analysis can be performed at this wavelength.
LITERATURE CITED (1) Chem. f n g . News, 53 (30),19-21 (1975). (2) S. D. Faust and H. M. Gomaa, fnvlron. Lett., 3 (3), 171-201 (1972). (3) M. Eto, “Organophosphorous Insecticides, Organic and Biological Chemistry”, CRC Press, Cleveland, Ohio, 1974, p 241. (4) S. D. Faust and H. M. Gomaa, Environ. Lett., 3 (3) 171-201 (1972). (5) H. A. Moye, J . Chromatogr. Sci., 13, 266-279 (1975). (6) D. H. Rodgers, Am. Lab., 9 (2), 133-138 (1977). (7) C. M. Sparacino and J. W. Gines, J. Chromatogr. Sci., 14, 546-556 (1976).
(8) (9) (10) (11) (12)
G. A. Junk et al., J. Chromatogr., 99, 745-762 (1974). A. K. Burnham et al., Anal. Chem., 44, 139-142 (1972). Chem. Eng. News, 54 (IS), 35-36 (1976). C. G. Creed, Res./Dev., 27 (9),40-44 (1976). Eugene Mossbacker, extension advisor for agrlculture, McLean County Cooperative Extenslon Service, Private Comrnunicatlon.
RECEIVED for review April 11, 1977. Accepted June 23,1977. Work supported in part by an institutional grant from Illinois State University (No. 75-32).
Laser Two-Photon Excited Fluorescence Detection for High Pressure Liquid Chromatography Mlchael J. Sepaniak and Edward S. Yeung” Ames Laboratory-ERDA and Department of Chemistry, Iowa State University, Ames, Iowa 500 11
A laser two-photon excited fluorometric detector for high pressure liquid chromatography Is described and characterized for the separation of the oxadiazoies PPD, PBD, and BBD. Excitation is provlded by the absorption of two photons of radiatlon at 5145 A from an argon ion laser. The detectlon limits, linearity of response, precision, and selectivity are reported and are found to compare favorably with other UV detection methods.
While the fluorometric high pressure liquid chromatography (HPLC) detector is not as commonly used as the UV absorbance detector (1, 2 ) ) it does possess some definite advantages, namely higher sensitivity for those compounds with an appreciable fluorescence quantum efficiency and greater selectivity since relatively few of the molecules that absorb UV radiation actually fluorescence. Selectivity is also enhanced by the fact that fluorescent molecules have both an excitation and emission spectrum that can be scanned ( 3 , 4 ) . The present paper describes a fluorometric detection method that has two unique features. First, excitation is provided by an argon ion laser capable of 4 W of radiation at 5145 A. Second, the excitation process is the result of the absorption of two photons of the 5145-A light. In 1931Maria Goppert-Mayer realized that a molecule could absorb two photons simultaneously to achieve a change in its quantum level (5). The process involves some distinctive selection rules and represents a way for spectroscopists to find and describe new molecular states (6). The value of the two-photon process in fluorometric HPLC detection lies in its improved selectivity. The fact that two-photon absorption involves different selection rules than one-photon absorption results in different one-photon and two-photon absorption spectra, and this produces an additional variant in the selective detection of fluorescent molecules. Two-photon fluorescence detection is somewhat limited by the small size of the twophoton absorption strength (6) and it is only with the high output power of a laser that measurable fluorescence signals can be obtained. The two-photon absorption strength is defined by the relationship
AP = P’CLA-‘ 6 1554
ANALYTICAL CHEMISTRY, VOL. 49, NO. 11, SEPTEMBER 1977
where AP is the absorbed optical power, P is the optical power, C is the solute concentration, L is the path length, and A is the optical beam cross-sectional area (6). Typical values for 6 are 110-4scm4 s photon-I molecule-’. The fraction of the absorbed optical power that is actually detected as fluorescence ( R ) can be calculated from Equation 2
R
= Qkq
where Q is the solute fluorescence quantum efficiency, k is the optical collection efficiency, and q is the detector quantum efficiency. This report characterizes a laser two-photon excited fluorometric (LTPEF) detector used in the two-photon detection of PPD, PBD, and BBD. These oxadiazoles have the general structure N-N
R-c,
II
II
0
,C-R
,
where R and R’ are either phenyl or biphenyl groups. A UV absorbance detector is used for the comparison of detection limits, linearity of response, and selectivity.
