chemistry Formal Report FOR A-PLUS WRITER

http://www.jce.divched.org/Journal/Issues/2007/Feb/abs336.html

In the Laboratory

www.JCE.DivCHED.org • Vol. 84 No. 2 February 2007 • Journal of Chemical Education 337

more extensive characterization of the I2 absorption in this region has been given by Forkey et al. (10), who have listed all the detectable lines in absorption from υ� = 0, 2, and 3 in a ∼2-cm–1 interval.2 In fact, I have succeeded in photo- graphing spectra that are clearly attributable to these three υ� levels. Results are illustrated in Figure 1. The confirma- tion of excitation from the two excited levels (υ� = 2 and 3) is the observation of two and three anti-Stokes lines, respec- tively (anti-Stokes meaning lines falling at wavelengths shorter than that of the exciting line).

Spectra of green laser pointers confirm that the output varies progressively rather than randomly with time. It seems likely that this behavior relates to thermal changes in com- ponents of the optical cavity or power supply as the laser warms up under power. As one indication of this, if the laser is allowed to operate for several minutes and is then turned off and on again quickly, it appears to cycle rapidly through its previous tuning cycle, producing two or three fluorescence “flashes” before it stabilizes again near its performance when turned off. In one particularly dramatic instance of such be- havior, through the spectroscope I observed in repeated fash- ion spectra like those illustrated in Figures 1B and 1C, with a cycle time of ∼2 s.3

The logic for high spectral purity is probabilistic in na- ture. The I2 B ← X absorption spectrum is so dense in this region that excitation at random wavelengths will produce strong fluorescence about half the time. For example, in early attempts to measure the continuous absorption between the lines of the B ← X system using atomic line sources, discrete absorption was evident for all but three of a dozen Ar atomic lines (11). In a later, more extensive study, almost all mea- surements had to be corrected for discrete absorption (9). When the present fluorescence “turns off ”, the beam can com- pletely disappear in a dark room; from observations of the strong fluorescence through neutral density filters, this means a drop in fluorescence intensity by a factor of at least 104. If the GLP emitted light in broad lines or in too many different modes, such “off ” behavior would be improbably rare.

Green laser pointers operate on a doubled transition of Nd doped in a host crystal. The latter is often YVO4 (yttrium

orthovanadate) instead of YAG (yttrium aluminum garnet, Y3Al5O12), because YVO4 is said to better couple optically with the 808-nm radiation from the laser diode pump (12). The IR laser output at 1064 nm is doubled in a crystal of KTP (potassium titanyl phosphate, KTiOPO4), which is contained within the laser optical cavity, where it is exposed to the high optical powers needed to achieve frequency doubling. The IR output is blocked with a filter, leaving only 532-nm radia- tion in the output beam.4 The lasers are said to operate in the single transverse optical mode TEM00. Longitudinal modes occur at frequency intervals of c �2d, where c is the speed of light and d is the length of the optical cavity (13). For the typical cavity lengths of ∼5 mm (12), this translates into a wavenumber spacing of ∼1 cm�1 in the IR fundamental, 2 cm�1

in the doubled visible output. Since the Nd gain profile is only several cm�1 wide, it is reasonable for the GLP to oper- ate on only one or two modes, as required probabilistically for the observed “off ” behavior of the fluorescence.

Green laser pointers may operate either pulsed or CW (continuous wave). Pulsed operation can be demonstrated by sweeping the beam across a wall rapidly, giving a “dashed- line” display.

Spectra of Green Laser Pointers

Spectra were recorded at high dispersion (0.04 nm�mm) for three of the pointers (numbers 1, 2, and 4), revealing a surprising range of behaviors, with each displaying a spec- trum clearly distinct from those of the others. (Lasers are iden- tified in the Materials section.) Spectra were recorded with short exposures (0.3–1 s), slits almost closed (∼2 µm), and neutral density filters to further attenuate the light so as not to saturate the CCD array detector. For all three GLPs, some spectra were recorded in time sequence (< 1 min apart). Typi- cal results are illustrated in Figures 2 and 3.

