Phase Resolved Phosphorimetry J. J. Mousa' and J. D. Winefordner2 Department of Chemistry, University of Florida, Gainesville, Fla. 3267 7
The frequency and phase characteristics of several phosphorescent species have been studied with emphasis upon the change in phase and amplitude with frequency of modulation and lifetime of the phosphorescence. A phase sensitive detector has been employed to study the possibilities of phase resolution as an analytical tool in phosphorimetry. It was shown that phosphorescence emission and excitation spectra of molecules which show severe overlap can be phase resolved into the spectra of the individual components. Fluorescence emission was also phase resolved from phosphorescence emission. These resolutions were demonstrated with dilute solutions of pure compounds and synthetic binary mixtures. Quantitative analysis of synthetic binary mixtures by phase resolution was found to be accurate and precise.
The first reported use of the phosphorescence of organic molecules as a means of chemical analysis was in a study by Keirs, Britt, and Wentworth in 1957 ( I ) . Since this work appeared, other workers have developed theories and experimental systems which provided the basis for the exploitation of the phosphorescence lifetime as a method for the resolution and analysis of mixtures of organic phosphors (2-5). The advantages of a pulsed source, time resolved system for phosphorimetry were discussed by Winefordner ( 6 ) , and the subsequent experimental work by Fisher and Winefordner (7) demonstrated the actual application of these ideas to the resolution and analysis of synthetic binary mixtures of phosphorescent species. This same technique, with improvements in instrumentation, was employed by O'Donnell, Harbaugh, and Winefordner for the determination of phosphorescence lifetimes (8) and for the resolution and analysis of mixtures of halogenated biphenyls ( 9 ) . Time resolved phosphorimetry is a stroboscopic technique which uses a repetitive, short duration, high intensity flash of exciting radiation, coupled with sampling of the luminescence signal during selected time intervals after termination of the exciting light. Gated detectors or signal averagers are used to monitor either portions of or the entire decay curve. Time resolved determination of fluorescent species is a more difficult experimental technique because of the extremely short fluorescence lifeP r e s e n t address,
Del.
E. I.
D u P o n t d e N e m o u r s & Co., W i l m i n g t o n ,
Author t o whom r e p r i n t requests s h o u l d be sent. (1) R. J. Keirs, R. D. Britt, Jr., and W. E. Wentworth. Anal. Chem.. 29, 202 (1957). (2) P A. St. John and J. D. Winefordner, Anal. Chem., 39,500 (1967). (3) H . C. Hollifield and J. D . Winefordner, Chem. lnstrum., 1 , 341 (1969). (4) T. C . O'Haver and J. D. Winefordner, Anal. Chem., 38,602 (1966). (5) T. C. O'Haver and J. D. Winefordner, Anal. Chem., 38, 1258 (1966). (6) J. D. Winefordner. Accounts Chem. Res., 2 , 361 (1969). (7) R P. Fisher and J. D . Winefordner, Anal. Chem., 44, 948 (1972). (8) C. M. O'Donnell. K. F. Harbaugh, and J. D. Winefordner, Anal. Chem.. 45, 381 (1973). (9) C. M. O'Donnell, K. F. Harbaugh, and J. D. Winefordner. Anal. Chem.. 45, 609 (1973).
times (order of nanoseconds). The experimental difficulties associated with high intensity, short duration sources and fast response detectors had to be surmounted so that this technique could be used successfully. Advances in technology and instrumentation have overcome most of these problems in recent years. Previously, fluorescence lifetime measurements were made with phase or modulation fluorometers. Both phase and modulation fluorometers employed a continuously operated (cw) excitation source whose intensity is modulated sinusoidally a t a high frequency (in the MHz range). The result is the excitation of a sinusoidally varying luminescence from the sample, The amplitude and phase of the luminescence signal is a function of the frequency of modulation and the lifetime of the luminescence. By measuring the phase shift or degree of modulation of the signal, fluorescence lifetimes can be calculated. A review by Birks and Munro ( I O ) covers the historical development of a phase, modulation, and time resolved fluorometers describing some of the experimental systems employed and giving the theoretical basis for each technique. Phase or modulation fluorometers have been used primarily to study single exponential decay processes when lifetime measurements were desired. The analysis of complex decays by the phase fluorometer was complex and difficult. Birks and Munro (10) give reference to several fluorometers which operated a t several different frequencies, but it is pointed out that none of these were used to study complex decays. The possibility of obtaining the Fourier transform of the exponential decay process by phase measurements a t different frequencies was also mentioned. Schmillen (11) performed fluorescence decay time measurements on hydrocarbon crystals and obtained evidence for several exponential decay processes occurring in anthracene, This evidence was obtained by performing phase measurements and Fourier analysis of the decay processes. In 1968, Doi and Toshinai (12) published a theoretical paper which involved the evaluation of transient phenomena, such as fast decaying phosphor luminescences, by their frequency characteristics. Applying a Fourier transformation to exponential build-up and decay processes, they derived expressions which were equivalent to the equations describing the phase and amplitude behavior of luminescent species under sinusoidal excitation as in a phase fluorometer. In a later paper ( I 3 ) , these same authors applied their theory to the determination of the amplitude transfer characteristics of inorganic, red-emitting phosphors. The parameter which they labeled as the amplitude transfer characteristic corresponded to the parameter called the degree o f modulation in the theory of phase and modulation fluorometers. Although in their theoretical paper, these workers predicted a phase shift parameter which they called the phase transfcr character(10) J. B. Birks and I. H. Munro, in "Progress in Reaction Kinetics." Vol. 4. G. Porter, Ed., Pergamon Press Ltd., London, 1967. (11) A. Schmillen, in "Luminescence of Organic and Inorganic Materials,'' H. P. Kallmann and G. M . Spruch, Ed., John Wiley and Sons, Inc., New York, N.Y.. 1962. (12) K. Doi and A. Toshinai, Jap. J. Appl. Phys., 7 , 1504 (1968). (13) K. Doi and A. Toshinai, Appl. Opt., 9, 2762 (1970).