EXPERIMENTAL Chromatographic System. The liquid chromatographic system was composed of a LDC, Riviera Beach, Fla., minipump capable of delivering 16-160 mL/h of eluent at pressures up to 5000 psi, Rheodyne, Berkeley, Calif., injection valve with a 100-pL sample loop, and a Waters Associates, Milford, Mass., p-Bondapak CIS column (30 cm long X 3.9 mm id.). The eluent used for all separations was 60/40 UV grade tetrahydrofuran/water. The oxadiazoles were from Pfaltz and Bauer, Inc. Separations were all at ambient temperature with a flow rate of 2.0 mL/min and an injection volume of 100 pL. UV Absorbance Detector. The UV detector used for comparison purposes was a Chromatronixs,Berkeley, Calif., Model 230 mixed wavelength detector. The detector was operated at 280 nm where background noise was smallest and oxadiazole absorptivities greatest. LTPEF Detector. The fluorometric detector (see Figure 1) is composed of a light-tight cubic metal box containing a 1-mm i.d. X 3-mm 0.d. quartz flow cell. The 5145-A laser radiation of a Control Laser model 553 argon ion laser passes through a 0.5-inch aperture, then two Corning 3-71 sharp cut-off filters, and one Corning 4-96 wide bandpass filter, before being focused on the center of the flow cell by a 50-mm focal length X 25-mm diameter
Table I. Detection Limitsa UV absorbance LTPEF detector, ng detector, ng PPD 13 150 PBD 10 9 BBD 76 31 a Detection limit taken as amount of solute that gives SIN = 3 where S is peak height and N is peak to peak background noise. a stagnant solution of M PPD in the flow cell, then adjusting the optics (the lens holders shown in Figure 1 provide slight adjustment of the optics) to give a photon count of 20000 counts/s at a laser power of 1 W. The focusing lens adjustment was critical but, once set, the system remained stable until other experimental operations required moving the laser. Primarv Filters
I
Outlet
Figure 1. Schematic diagram of laser two-photon excited fluorometric detector. The Ortec photon counting system is outlined by the dashed line 100
z 0
10
E
w
2 4
a
I-
1.0
I-
B W
n
0.1
i
BBD
I
5--I1 I PED I L1---L
Excit tion waveqength
I
PPD
7,
IAr'plasma line re&?
5145;
11°C
I
3000
4000
5000
6000
7000
(81 Figure 2. Spectral considerations for the fluorometricdetector. Upper graph: Transmission curves for primary filters (3-71 and 4-96) and secondary filters (7-51 and 7-54). Lower graph: Fluorescence emission regions for PPD, PBD, and BBD. The super radiance of the argon ion laser contains Ar' plasma line emission in the region from 3500 to 5145 WAVELENGTH
A
quartz focusing lens. The fluorescence emanating from the flow cell is collected and collimated by a 38-mm diameter quartz lens with a f-number of 1. The collimated fluorescence is passed through three UV bandpass filters (two Corning 7-51 filters and one Corning 7-54 filter), then onto the photocathode of an Amperex, Hicksville, L.I., N.Y., 56-DVP photomultiplier tube. The purpose of the Corning 3-71s is to eliminate laser plasma lines present in the laser super radiance (7), and the Corning 4-96 is used to block orange fluorescence, from the Corning 3-71s, which is passed by the secondary filters. This and other spectral considerations are illustrated in Figure 2. Signal Processing. The fluorescence signal was counted in 0.5-s intervals with an Ortec, Oak Ridge, Tenn., photon counting system (see Figure 1). A latching circuit and Ortec D/A converter were used to convert the digital signal to an analog readout on an Omini-scribe stripchart recorder. The D/A converter time constant was set at 0.1 s, since it was found that longer time constants decreased detector precision. Detector Optimization. Prior to chromatographicoperation, the detector optics were optimized and standardized by placing