Figure 2 shows a behavior more complex than antici- pated. The strong modes are spaced ∼2.7 cm�1 apart, imply- ing a cavity length of 3.6 mm. However, there are other peaks, usually weak, that do not fit this pattern. The extra peaks do not appear to be due to different transverse modes; these

Figure 1. Calibration and I2 fluorescence spectra for excitation with a green laser pointer: (A) reference spectra from Hg [bright green (546.1 nm) and yellow (577.0 and 579 nm)], Ne (red, from He– Ne laser at 632.8 nm), and from the green laser pointer (532 nm). Spectra (B–D) are I2 fluorescence spectra recorded at different times. (B) and (C) are dominated by fluorescence excited mainly from υ� = 0 and 2, respectively; (D) includes a detectable component ex- cited from υ� = 3 (note “blue” line at top).

Figure 2. Representative spectra for GLP 4, as recorded for differ- ent “on” times and battery powers. These spectra were calibrated with Ne lines at 533.0778, 534.1094, and 534.3283 nm (stan- dard air), giving 532.05(1) nm or 18790.0(4) cm�1 for the zero point on the wavenumber scale.

http://pubs.acs.org/action/showImage?doi=10.1021/ed084p336&iName=master.img-000.jpg&w=215&h=120

http://pubs.acs.org/action/showImage?doi=10.1021/ed084p336&iName=master.img-001.jpg&w=234&h=131

In the Laboratory

338 Journal of Chemical Education • Vol. 84 No. 2 February 2007 • www.JCE.DivCHED.org

would be expected to vary spatially in the laser spot, but no differences were observed in spectra obtained by directing different regions of the spot onto the spectrometer slit. Time sequence spectra for this GLP showed a smooth progression of changes, with modes drifting to lower wavenumber ac- companied by slow intensity changes. With fresh batteries, the longtime pattern was typically one strong mode and ei- ther one somewhat weaker mode (∼50% of main) or two much weaker modes (each ∼5% of the main mode) on ei- ther side.

Figure 3 shows the much different output of the other two GLPs. GLP 1 displayed multiple peaks displaced con- siderably to the blue (spectrum C) when operated with weak batteries. On the other hand, with fresh batteries, only one of the many recorded spectra (D) showed more than one peak, and that an order of magnitude weaker than the main one. It is hard to interpret the peak spacing in spectra C and D in terms of longitudinal modes. Similarly, the mode pat- tern in the spectra of GLP 2 (A and B) is not simple; and in this case the pattern is similar for weak and fresh batteries and shows many peaks of significant intensity. Not surpris- ingly, the fluorescence beam excited by GLP 2 never weak- ened to the vanishing point in a dark room.

It is noteworthy that GLP 1 operated CW while the other two were pulsed. To check on the possibility that the single-mode character was tied to this property, I recorded spectra for another CW GLP (not used in the fluorescence experiments) but found its spectrum more like that of GLP 4 in Figure 2.

LIF in the Gas Phase: What Else Is Required?

The green-laser-induced fluorescence dramatically illus- trates the prime condition necessary for LIF: the radiation must be absorbed by the fluorescing species. That this con- dition is necessary but not sufficient is nicely illustrated by directing the same laser beam through a cell containing a sig- nificant pressure of another gas in addition to the I2, as shown

in Figure 4. I2 in its B electronic state is known to be de- stroyed by collisional predissociation (14),

I + I + I2I2(B) + I2 (1)

and

I + I + Ar I2(B) + Ar (2)

These processes represent radiationless decay of the B state, so the fluorescence is said to be quenched. Actually, the beam can be perceived weakly in a high-pressure cell like that shown in Figure 4, since the quenching is not 100% efficient. And interestingly, this weak beam seems less subject to the tem- poral variations of the exciting laser. This behavior, which has not yet been quantified, could be a consequence of pres- sure broadening of the I2 absorption lines, leading to a re- duced specificity in the excitation process, due to more extensive overlap of the absorption lines.

So I2 visible fluorescence requires absorption into the B state at low pressures, in the absence of quenching gases. Is that sufficient? The answer is “No,” as is illustrated by Fig- ure 5, which shows several beams from an Ar-ion laser oper- ated multiline, directed through the same low-pressure I2 cell. Only the longest wavelength line, at 514.5 nm, excites fluo- rescence (spectrum in Figure 6). With a power meter, one can confirm that the other wavelengths are absorbed (9). However, all but the “green” line excite I2 above the B-state photodissociation limit, resulting in direct dissociation (hence no fluorescence),

I + I*I2(X) + hν (3)

where I represents (as before) the ground state (2P3/2) of the atom and I* the spin-orbit excited state (2P1/2).

