(l) state. Evidence has been given above that the perturbing state in the 3-0 band is the v' = 13 level of the b'311(l) state or possibly a level of the 3A1 state. EXCITED STATES OF S N O A good deal of information now is available on the excited states of the Group IV oxide radicals CO, SiO, GeO and SnO. A common characteristic is the nesting of numerous excited states with the principal A'II state; these nested states include 311, 3A; 3C+ and 'A, 'Z-, 3C-, all of which correlate with the same atomic configurations.The details of the perturbations vary considerably from molecule to molecule. Thus, for example, the major perturbations in SO2" and SnO are dissimilar. These differences are not surprising, since the positions of the curve-crossings and hence the Franck-Condon factors for the relevant perturbations, are sensitively dependent on small differences in Y, values between the states. In addition, there is a trend towards case (c) coupling with increasing molecular mass, thus favouring triplet-singlet per- turbations for the heavier members of the series CO, SiO, GeO and SnO. No doubt other perturbations in SnO will be observed, when vibrational levels of SnO A'II, other than u' = 1, 3 and 4, are investigated by high-resolution laser spectroscopy, The identification of states should be considerably aided by using life- time measurements as an additional diagnostic, as in the present study for the A'Il and F3n(l) states.We thank Arthur Fontijn for helpful discussion, and Richard Barrow for very use- ful correspondence. We gratefully acknowledge support of this work by the Royal Society, the S.R.C. and the U.S. Air Force Office of Scientific Research (Grant AFOSR-78-3 507). See: G. Herzberg, Spectra and Structure of Simple Free Radicals (Cornell University Press, Ithaca, 1971). M. C. Lin and J. R. McDonald, in Reactive Intermediates in the Gas Phase, ed. D. W. Setser (Academic Press, New York, 1979). M. A. A. Clyne and I. S. McDermid, Adv. Chem. Phys., to be published. P. C. Mahanti, 2. Phys., 1931, 68, 114; also P. C. Mahanti and A. K. Sen Gupta, 2. Phys., 1938, 109, 39. F. C. Connelly, Proc. Phys. SOC., London, 1933, 45, 780. E. W. Deutsch and R. F. Barrow, Nature (London), 1964, 201, 815. ' A. Lagerqvist, N. E. L. Nilsson and K. Wigartz, Arkiv Fysik, 1959, 15, 521. a J. J. Smith and B. Meyer, J. Mol. Spectrosc., 1968, 27, 304. lo W. Felder and A. Fontijn, Chem. Phys. Lett., 1975, 34, 398. l1 W. Felder and A. Fontijn, J. Chem. Phys., 1978, 69, 1112. l2 D. M. Manos and A. Fontijn, J. Chem. Phys., 1980,72, 416. l3 M. A. A. Clyne and M. C. Heaven, Chem. Phys., 1980, 51, 299. l4 M. A. A. Clyne and I. S. McDermid, J. Chem. SOC. Faraday Trans. 2, 1978,74, 1376. l5 M. A. A. Clyne and M. C. Heaven, J. Chem. SOC., Faruday Trans. 2, 1978,74, 1992. l6 K. P. Huber and G. Herzberg, Molecular Spectra and Molecular Structure. l7 S. G. Tilford and J. D. Simmons, J. Phys. Chem. Ref. Data, 1972, 1, 147. la For references, see: M. A. A. Clyne and M. C. Heaven, J. Chem. Soc. Furaduy Trans. 2, 198 I , l9 J. A. Coxon, Molecular Spectroscopy (Spec. Period. Rep., The Chemical Society, London, 2o J. A. Coxon, J. Mol. Spectrosc., 1974, 50, 142. G. A. Capelle and C. Linton, J. Chem. Phys., 1976, 65, 5361. Part B.-Constants of Diatomic Molecules (Van Nostrand-Reinhold, New York, 1979). 77, 1375. 1973), vol. 1, p. 177. I. KovAcs, Rotational Structure in the Spectra of Diatomic MoIecules (A. Hilger Ltd., London, 1969).M. A . A . CLYNE A N D M . C. HEAVEN 23 1 22 M. A. A. Clyne, M. C. Heaven and E. Martinez, J. Chem. Soc., Farachy Trans. 2, 1980,76, 177. 23 S. J. Harris, W. C. Natzle and C. B. Moore, J. Chem. Phys., 1979, 70, 4215. 24 R. W. Field, A. Lagerqvist and I. Renhorn, Phys. SCY., 1976, 14, 298. 25 R. F. Barrow, M. A. A. Clyne and M. C . Heaven, to be published. 26 T. Torring, Z . Naturforsch., Teil A , 22, 1234.

ISSN:0301-7249    

DOI:10.1039/DC9817100213

出版商:RSC    

年代:1981

数据来源: RSC