RESULTS AND DISCUSSION Detection Limits. The concentration necessary to detect a particular solute with a given response can be calculated using Equations 1 and 2. For the present detection system, R is approximately L, the diameter of the flow cell, is 0.1 cm; and A , the cross-sectional area of the focused laser beam, is roughly cm2. Using these quantities the concentration necessary to register a photon count of 100 counts/s, at a power of 4 watts (5145 A) for a solute with an absorption strength of cm4 s photon-l molecule-1, is 1.5 X M. As seen from this calculation, sub parts-per-million concentrations can, in principle, be detected even for solutes with moderate two-photon absorption strengths, providing high laser output power can be obtained a t the proper absorbing frequency. In the previous calculation a relatively small photon count of 100 counts/s was chosen as an easily detectable signal. This is mainly due to the ease with which optical filters can be used to reduce background signals. The laser light is better separated, spectrally, from the fluorescence (see Figure 2) than in normal one-photon excitation fluorescence (8). The favorable detection limits of the LTPEF detector are shown in Table I where detection limits for two of three oxadiazoles are better for the LTPEF detector. The peakto-peak background noise, and therefore the detection limits, of the W absorbance detector could be, in principle, improved by effective eluent pulse dampening, but the improvement is limited by detector drift and fluctuations other than the regular pumping noise of the system. And, in fact, the elimination of pumping noise, solvent peaks, detector drift, and problems with gradient elution are some of the distinct advantages of the LTPEF detector. There are a few LTPEF detector improvements that should in theory improve sensitivity. The first is to replace the primary filters by a good quality dispersive prism, which would remove plasma lines from the laser super radiance with approximately 40% less beam attenuation than with the fiiters. The second is to use a focusing lens with a shorter focal length. The theoretical limit of the cross-sectional area of the focused laser beam ( A in Equation 1)is dependent on the focal length of the focusing lens (9). With a shorter focal length lens, higher power densities, and therefore larger two-photon absorption, could be obtained. The quadratic dependence of two-photon signals on laser power is shown in Equation 1. A third improvement would therefore be increasing the laser output power by means of a mode locker. The reader is referred to Ref. 7, p 199, for a good description of mode locking. When mode locked, the continuous output of the argon ion laser is converted to a series of high energy pulses. Sensitivity could then be greatly enhanced by properly gating the detection sytem so that signals are counted only during and directly after a laser pulse. Even without gating the ANALYTICAL CHEMISTRY, VOL. 49, NO. 11, SEPTEMBER 1977
1555
I
5 MIN
4>
k 5 MIN 4
PBD I
PPD
I
a PBD
1
k 5
ib
MINI
t-5
MIN
1
Flgure 3. Chromatograms for laser two-photon excitation fluorometric detection of PPD, PBD, and BBD, 100-pLinjection volume, sensltivlty 100 counts per inch. (a) 2.4 X lo-’ M PPD, 1.5 X lo-’ M PBD, 4.0 X lo-* M BBD; (b) same oxadlarole concentratlons but with approximately IO-‘ M phenol, fluorene, chrysene, and anthracene
Flgure 4. Chromatograms for UV absorbance detection of PPD, PED, and BBD, 100-KL Injection volume, sensltivity 0.008 absorbance unit per Inch. (a) 4.0 X lo-’ M PPD, 7.0 X lo-’ PBD, 1.0 X lo-’ M BED; (b) same oxadiazole concentrations but wlth approximately M phenol, fluorene, chrysene, and anthracene