Figure 6 includes the spectra produced by exciting from υ� = 0 at two different wavelengths, from which it is clear that the wavelength range of the fluorescence is determined firstly by the excitation wavelength of the source.6 This seems

Figure 3. Spectra recorded for GLP 2 (A and B) and 1 (C–E). The wavenumber calibration is the same as in Figure 2.

Figure 4. A GLP beam traverses (right to left) two cells containing I2 vapor at room temperature. The cell on the right contains only I2, while that on the left has ∼500 torr Ar in addition. The laser was positioned ∼7 m from the cell to give a less focused beam.

http://pubs.acs.org/action/showImage?doi=10.1021/ed084p336&iName=master.img-002.jpg&w=229&h=123

http://pubs.acs.org/action/showImage?doi=10.1021/ed084p336&iName=master.img-003.jpg&w=191&h=143

In the Laboratory

www.JCE.DivCHED.org • Vol. 84 No. 2 February 2007 • Journal of Chemical Education 339

at odds with comments made in a recent discussion of fluo- rescence and scattering in this Journal (15), in which the fluo- rescence wavelength was stated to be independent of exciting wavelength. The difference is that the present gas-phase fluo- rescence spectra are obtained under essentially collision-free conditions in the gas, while the fluorescence spectra discussed in ref 15 were observed in solution, where internal relaxation occurs in the excited electronic state prior to emission. In that sense the present fluorescence spectra are more like the resonance Raman spectra described in ref 15; in fact workers with a primary background in Raman spectroscopy have of- ten referred to such resonance fluorescence spectra in exactly that way, as resonance Raman spectra.7

The intensity patterns in Figures 1 and 6 reflect the vary- ing Franck–Condon factors (FCFs) for the transitions, since the intensities are proportional to the FCFs (6). Figure 7 shows these FCFs for υ� = 43 and 32, the latter being the lowest υ� level excited at 532 nm (10). Note especially the strong alternation in FCFs for the latter in the region υ� = 7–15—a behavior that is clearly evident in the spectra of Fig- ures 1B and 6B.

Other Spectroscopic Processes in Gaseous I2 The phenomena discussed above all involve the B–X

transition in I2, which occurs conveniently in the visible spec- tral region. However, I2 has a rich emission–fluorescence spec- trum at shorter wavelengths, involving a number of other electronic transitions, some of which can be illustrated easily as a part of a demonstration of LIF in the B–X system. Thus, for example, if a Tesla coil is discharged in contact with the two cells shown in Figure 4, the resulting optical emissions are radically different (see the Supplemental MaterialW): The low-pressure cell gives a weak, yellowish-white glow, which on spectral inspection can be seen to display many of the same B–X bands that are observed in absorption (16). On the other hand, the high-pressure cell now displays an in- tense blue emission. Both cells show strong emission in the UV, but again with radically different spectra. The low-pres- sure spectrum is dominated by an extensive series of diffuse bands known historically as the McLennan bands (17, 18), while the high-pressure cell shows several discrete band sys- tems (19). Similar (but weaker) spectra can be excited by a low-pressure mercury discharge lamp like that used to ob- tain the calibration spectra in Figures 1 and 6.

The states responsible for the optical emissions in the UV are of ion-pair character and lie at much higher energies than the valence B state (20). Their behavior in the presence of inert gases (rare gases, N2) is in stark contrast with that of the B state. The low-lying ion-pair states are practically im- mune to quenching, permitting the excited molecules to re- lax into near-thermal υ and J population distributions prior to emission. The resulting fluorescence–emission spectra are relatively insensitive to excitation process, making this case more like that described for fluorescence in (15). This be- havior—immunity to collisional quenching—is also the hall- mark of the rare-gas halide (RgX) excimer lasers, which operate on similar charge-transfer electronic transitions. In- deed, the strongest transition in the high-pressure I2 emis- sion spectrum, known as D� → A�, was long ago made to lase in experiments like those which led to the development of the RgX lasers.

Figure 5. Several beams from an Ar-ion laser operated multiline transit the low-P I2 cell. The beams are dispersed by a prism lo- cated ∼25 cm to the right of the cell, the window of which is viewed from the inside, looking through the cylindrical wall. Fluorescence is excited by only the beam at 514.5 nm. The spots on the en- trance window to the left of the fluorescence beam mark the points of entry for beams at 501.7 nm, 496.5 nm, 488.0 nm, and 476.5 nm, from right to left. The lower spots are from reflected beams, and the faint green and blue lines are “ghost” images.

Figure 7. Franck–Condon factors for the lowest υ� level excited at 532 nm and for the υ� = 43 level excited at 514.5 nm.