A N A L Y T I C A L C H E M I S T R Y . VOL. 46, NO. 9, A U G U S T 1974
1195
istic, no attempt a t measuring this parameter was made in their experimental work. This phase parameter corresponded to the phase shift parameter in the theory of the phase fluorometer. Experimentally, their work involved the excitation of phosphors with a sinusoidally modulated electron beam, although they stated that the technique and principles involved could just as well be applied to optical excitation of luminescent species. The phenomenon of the phase shift of fluorescence has been used in several cases to study the kinetics of fluorescence quenching (14, 15). In a later paper, Veslova, Cherkosov, and Shirokov (16) indicated the possible resolution and recording of individual fluorescent spectra from each of two luminescent centers with overlapping emission spectra by means of a modulated light source and a phase sensitive detector. The use of a lock-in amplifier to determine the phase shift of luminescence during a chemical reaction by the null or quadrature technique has also been reported (17). The use of phase phenomena and frequency behavior has also been applied in the field of gas chromatography by Reilley et al. (18). These workers described the behavior of output signals when a periodic sampling function is used at the input of a gas chromatographic column. A similar application of the phase characteristics of periodically pulsed columns was given by Obst (19). Using the different phase and amplitude relationships of luminescence signals from species with different lifetimes, the purpose of this project was the investigation and evaluation of these relationships in analytical phosphorimetry for the resolution and analysis of mixtures of organic molecules. This new analytical method will be termed phase resolved luminescence spectrometry. The objective of the investigation was not the construction of the optimum experimental system, but rather the demonstration of the principles involved, showing the potentialities and weaknesses of this new technique as a method of resolution in phosphorimetry. The molecules studied were selected on the basis of the strength of their native phosphorescence signals and because their lifetimes were known via time resolved determinations. THEORETICAL CONSIDERATIONS Theory of Phase a n d Frequency Characteristics of Luminescence. The following derivation of the equations describing the phase and frequency behavior of luminescence will follow the procedure of Birks and Munro ( I O ) in their development of the equations for phase and modulation fluorometry. The use of periodic exciting light will be assumed, but this condition is in no way necessary to obtain the final equations describing the phase and frequency characteristics of the luminescence, as they can be independently obtained via Fourier transformation of the time dependency of luminescence law as described by Doi and Toshinai (12). The phosphorescence intensity as a function of time, Zp(t), will be a function of the time dependence of the exciting light, Z o ( t ) , and the phosphorescence signal decay behavior, ip(t ) . Although only phosphorescence will be discussed here, the expression below should also apply generally for fluorescent as well as phosphorescent species. The phosphorescence intensity can be expressed by a convolution integral as in Equation 1, (141 W. R . Ware. J . Amer. Chem. SOC.,83, 4374 (1961) (15) W. R. Ware, J. Phys. Chem., 66, 455 (1962). (16) T. V. Veslova, A. S. Cherkosov, and V. I. Shirokov, Opt. Spectrosc., 29, 617 (1970). (17) L. F. Phillips, Rev. S o . Instrum., 42, 1078 (1971) (18) C . N. Reilley. G. P. Hildebrand, and J. W. Ashley, Jr., Anal. Chem., 34, 1198 (1962). (19) D. Obst, J. Chromatogr., 32, 8 (1968)
1196
In this expression, Zp(t) is the phosphorescence signal intensity a t any time t and ip(t’) is the value of the exponential decay function at a time t’ relative to the start of the decay. l o ( t - t’) is the value of the exciting light function at the start of the decay process. The product of the exciting light function and the decay function is integrated over the interval from t’ = 0 to t’ = m . Assuming a periodic character for the exciting radiation, IO(t ) can be expressed as a Fourier series,
where ko is an abritrary constant and w j , k,, and 6, are, respectively, the angular frequency, relative amplitude, and phase of the j t h component. Taking the phosphorescence lifetime as T P , the decay behavior of the phosphorescence signal can be expressed as:
ip(t) = ipoexp(-t/~,)
(3)
where ipo is the phosphorescence signal at t = 0 and T P is the phosphorescence lifetime. Now, substituting for Z o ( t t’) and ip(t’), Equation 1becomes,
which reduces to
Equation 5 consists of a Fourier series similar to Z o ( t ) in which the phases and amplitudes of the various components depend upon those of the same frequency components in Z o ( t ) . Equations 2 and 5 are similar except for the constant multiplicative term i p o ~ p ,and the function (1 i w , ~ ~ ) - lThis . second function affects the amplitude of the various components of Zp(t) and introduces a phase and the shift with respect to Zo(t). The vector amplitude ‘2, phase 8, of the j t h component in Zp(t) relative to the equivalent quantities in Zo(t) are given by
+
-
m,
= (1
+
i~,7~)-’
( 6)
and
e, = tan-’(w,7,) The absolute value of the relative amplitude, by m, = (1 W , ~ ~ ~ ~ ) - - I / ~ or m, = (1 47r2fJ27p2)-1/2
+
+
(7)
z,, is given (8) (9)
where f, is the linear frequency, in Hz, and is equal to wJ/27r. If the exciting radiation is assumed to have a sinusoidal character, then the subscript j may be dropped because only the terms where j = 1must be considered. The parameter, m,, which shall be called the degree of modulation in this paper is essentially a measure of the amplitude of the ac component of the exciting radiation. It is seen that the value of this parameter depends upon the frequency of modulation and the lifetime of the phosphorescence. In Figure 1, a graphical illustration is given of the variation of m, with modulation frequency and
A N A L Y T I C A L C H E M I S T R Y , VOL. 46, N O . 9, AUGUST 1974
‘1
2
3
5
IO
20 33
50
100
200 300 500
1000
FREQUENCY ( H z l