detection system, the mode locker could theoretically produce approximately a two-fold increase in signal. Linearity of Response. Calibration plots for 100-pL injections of 3.0 X lo-’ M to 1.0 X M PBD were drawn for both detectors. The inner-cell effect oftentimes causes nonlinearity in fluorometricdetection (IO),but the small active cell-length (about 1 mm) and large incident radiation of the LTPEF detector result in a linear calibration curve for PBD, even at solubility limited concentrations. The linear regression constants for the calibration plots were 0.984 for the UV absorbance detector and 0.987 for the LTPEF detector, showing equally good linearity. Precision. The precision of the detection system was evaluated by making five injections of the three oxadiazoles at the concentrations listed in Figure 3. The peak heights were measured and the BBD peaks were used as internal standards. This was done so that only inconsistencies resulting from detector response would be considered. The relative standard deviation for the five PPD and PBD peaks were 6.7% and 10.1%, respectively. The LTPEF detector reproducibility compares favorably with that reported by Perchalski, Winefordner, and Wilder for a fluorometricHPLC detector (II), but is not as good as the reproducibility of a fluorometric detector reported by Cassidy and Frei (12). Inconsistencies in LTPEF detector response can be attributed, in part, to the nonlinear dependence of two-photon signals on laser power density, which makes beam and power stability critical to reproducibility. Better reproducibility could be attained if fluorescence signals were normalized to laser power output using Equation 1. Selectivity. The selective detection of the oxadiazoles PPD, PBD, and BBD is illustrated in Figure 3. Chromatograms 3a and 3b are essentially identical despite the fact that the sample injected in Figure 3b contains several Polyaromatic Hydrocarbons (PAH) at concentrations of approximately lo5 M. The same separations in Figure 4 show appreciable interference effects for the UV absorbance detector. In Figure 4b the PBD peak is almost totally obscured by the PAH and the PPD peak is unresolved as well. The
PAH were chosen so that the BBD peak was not interfered and could be used as an internal standard. At this point the selectivity of the LTPEF detector is limited by the availability of laser output frequencies. The present detector employs an argon ion laser which has several visible plasma lines between 4579 and 5145 A. To take full advantage of the uniqueness of two-photon spectra, lasing action must be attained at two-photon frequencies covering most of the near UV region of the spectrum. This could possibly be done with high output power tunable dye lasers. While two-photon absorption strengths are very small, molecular two-photon states are as common as one-photon states. This means that LTPEF detection can be applied to most fluorescent compounds provided a laser can be found with the proper output frequency and power. Of course many compounds would require laser output frequenciesand powers which are not available at this point but, with the steady advancement of laser technology and two-photon absorption research, more applications of LTPEF detection are possible.
1556
ANALYTICAL CHEMISTRY, VOL. 49, NO. 11, SEPTEMBER 1977
LITERATURE CITED (1) J. N. Done, J. H. Knox, and J. Loheac, “Applications of High Speed Liquld Chromatography”, J. Wiley and Sons, New York, N.Y., 1974, p IO. (2) M. Krejcl and N. J. Posplsllova, J . Chromafogr., 73, 105 (1972). (3) S. 0. Perry, R. A m s , and P. I. kewer, “Practical Uquld Chromatography”, Plenum Press, New York-London, 1972, p 194. (4) E. D. Pelllzzarl and C. M. Sparachlno, Anal. Chem., 45, 378 (1973). (5) M. Goppert-Mayer, Ann. fhys., 9, 273 (1931). (6) W. M. McClaln, Acc. Chem. Res., 7 , 199 (1974). (7) 6. A. Lergyel, “Lasers”, J. Wiley and Sons, New Yak, N.Y., 1971, Chapter 9. (8) G. G. Guilbauk, “Ractlcal Fluorescence”, Marcel Dekker, New Yo&, N.Y., 1973, p 138. (9) A. E. Siegman, “An Introduction to Lasers and Masers”, McGraw-HIII Book Co., New York, N.Y., 1971, p 317. (IO) D. R. Baker, R. C. Williams, and J. C. Steichen, J . Chromatogr. Scl., 12, 505 (1974). (11) R. J. Perchalskl, J. D. Winefordner, and B. J. Wilder, Anal. Chem., 47, 1993 (1975). (12) M. Cassidy and R. W. Frel, J . Chromafogr., 72, 293 (1972).
RECEIVED for review March 30,1977. Accepted June 6,1977. Work performed for the U.S. Energy Research and Development Administration under Contract No. W-7405-eng-82.