Figure 6. I2 fluorescence spectra excited by the Ar-ion laser at 514.5 nm (A) and, for comparison, spectrum (B) from Figure 1. Reference spectra are from Hg and (in A) from the laser, at wavelengths noted in Figure 3 (except missing the last).

http://www.jce.divched.org/Journal/Issues/2007/Feb/abs336.html

http://pubs.acs.org/action/showImage?doi=10.1021/ed084p336&iName=master.img-004.png&w=215&h=138

http://pubs.acs.org/action/showImage?doi=10.1021/ed084p336&iName=master.img-005.jpg&w=191&h=143

http://pubs.acs.org/action/showImage?doi=10.1021/ed084p336&iName=master.img-006.jpg&w=232&h=188

In the Laboratory

340 Journal of Chemical Education • Vol. 84 No. 2 February 2007 • www.JCE.DivCHED.org

Observing and Photographing the Fluorescence Spectra

The reader may have noticed that the fluorescence beam in the low-pressure cell in Figure 4 appears as an elongated cone, with intensity clearly diminishing with distance from the entrance window. The laser light is absorbed so strongly that it is significantly attenuated by the time it exits the cell. For absorption near the centers of lines originating in υ� = 0, this attenuation is much greater than the ∼25% one might predict from the average molar absorptivity of ∼800 L mol�1

cm�1 (9), for a 10-cm cell at P = 0.3 torr (the approximate vapor pressure of I2 at room temperature). On the other hand, for observing the fluorescence spectrum, it is good to use a disperse laser beam (as in Figure 4) rather than a focused one, and then to view the fluorescence near the entrance window, where it is most intense. Lasers that operate single mode with high spectral purity can interact with only small velocity groups of molecules, so a focused beam can effectively de- plete the population of absorbers as it “burns” its way through the cell. The result is reduced absorption and fluorescence (and violation of Beer’s Law, which assumes constant con- centration of the absorbing species). In essence, this phenom- enon is the basis of one method of sub-Doppler spectroscopy, known as hole-burning. The gross attenuation of the beam seen in Figure 4 is also a dramatic illustration of the origins of nonlinear behavior in analytical fluorimetry: when the fluo- rescence intensity is observed some distance from the entrance window, it will first rise with increasing pressure of absorber, then decline as the absorption attenuates the exciting light before it reaches the observation region.

Spectra like those shown in Figures 1 and 6 are readily observed “live” using a hand spectroscope; however, record- ing them required much trial-and-error work with a digital camera. This task provides a golden opportunity to get ac- quainted with the more subtle options in the camera’s menu. A tripod is necessary, since exposures of 1–8 s may be needed. It is helpful to use a bright source to achieve proper align- ment of the source, spectroscope, and camera. The fluores- cence excited by the Ar-ion laser can be bright enough for this purpose (depending on the laser power), but that pro- duced by the green laser pointer probably will not be. I used a mercury Pen-Ray lamp, the strong green and yellow lines from which could be seen easily on the camera’s LCD moni- tor at lowest magnification. I could then zoom in on these lines, adjusting the alignment to maintain good images. Manual focus is needed for this operation; about 60 cm was the optimal stated distance in my case, though the lens of the camera was actually only about 1 cm away from the eye- piece of the spectroscope. Manual settings for both the aper- ture and shutter speed are also required, and I found it necessary to operate at largest aperture (F2.8) to obtain good image quality. For the longer exposures it is also necessary to invoke the “noise reduction” option. Alternatively one may use the “night scene” setting, which automatically employs noise reduction. Finally, I had to tweak the white-balance to render the yellow Hg doublet as yellow; an enhancement of the red component sufficed. However, I was never able to record a full spectrum with the nuanced differences that are obvious to the eye, for example, in the several laser lines to shorter wavelength from the green line in the Ar-ion laser output. In part, this is likely a problem of color/intensity in-

terdependence, which the eye can handle much better than the camera.

Materials

I observed the on–off behavior for all four of the laser pointers I tried: (1) from Sean and Stephan Corp. (Taiwan); (2) Model GP4 from Limate Corp. (Taiwan); (3) AltasNova (Mead, WA and Taiwan); and (4) Beta Electronics (Colum- bus, OH). All of these pointers are stated to be less than 5 mW, but some evidently push this limit harder than others, judging from differences in the observed fluorescence inten- sities. Pointer 1 appeared to be best overall. Pointers 1 and 3 appeared to operate CW, while pointers 2 and 4 were pulsed, with a much faster repetition rate for pointer 2. For pointer 1 the stated wavelength was 532 ± 10 nm, while that for pointer 2 was 500–550 nm! However, as shown above, all operate near the stated doubled Nd wavelength of 532 nm (12).