Figure 1. Semi-logarithmic plot of theoretical variation of degree of modulation, m, with the frequency of’modulation phosphorescence lifetimes in the millisecond range. At low modulation frequencies. the value of mj approaches unity, which corresponds to the maximum signal level, and gradually approaches a limiting value of zero as the frequency is increased. If the exciting radiation is modulated a t a sufficiently high frequency, short-lived phosphors have a relatively greater luminescence signal than long-lived phosphors (see Figure 1). This behavior can be used advantageously when short-lived phosphors are to be determined in the presence of long-lived background interferences. The phase shift angle, 0,, is also a function of frequency and lifetime as illustrated in Figure 2. The phase angle is measured with respect to the phase angle of the exciting light and as Figure 2 shows, varies from 0 to n / 2 radians (0 to 90”). At low frequencies, the luminescence signal is essentially in phase with the exciting light and as the frequency increases, the luminescence signal becomes nearly 90” out of phase. At a constant frequency setting, the phase shift angle, O,, will vary according to the phosphorescence lifetime of the molecule. Phase Resolved Phosphorimetry. In phosphorimetric analyses, it is frequently found that the emission and excitation spectra of similar molecules overlap severely. Thus, an analysis of one of the components in a mixture of similar species could be subject to severe interference. However, in many cases, the phosphorescence lifetimes for these same molecules are quite different. Therefore, analytical techniques which exploit the lifetime of the phosphorescence for the resolution of the signals from similar phosphors should be quite useful. In the preceding section, it was shown that a molecule which is capable of luminescence, when excited with light of a periodic intensity, will emit a luminescence with a periodic variation in intensity. The amplitude and phase of this luminescence with respect to these same parameters in the excitation light will be functions of the frequency of modulation and the lifetime of the luminescence. This statement applies, of course, to all the different types of luminescence which a molecule is capable of emitting. Thus, for a molecule which displays fluorescence as well as phosphorescence, a periodic signal from both of these phenomena would be observed. In standard phosphorimetry, where a rotating can phosphoroscope is used, no fluorescence emission is observed. and only the phosphorescence is measured. In time resolved phosphorimetry, the fluorescence is eliminated by gating the detector to observe the signal after a delay time, during which the fluorescence has decayed completely. In phase resolved phosphorimetry, however, this is not the case as no phos-
FREQUENCY 1 nz )
Figure 2. Theoretical variation of phase shift angle, (I,, wlth the frequency of modulation phoroscope is used, and the fluorescence emission must also be considered. The exciting light function can be expressed as the sum of a constant intensity term and a sinusoidally varying intensity term,
+ I,”
I ” ( t )= I,’
cos ut
(10)
The amplitude of this function is IO” and its phase has a value of zero. The expression for the emitted luminescence, according to Equations 5 , 6, and 7 , will have the same form as Equation 10 and can be represented as
I,
=
hLI”/
+ mLhI,13” cos (ut - 01,)
(11)
where h l , is a constant factor taking into account the quantum efficiency and concentration factors, m I , is the degree of modulation with respect to IO, and 01. is the phase shift angle between luminescence and excitation. The total intensity from the sample, I r , will be made up of the phosphorescence intensity, I p , the fluorescence intensity, I F ,and a scattered or stray light component I,.
I,
=I
,
+ I, + I,
(12)
These individual intensities upon excitation with a sinusoidal excitation radiation are given for low concentrations of analyte by
I,
= hpIO/
I F
= kFI0’
I,
=
and
hsI(
+ m p h p l ( ‘ cos (ut - Oil) + mFhFI,” cos (ut - 8,)
(13)
+ msh,I,”
(15)
cos (ut - Os)
(14)
where k~ = (2.3)Yptbc and k.17 = (2.3)YFcbc.Here, Yp and YE. are the relative quantum efficiencies for phosphorescence and fluorescence, respectively. the t is the molar absorptivity of the analyte absorber, b is the absorption path length, and c is the molar concentration of the absorber. The parameter, hs, is that fraction of the excitation radiation detected as stray light or scatter. The degree of modulation of the fluorescence and phosphorescence is given by m F and m P ,respectively, and the relative phase shift by OF and Op. The scattered light will also have a degree of modulation, m8, and a phase shift, 05, which should be essentially that of the exciting light. For the purposes of the following discussion, it will be assumed that the stray light is negligible, and the total luminescence intensity will be given by
IT= hJ,/
+ mphpI,,” cos (ut - 8,) + hFI; + mFhFIO”
COS
(ut - 8,)
A N A L Y T I C A L C H E M I S T R Y , VOL. 4 6 , NO. 9 , A U G U S T 1 9 7 4
(16) 1197
and, rearranging,
I,
= (kpZo'
+ kFZ() + m p h p I ~ 'X cos (at - OP) + m F k F I i 'cos (at - 8,)
(17)
Thus, the total luminescence intensity will be given by a constant dc component, (IzPIo' + k&'), and an ac component due to the sum of the cosine functions for the fluorescence and phosphorescence. Depending upon the frequency of modulation, the parameters m p and O p , and mF and O F can be widely different. In the frequency range to be covered in this work ( 2 to 1000 Hz), the value of mF is equal to unity, and the phase shift parameter is zero; these fluorescence parameters do not begin to vary until the MHz frequency range is approached. Thus, under the conditions of low frequency operation, the fluorescence ac term will be insensitive to variations in frequency and will have the same phase as the exciting light. In phase resolved phosphorimetry, a frequency and phase selective detection system is employed; thus only the ac terms of the luminescence intensity will be observed. Equation 17 thus reduces to
where $p1 and 4 p 2 are the relative phases' of phosphorescent components 1 and 2 , respectively, with respect to the reference signal, and E p ( 1 ) and E p ( Z i are the average input signals from components 1 and 2 . The phase resolution of this signal into its two components can be accomplished in two ways. One of these methods shall be called the phase method. In this technique, if the phase of the reference signal 4~ is made equal to 90" f o r 2, then the signal from one of the two components may be "nulled out" and only the signal from the other component mea then sured. For example, if 4~ = 90" 4~1,
+
E,
=
Y IE , y l,
COS
(90"
+
@PI
--@ P I ) +
cos (90"+ 4 P 1 -
Ep12,
Ep
=
Y E P i 2 , cos (90"
+
=
Y E P r J i sill
-
or, if 4~ = 90"
E,
=
(@pz
@p1
-
@P?)