The spectra illustrated in Figures 2 and 3 were obtained using a Jobin–Yvon 1.5-m spectrometer equipped with a 3600-groove�mm holographic grating and a CCD array (Photometrics CCD9000, 27-mm pixel width) as detector. The reciprocal dispersion in this region was obtained by re- cording several Ne lines (noted in Figure 2) from a discharge source.

The spectroscope contains a prism as dispersing element and has an adjustable slit, which was set as narrow as pos- sible for recording the spectra. When focused, with narrow slits, this instrument easily resolves the Hg yellow lines, which are separated by 2.0 nm. The specific manufacturer’s infor- mation has been lost, but a very similar spectroscope is sold by Edmund Industrial Optics for less than $300. Cenco (Sargent–Welch) offers a prism model with adjustable slit and focus for less than $60, but the resolution is not stated.

An Olympus Camedia model C-4000 (3× optical, ∼10× total zoom) was used to record the spectra. Procedures and settings have already been described, and similar procedures should work for other digital camera models. The mercury discharge source used for alignment and for reference–cali- bration was a Pen Ray lamp (Ultra-Violet Products, Inc.). The Ar-ion laser was a model Innova 300 from Coherent (15 W maximum power all lines). The Tesla coil was a model BD-20 from Electro-Technic Products. The cells were silica and were charged with I2 (and Ar, for the high-P cell) on a vacuum line and sealed off with a torch.

Hazards

It is never good to stare directly into any laser, includ- ing the Class IIIa devices that are approved for laser point- ers. However, numerous studies have shown that momentary exposure to such light is no sight hazard, and certainly not at the distances involved in purported cases of airplane “spot- ting” that made news not long ago (21). Ar-ion lasers, on the other hand, can operate at powers that represent great sight hazard, so any use of the 514.5 nm line from such a laser should be carefully planned so that observers are not exposed to either direct or reflected beam light. The Hg Pen- Ray lamp produces strong UV light at 253.7 nm, which can damage the cornea of the eye. Both glass and plastic safety glasses block this radiation, but side exposure should be avoided, also. Operation of the source at a distance of 1 m

In the Laboratory

www.JCE.DivCHED.org • Vol. 84 No. 2 February 2007 • Journal of Chemical Education 341

or more for only short times (1–2 min) should result in mini- mal exposure, with or without safety glasses.

The Tesla coil output is very high voltage (up to 50 kV) but also very high frequency, so the spark is not dangerous (22). Instructors can “wow” students by boldly taking a ∼0.5- inch spark from the tip, though this is more comfortably done if the spark is directed to a coin held between thumb and forefinger.

Both the Tesla coil and the Hg discharge lamp produce ozone at detectable levels (by smell), so should be left on only while needed.

Summary

A green laser pointer can be used in a “flashy” demon- stration of laser-induced fluorescence in the gas phase, by di- recting the beam of the laser through a cell containing I2 at its room temperature vapor pressure. The demonstration is a good one to provoke discussion, and the explanation of the on–off behavior provides valuable insight into the require- ments for LIF and the properties of lasers. On the other hand, if this source is to be used to record fluorescence spectra for quantitative analysis, it will be necessary to use an array-type detector and to be aware that the time-integrated spectrum may contain contributions from numerous (υ�, J�) levels.

Under the assumption that it is changing cavity tem- perature that is responsible for the varying wavelength of the laser, it is conceivable that one can modify such a pointer to permit control of its temperature and hence its wavelength. I am looking into this possibility.8

WSupplemental Material

Color versions of Figures 1, 4, 5, and 6 and four QuickTime movies illustrating the temporal behavior of the fluorescence and of the discharges obtained by bringing the low- and high-pressure cells into contact with an operating Tesla coil are available in this issue of JCE Online.

Acknowledgment

I thank Tim Hanusa for providing two of the green la- ser pointers and Mike Bowers and Laura Swafford for help setting up and operating the Ar-ion laser.

Notes

1. This behavior is so astounding to observers that it readily elicits hypothetical explanations. In my experience on demonstrat- ing the phenomenon to a number of students and faculty, few of these are close to reality.