(23
+ ~ P Zthen ,
YIEIWl COS (90"
+
$1'2
-
+
$Pi)
cos (90"+
Ep(21
+
With the proper choice of excitation and emission wavelengths and by using molecules which are strongly phosphorescent, the fluorescence terms in Equation 14 can be made negligible and only the phosphorescence observed. In many cases, this selectivity cannot be achieved, and the fluorescence must still be considered. For the sake of simplicity, it will be assumed that for the molecules used in this study, the above condition holds, and only the phosphorescence is measured. Thus, Equation 18 can be rewritten as
I,
= m p l p n cos (ut
-
8p)
(19)
where I p o = kplo". In this work, a lock-in amplifier detection system is employed so that only ac signals having the same frequency as the reference signal are selectively demodulated and measured. The phase of the internal reference signal can be adjusted so as to give maximum response to a signal having the same frequency and phase characteristics. The output dc signal is related to input ac signal by the following relationship
@P2)1
E,, = Y E , , , , cos (90" @ p i E ~=, y E p i l sin , - @P2)
@IJ)
-
@P?)]
@PI)
(24)
It is observed that the output signal becomes proportional to the sine of the phase angle difference between the two phosphors. Therefore, the best results are obtained when the difference in phase angles is the greatest so that the sine term is maximized. It should be noted that one of the sine terms can be negative; this will give rise to a negative output signal on the lock-in amplifier. This presents no problem as a 180" shift of the reference phase can be used to give a positive output signal. The phase technique is utilized a t a fixed frequency, and therefore the magnitudes of the signals, Ep(1, and Eplzi, will be affected by their respective modulation parameter, m,, a t that particular frequency. The other resolution method shall be called the frequency method. In this method, the phase angle, 8 , is related to the frequency of modulation by $ = tan-1 U T . Therefore, if we set the reference phase angle a t a con, will be a certain frequency a t which stant value, 4 ~ there 4~ = 90" f 4 ~ 1 and , another frequency a t which 4~ = 90" f 4p2. If, for example, the frequency is such that & = 90" 4F11, then the same situation exists as before in the phase resolution technique, i.e.,
+
where is the average input voltage, y is an amplification factor and ( 6 -~d i i n , ) is the phase difference between the reference and input signals. In normal operation, 4~ is made equal to c$lln,, and the maximum signal is obtained. If 4~ = 90" + $ i l l , ) , then the cosine term goes to zero, and no output signal is observed. This condition is called the null or quadrature condition. Consider now a phosphorescence signal Er.lll,,from a solution containing one phosphorescent species applied to the input of a lock-in amplifier. The output signal will be given by
where 4 p is the instrumentally measured phase angle of the phosphorescence signal with respect to the reference signal. It is directly related to B P , which is the phase shift with respect to the excitation light. Consider now the situation when a signal from a binary mixture is considered
1198
=
?EI,,?,sin (@p2 - @ p l )
at w1
(25)
at
(26)
and vice versa if dR = 90" + 62, then
E ~ =, YE,,,,sin
(@PI
-
@pi)
The same sine relationship holds as before, but, in this case, the measurements are made a t two different frequencies and so the magnitudes of E , , l l , and Ep(2, are also affected. This effect upon the degree of modulation with a change in frequency can place a limit on the utility of this method, if the frequencies are such that the amplitudes are greatly reduced. An expression for an ac signal composed of a phosphorescent and fluorescent emission is given by
This equation describes the output signal from the detector when a single species which fluoresces and phosphoresces is present. The phase resolution technique can be applied in this situation to resolve overlapping or interfer-
A N A L Y T I C A L C H E M I S T R Y , VOL. 46, N O . 9, AUGUST 1974
P
-
Figure 4. Figure 3 .
C, = 4000 M F D . 40 W V D C
Block diagram of instrumental system
a, = 2 ~ 5 8 8 1
A = 0-40V, 0-30A DC power s u p p l y (Harrison Model 6268A) B = starter circuit C = 0-100V. 0-0.2A DC power supply (Harrison Model 6116A) D = summing operational amplifier and current booster (Heath EUW-19)
a2 = 2 ~ 5 8 8 5 = 2N1479 R 1 = l a dz 5%. 240 W
Model
modulation circuit excitation and emission monochromators G = photomultiplier tube and housing H = high voltage power s u p p l y I = load resistors J = differential amplifier (Tektronix Model 1A7A) K = amplifier (optional) (PAR Model 211) L = lock-in amplifier M = strip-chart recorder (optional) N = X-Y recorder (optional) P = xenon arc lamp S = sample compartment E= F=
EXPERIMENTAL A p p a r a t u s . The experimental system employed in this work consisted of several basic components: a continuum light source whose intensity may be modulated in a periodic manner; a means of selecting the appropriate wavelengths of light for excitation of the sample and observation of the luminescence emission; and a phase and frequency selective detector in t h e signal measurement system. A dc detection system is also used a t certain times to obtain modulation data. These basic components are very similar to the types of instrumentation used by the workers in the field of phase and modulation fluorometry as described in the introduction to this work. The aim of this work as far a s the instrumentation was concerned, was to construct the simplest analytically useable experimental apparatus possible. A block diagram of the instrumental system is given in Figure 3. The source of excitation radiation in this work is a 150-W xenon arc lamp (type 901C-11. Hanovia L,amp Div., Conrad Precision, Inc., Newark. N..J.). The source is enclosed in a lamp housing (Schoeffel Instrument Corp.. Westwood, N.J.) with a n adjustable condensing lens attachment for focusing the source radiation upon the entrance slit of a monochromator. The normal operating conditions for this lamp call for operation at 20 V and 7 . 5 A. The current for the operation of the source is provided by a Harrison 6268A dc power supply (Hewlett-Packard, Palo Alto, Calif.) operated in its constant current mode. The lamp is started using a starter as described by Zweidinger and Winefordner (20). A schematic diagram of the starting and modulation circuit is given in Figure 4. The circuit consisted of two sections; a starting section designed for normal startup of the lamp and transfer of control to the modulation section, and the modulation section itself which is designed to keep the arc operating a t a constant intensity level and to add on a sinusoidal variation in intensity. These two sections are designed such that the lamp could be started with the and J. D . Winefordner, Anal. Chem.. 42,
RP = 1 Q f 5%
*
RJ = 0-5 ki2 5% R4 = 100!1 k 5 % Rs = 1 k!l k 5% Re = 10 k!! f 5% RI = 100 kf2 5% R g = 100 kl2 f 5% S1 = SPST switch
S2 = DPST switch S3 = SPST switch S4 = SPST heavy duty switch M = Simpson 0-10A DC ammeter A,B.C,D,E,L = (See Figure 3)
ing fluorescence and phosphorescence emission spectra. In these situations. the fluorescence emission will have a phase angle, @F, which will be identical to the phase angle for the exciting radiation. Setting @R equal to 90" + @ F will phase out the fluorescence and scattered exciting light, and only the phosphorescence will be observed.
(20) R . Zweidinger
Schematic diagram of starting and modulation circuit.
643 (1970)
modulation circuit switched out, which prevented destruction of the semiconductor devices by the high energy ac pulse used to start the lamp. A Harrison Model 6116A dc power supply (Hewlett Packard, Palo Alto, Calif.) supplied the constant dc voltage level to the lamp modulation circuit. This dc voltage together with a sinusoidal signal is applied to the input of a summing operational amplifier and current booster (Model EUW-19, Heath Co., Benton Harbor, Mich.) where the voltages are amplified and added. The output of this amplifier drives the circuit which controls the current through the light source. Details of the operation of this starting and modulation circuit can be obtained by request. The modulation signal is provided by the internal oscillator of the lock-in amplifier to be used as a detector. This is similar to the experimental set-up used by Phillips (17) to modulate microwave powered discharges. The sinusoidal signal is taken from the reference in/out jack of a PAR Model 220 lock-in amplifier (Princeton Applied Research, Princeton, N . J . ) and applied to the input of the summing amplifier along with the dc voltage. The feedback arrangement of the resistors Re, R7. and Rs allows for a X10 amplification of the sinusoidal signal but no amplification of the dc level. T h e signal level available a t the reference in/out jack can be adjusted and varied from 1 V rms to 0.5 mV rms. This signal level is adjusted until the desired degree of modulation is obtained in the exciting radiation. When the modulation circuit is in operation, the starter circuit is switched out of the circuit by closing switch Sq to minimize any high frequency limitations on the amplitude of modulation which may be introduced by the RC circuits in the starter, In operation, the modulation circuit described performs satisfactorialy. The degree of modulation is selected so as to place the peak lamp current value within the current level set by the front controls of the Harrison 6268A power supply. Too high a degree of modulation causes instability and results in the extinguishing of the lamp. The lower current value also is limited, especially in operation a t the lower frequencies. Too low a current level will also cause instability in the lamp output. Of course, the degree of modulation possible will be affected by the value of the dc voltage supplied along with the modulation signal. This dc voltage will determine the constant current level for the lamp operation. and the superimposed ac signal will be centered at this value. In this work, it has been possible to modulate the light intensity up to 75% with no resulting instabilities. The degree of modulation was constant from 1 Hz through approximately 1000 Hz. Abow 1000 Hz, the amplitude of the lamp modulation began to decrease. This decrease is probably limited by the power supply.
A N A L Y T I C A L C H E M I S T R Y , V O L . 46, NO. 9 , AUGUST 1974
* 1199
A note of caution should be introduced here about the effect of the starting pulse for the lamp upon electronic components nearby. I t was found early in this investigation t h a t this high energy pulse could destroy sensitive F E T components in the lock-in a m plifier. Thus, whenever the lamp is started, all connecting cables to all the inputs of the lock-in amplifier were removed, and the lock-in turned completely off. The optical system consists of an Aminco S P F Spectrophotofluorometer (American Instrument Co., Silver Spring, Md.) which has been modified for the present studies. The usual lamp housing is removed and a slit holder with fixed slits is attached at the entrance of the excitation monochromator. The light from the source is focused on this entrance slit a s stated previously. Although the excitation and emission monochromators are baffled to reduce stray light, it is necessary to add an extra baffle in the interior of the excitation monochromator to reduce stray light even further. The gratings are blazed a t 300 n m for the excitation monochromator and a t 500 nm for the emission monochromator. The usual slit holder a t the exit slit of the excitation monochromator is modified to hold a 1-in. by 1-in. square Corning 7-54 filter. This filter prevented stray visible radiation from entering the sample compartment. The sample compartment is the standard Aminco phosphorescence cell compartment with the rotating can phosphoroscope removed. The sample solution is contained in a quartz sample cell (5-mm o.d. X 3-mm i.d.i 25 cm long which is held in place via a modified Varian A-60A N M R