2. This wavelength region corresponds to a Franck–Condon “hole” in the absorption from υ� = 1, so no lines from this level are listed.

3. After about two minutes of operation, Pointer 1 produced a particularly intense spectrum that was dominated by excitation from υ� = 0. Then in the next minute it slowly drifted off this ex- citation (or excitations) and onto one involving predominantly υ� = 2. When the laser was clicked off and on again, it cycled back to the latter spectrum in about two seconds, displaying the former spectrum for about 1 second in the process.

4. Reference 12 includes many very instructive photographs and diagrams of green laser pointers and their components, in ad- dition to the diagram cited in the reference.

5. There is also a weaker continuous absorption in this same spectral region, designated 1Πu ← X, which yields two ground-state I atoms via photodissociation (9, 11). However, this transition never yields fluorescence and is in no way sensitive to the B-state photo- dissociation limit.

6. It is sheer coincidence that the 532-nm wavelength so nearly coincides with the third line in the fluorescence spectrum excited at 514.5 nm: the energy difference in the two wavelengths just happens to match closely the energy difference between υ� levels 0 and 3 in the X state. This is perhaps unfortunate, as it suggests a “symmetry” in nature that is not there.

7. The weak fluorescence that persists at high pressures shows less dependence on exciting wavelength, as the surviving B-state molecules have mostly relaxed into low υ� levels. This situation is thus closer to that described for fluorescence in ref 15.

8. Additional spectra recorded at lower dispersion show that GLP 1, which appears to operate single mode in Figure 3, actually operates on one or two additional modes spaced 2 nm (70 cm�1) apart most of the time. Such broad spacing is hard to explain from expectations for both the gain profile and the cavity length.

Literature Cited

1. Shoemaker, D. P.; Garland, C. W.; Nibler, J. W. Experiments in Physical Chemistry, 6th ed.; McGraw–Hill: New York, 1996; pp 425–434.

2. Stafford, F. E. J. Chem. Educ. 1962, 39, 626–629. 3. Steinfeld, J. I. J. Chem. Educ. 1965, 42, 85–87. 4. Hollenberg, J. L. J. Chem. Educ. 1970, 47, 2–8. 5. D’alterio, R.; Mattson, R.; Harris, R. J. Chem. Educ. 1974,

51, 282–284. 6. Tellinghuisen, J. J. Chem. Educ. 1981, 58, 438–441. 7. Muenter, J. S. J. Chem. Educ. 1996, 73, 576–580. 8. Garland, C. W.; Nibler, J. W.; Shoemaker, D. P. Experiments

in Physical Chemistry, 7th ed.; McGraw–Hill: New York, 2003; pp 423–432.

9. Tellinghuisen, J. J. Chem. Phys. 1982, 76, 4736–4744. 10. Forkey, J. N.; Lempert, W. R.; Miles, R. B. Appl. Opt. 1997,

36, 6729–6738. 11. Tellinghuisen, J. J. Chem. Phys. 1973, 58, 2821–2834. 12. Goldwasser, S. M. Sam’s Laser FAQ. http://repairfaq.ece.

drexel.edu/sam/laserssl.htm#sslafgl. See especially http://repairfaq. ece.drexel.edu/sam/dpss1.gif for a diagram of the optical cavity (both accessed Nov 2006).

13. Svelto, O. Principles of Lasers; Plenum: New York, 1976. 14. Masiello, T.; Vulpanovici, N.; Nibler, J. W. J. Chem. Educ.

2003, 80, 914–917. 15. Clarke, R. J.; Oprysa, A. J. Chem. Educ. 2004, 81, 705–707. 16. Singh, S. M.; Tellinghuisen, J. J. Mol. Spectrosc. 1973, 47, 409–419. 17. McLennan, J. C. Proc. Roy. Soc. 1913, A88, 289–291. 18. Tellinghuisen, J. Chem. Phys. Lett. 1974, 29, 359–363. 19. Guy, A. L.; Viswanathan, K. S.; Sur, A.; Tellinghuisen, J. Chem.

Phys. Lett. 1980, 73, 582–588. 20. Mulliken, R. S. J. Chem. Phys. 1971, 55, 288-309. 21. The Register. http://www.theregister.co.uk/2005/01/06/

laser_man_letters/ (accessed Nov 2006). 22. McKeever, M. R.; Sur, A.; Hui, A. K.; Tellinghuisen, J. Rev.

Sci. Instrum. 1979, 50, 1136–1140.