spinner assembly (Varian Instruments, Palo Alto, Calif.) as described by Lukasiewicz et al. ( 2 2 ) .The cell is placed into a liquid nitrogen dewar which fitted into the Aminco sample cell assembly. In these experiments, the sample cell is not rotated. The luminescence emission from the sample is detected by an RCA 1P21 photomultiplier tube (American Instrument Co., Silver Spring. M d . ) which is operated a t a constant voltage of 700 V supplied by a Heath high voltage power supply (Model EU-42A, Heath Co., Benton Harbor, Mich.). The signal from the photultiplier is directed to a load resistor which converted the signal to a voltage. The output from the load is then applied to the input of a high gain, differential amplifier (Type 1A7A. Tektronix, Inc., Portland, Ore.). The output of this amplifier is then applied to both the input of a lock-in amplifier (Model 220, Princeton Applied Research, Princeton, N.J.) and the input of a dc electrometer (Model 610BR, Keithley Instruments, Inc., Cleveland, Ohio). The dc electrometer is use,d to measure the average dc level of the signal whenever the degree of modulation is measured. The lock-in amplifier measured the average ac component of the incoming signal. The lock-in amplifier is equipped with a variable frequency control and a calibrated, adjustable phase shifter. A phase quadrant switch allowed shifting the phase of the internal reference by go", M O O , or 270". The output of the lock-in amplifier is displayed either on a Moseley X-Y recorder (F. L. Moseley, Pasadena, Calif.) or a Sargent Model T R strip chart recorder (E. H. Sargent & Co., Chicago, Ill.). For the phase resolution measurements, a n additional preamplifier (PAR Model 211, Princeton Applied Research, Princeton, N.J.) is placed between the lock-in and the 1A7A amplifier. This provided a n additional measure of sensitivity. Reagents. Reagents used without further purification are: 2bromobiphenyl, 3-bromobiphenyl, 4-bromobiphenyl, 4,4'-dibromobiphenyl. and 4-iodobiphenyl (Pfaltz and Bauer. Flushing, N.Y.): 4,4'-bisdimethylaminobenzophenoneand 4'-hydroxybutyrophenone (J. T. Baker, Phillipsburg, N.J. j; anthraquinone (sublimed) Eastman Organic Chemicals, Rochester, N.Y .): benzophenone (Fisher Scientific Co., Fair Lawn, N . J . ) . The solvent used for this study is ethyl alcohol obtained by special distillation of 95% v j v ethanol (U.S. Industrial Chemicals Co.. S e w York. S . Y . ) . Absolute alcohol is obtained by a modification of the procedure described by Lund and Bjerrum (22). Procedure. The modulation behavior for the selected phosphorescent species is determined by measuring the ac component of the luminescence signal while changing the frequency of modulation of the exciting light. The parameter m (the degree of modulation) is measured by comparing the ratio of the ac to dc component in the luminescence signal (ac,.,'dcr.j to the same ratio (acS:/ dc,*j in the excitation light.
m = (ac, / dc,,)1(ac, /dc,) (21) R J. Lukasiewicz, P. A Rozynes, L. B. Sanders, fordner, A n a / . Chem., 44, 237 (1972). (22) H . L u n d and J. Bjerrum, Ber.. 6 4 8 , 210 (1931).
1200
(28) and
J. D.
Wlne-
The excitation radiation is sampled by setting the excitation and emission monochromators to a wavelength in the visible region, say 450 nm, and measuring a portion of the scattered radiation from the sample cell. When the scatter signal is being measured, the 7-54 filter is removed and a slit width of 1 m m was used in front of the sample cell compartment. The exit slit width from the cell compartment is also reduced to 1 m m in order to limit the intensity of the scatter signal. The dc component is measured by the dc electrometer and the average ac component by the lock-in amplifier. The dc component of the luminescence signal is measured a t 50 Hz a s were also the ac and dc components of the excitation light. Lifetime values by the modulation method are calculated from the data a t 50 Hz. Phase shift angle information is obtained by measuring the instrumental phase angle of the exciting radiation on the lock-in amplifier and comparing this value to the value of the instrumental phase angle of the luminescence signal. The instrumental phase angle of the input signal was taken to be equivalent to the reference signal which gave peak response to that input signal. The peak reference phase angle is determined by rotating the quadrant switch and phase dial on the lock-in amplifier so that a null output signal is obtained for the input signal. Rotating the quadrant switch by 90" allows one to measure the peak signal and determine its phase angle. The same slits as before are used when measuring the scatter signal to determine the relative phase angle of the exciting radiation. This measurement is made a t every frequency a t which the phase angle for the luminescence is determined. Thus, the absolute phase shift angle, 8 , presented in this work represents the difference angle between the instrumental phase angle of the exciting light and the instrumental phase angle of the luminescence signal. T h e lifetime values evaluated by the phase angle method are also calculated from the data a t 50 Hz. In both the modulation and phase measurements. the read-out system did not include the PAR 211 preamplifier or the X-Y recorder. Spectral resolution of the overlapping luminescence spectra of two phosphors is accomplished by two methods: the phase method and the frequency method. The phase method involved the measurement of the instrumental phase angles for standards of each of the two compounds in the mixture. The phase quadrant and phase shifter dials on the lock-in amplifier are then set a t 90" + $PI, where $PI is the phase angle of one of the compounds. The spectrum of the mixture of components is then plotted a t this phase angle setting and frequency. The phase dials are then set so that the reference phase angle was equal to 90" + Q P Z , and the mixture spectrum is again obtained. In these measurements. the frequency of modulation is kept constant. The frequency method involved setting the phase dial and quadrant switch a t 90" &, where & is some selected phase angle. One of the standards is then run and the frequency of modulation is varied until the phase angle for this compound, $ P I r becomes equal to &. A zero output signal reading will be obtained for this standard a t this frequency. f l . A standard solution for the other compound in the mixture is then measured. and the frequency, f z , a t which its output signal goes to zero is noted. The spectrum of the mixture is then determined a t f l and fi, keeping the phase angle setting constant. For quantitative determinations of binary mixtures by the phase method, standards having concentrations near to that of the samples to be determined are measured as above in the spectral resolution studies, and the peak instrumental phase angle is determined for each of the compounds. For maximum accuracy, the most sensitive scale settings are used to determine the null point for the signals. However, the sensitivity which can be used in practice is limited by the noise and fluctuations in the total signal due to both components. Once the peak phase angles for each of the components in the mixture are known, the phase dial is set as before a t 90" f mpl, where Gpl is the peak phase angle of one of the components. A standard analytical curve is determined for the other component a t this phase setting. Xext, the reference phase is set a t 90" f and an analytical curve is measured for the other component, The signal from the binary mixture is then determined a t both phase angle settings, and the concentration of each component determined from the appropriate analytical curves. In the jrequencj method experiments. the reference phase angle, &, is set at a constant value which is equal to 90" + $s, where $> is some selected phase angle, A standard solution for each of the components in the mixture is measured. and the frequency a t the quadrature or null point is noted for each standard.
A N A L Y T I C A L C H E M I S T R Y , VOL. 46, N O . 9 , AUGUST 1974
+
The analytical curve for component 2 is determined at frequency where the signal from component 1 is nulled out. Conversely, standards for component 1 are measured at the frequency i z , where the signal from component 2 is nulled out. Again, the concentration of each component is determined from the appropriate analytical curves. In many of the phase resolution experiments, a negative signal is often obtained for one of the component signals. A positive signal can be obtained by changing the reference quadrature switch
f13
by 180".
When phasing out a fluorescence signal, the phase angle of the fluorescence signal is taken to be the same as that for a scatter signal. The phase angle, @SR, for the exciting light is measured, and the reference phase angle set at 90" f $JSR. The spectrum of the compound is then determined at this phase setting.
RESULTS AND DISCUSSION Optimization Studies. During studies, it was found that the amplitude .gain of the PAR lock-in amplifier was not constant with frequency. This behavior was confirmed by an experiment in which the response of the lock-in to a scatter signal was measured us the frequency of modulation. For the present work, the gain of the lock-in amplifier was determined to give peak response in the 10- to 100-Hz range. For the sake of convenience and accuracy when determining modulation parameters and calculating lifetimes, the response at several frequencies was assigned correction factors which ranged from 0.894 a t 2 Hz to 1.09 a t 1000 Hz. These correction factors were calculated on the basis of the response a t 50 Hz having a correction factor of 1.00. The phase response of the lock-in amplifier us frequency was also found to vary with frequency. Therefore, the phase angle of the scattered light was always measured whenever a phase measurement on a signal was desired. The per cent modulation of the exciting light is defined as the ratio of the ac to the dc component of the exciting radiation times 100%. Most of the experiments in this work were conducted with a source modulation of approximately 53%. This degree of modulation was found to produce a stable output from the lamp and preserved its life. For some of the spectral resolution experiments, a degree of modulation of 75% was used; the greater the per cent of modulation of the exciting light, the more sensitive will be the measurement of the phosphorescence. However, this work was not concerned with limits of detection but rather demonstration of a principle, and thus the degree of modulation which gave the most stable signal was used. The effect of changing the load resistance upon the measured phase angle of scatter was also studied; a t frequencies below 500 Hz, the load resistance had little effect upon the instrumental phase angle. The gain setting on the 1A7A amplifier also produced a change of the phase angle at the high gain settings, and so high gain scales were avoided when phase measurements were taken and greater use was made of changing the load resistor. When the PAR Model 211 preamplifier was employed, its gain was set a t X 1 0 and not varied; the phase angles for the scatter signal appeared to vary considerably with gain setting on this amplifier. As mentioned before, the solvent used in these studies was absolute ethanol. During the first experiments, it became apparent that the condition of requiring a glassy, uncracked matrix for measurement of phase and modulation was critical. Phase angle values could vary as much as 6" if the sample was cracked rather than clear. This variation could possibly be due to an increase in scattered light from a cracked sample over a clear sample. I t was
h
\4 joo
FREOUENCY
(Hz)
Figure 5. Variation of phosphorescence signal with the frequency of modulation for several phosphors A = Benzophenone B = 4-Bromobiphenyl
C = 4-Hydroxybutrophenone D = 4,4'-Bis(dirnethylarnino) benzophenone
also noted that the relative signal increased when the sample was cracked. The variation of the phosphorescence signal with modulation frequency is illustrated in Figure 5 for some of the molecules measured. In all cases, the measurements were made a t the peak phosphorescence excitation and emission wavelength. The curves for benzophenone and 4-bromobiphenyl followed the expected behavior for molecules of their respective lifetimes, with a leveling off a t low frequencies and a linear portion extending to higher frequencies. The curve for 4-hydroxybutyrophenone displays a deviation in slope a t the higher frequencies. The curve for 4,4'-bis(dimethy1amino)benzophenone seems quite anomalous, Benzophenone is known to show strong phosphorescence and no fluorescence. In the case of the 4-bromobiphenyl, again the phosphorescence is strong, but a fluorescence emission is also observed. By choosing the appropriate excitation and emission wavelengths, the fluorescence was reduced considerably. Traces of this fluorescence could possibly account for the slight upward deviation of the curve a t the very high frequencies. For 4-hydroxybutyrophenone, selection of wavelengths was also employed, but in this case, a stronger fluorescence is present, and the deviation occurs a t lower frequencies. The rapid leveling off of the 4,4'-bis(dimethy1amino)benzophenone (4,4'-DMAB) curve is due to the fact that the fluorescence and phosphorescence emission spectra for this compound overlap severely. The behavior of the 4,4'-DMAB illustrates one of the major limitations of the experimental technique and theory of phase resolution in phosphorimetry. Because the total luminescence emission is observed, a molecule which shows strong fluorescence and weak phosphorescence and/ or which shows severe overlap of the fluorescence and phosphorescence emission spectra results in a severe deviation in the signal us frequency curve. This. of course, limits the determination of accurate modulation factors and phase angles, and subsequently an accurate lifetime value cannot be determined. In quantitative measure-
A N A L Y T I C A L C H E M I S T R Y , VOL. 46, NO. 9 , A U G U S T 1974
*
1201
90
:80
$ 70 9
60
L z
40
4
30
h
20
50
10 '1
2
3
5
10
20
30
FREQUENCY
FREPUE NCY
A = Anthraquinone, TP = 3.0 msec B = 4-lodobiphenyl, ~p = 3.5 rnsec C = Benzophenone, 7 p = 6.0 msec D = 4,4'-Dibromobiphenyl,7 p = 1 2 rnsec E = 4-Bromobiphenyl, ~p = 17. rnsec F = 3-Bromobiphenyl. 7p = 55. rnsec
T a b l e I. Phosphorescence Lifetimes f r o m Phase a n d Modulation D a t a a Lifetime (rns)* Molecule
PhaseC
Modulationc Time resolvedd
6.1 2.7 3.3 20. 59. 13. Relative errors in lifetimes are *lo%. 5.9
3.0 3.5 14. 55. 12.
Solvent: ethanol. culated from measurements at 50 Hz. a
7. 3.6 3.2 17. 58. 12.
All values calData taken from references (8)and
(9).
ments, the presence of the fluorescence emission places a lower limit to the sensitivity of the phase resolved phosphorescence technique because only one signal can be phased out a t any one time. Thus, in resolving two phosphorescent species, the limit of detection would be limited by a constant fluorescence signal. If the fluorescence signal is weak, however, the phase settings used in the experiment could reduce the fluorescence signal and improve the analysis. In order to simplify the measurements of modulation and phase parameters, the molecules selected for further study were limited to those in which the fluorescence interference was small. P h a s e and Modulation Characteristics. A determination of the degree of modulation parameter, m, and the phase angle parameter, 4, was made for a series of molecules (see Figures 6 and 7 ) . The modulation curves for these molecules show the expected variation of amplitude with modulation frequency. From Figure 6, it can be seen that the degree of modulation for long lifetime molecules decreases more rapidly than the phosphorescence from short lifetime molecules. All the curves tend to approach a maximum value for m of 1.0 a t low frequencies and approach a value of zero as the frequency is increased. The lifetime of these molecules may be calculated from the value of m a t any frequency using Equation 9. However, the most accurate values of the degree of modulation are obtained in that portion of the curve with the greatest 1202
100
200 300 500
1000
(Hd
Figure 6. Variation of experimentally determined degree of modulation, m, with the frequency of modulation for several phosphors
Benzophenone Anthraquinone 4-Iodobiphenyl 4-Bromobiphenyl 3-Bromobiphenyl 4,4'-Dibromobiphenyl
50 (Hzl
Figure 7. Variation of experimentally determined phase shift angle, 8, with the frequency of modulation for several phosphors A = Anthraquinone, T = 3.0 msec B = 4-lodobiphenyl, 7 = 3.5 msec C = Benzophenone, T = 6.0 msec D = 4,4'-Dibromobiphenyl, T = 1 2 msec E = 4-Bromobiphenyl. T = 17. msec F = 3-Bromobipheny1, 7 = 55. msec
slope. Therefore, the lifetime values were determined a t 50Hz. The modulation curves indicate enhanced response for a short lifetime molecule over that for a longer lifetime molecule. This is especially important when the background phosphorescence from the solvent is of a long-lived nature. The curves also show a disadvantage of this experimental technique in that only signals from phosphorescent species with lifetimes shorter than approximately 50 msec can measured easily-ie., frequencies below about 10 Hz result in measurement difficulties such as low frequency fluctuations in the signal, due to bubbling in the liquid nitrogen dewar, and the need for long time constants when making these measurements. The phase shift angle measurements demonstrate the expected behavior for molecules of different lifetimesi.e., the phase shift for the luminescence signal varies from 0" a t low frequencies to 90" a t high frequencies. It should be noted here that the error in these measurements is greatest a t the very low ( < l o Hz) and high frequencies (>lo0 Hz).At the low frequencies. bubbling noise made the determination of null signals very difficult. A t the high frequencies, the determination of the phase angle became difficult because of the low signal levels involved and the increased possibility of observing scattered and fluorescence radiation. The leveling off of these curves a t approximately 85" could be attributed to these factors as well as to the characteristics of the lock-in amplifier at these higher frequencies. Phase shift angles were improved somewhat when the 7-54 filter was used in the excitation path which indicated some visible stray light contribution a t the higher frequencies. Lifetime data can be calculated from the phase angle value a t any frequency, but here again, as in the modulation experiment, the most accurate values are those which lie on that portion of the curve with the greatest slope. The lifetimes were calculated from the data a t 50 Hz. In Table I, the lifetimes calculated from the phase and modulation data are given, these values compare well to the lifetimes reported by time resolved phosphorimetry . The lifetimes were calculated from the phase data using the equation
ANALYTICAL C H E M I S T R Y , VOL. 46, NO. 9, A U G U S T 1974
8
=
tan-' 2 ~ f r
(29)
80
80
?\
70
*
k rn
60
> k
50-
50-
I-
40 -
w
w
2
5
30-
4w
z
I-
z L
40 -
30-
I-
K
-I
w
a
60 -
20 -
20 IO IO
-
01 200
OL 200
400
300
WAVELENGTH (nm)
Figure
Phase resolved emission spectrum of a mixture of 2.2 benzophenone and 2.9 X 1 0 - 5 M 4-bromobiphenyl at 25 Hz. Excitation wavelength is 275 n m 8.
X 10-5M
A = Mixture spectrum at peak phase angle, @R = 270' 6 = Mixture spectrum, @R = 0" 26'
+
C=
+
+ 39"
Mixture spectrum, @R = 180' 52" ( - - - ) = Spectrum of benzophenone standard, @R = 270" 52" (- - -1 = Spectrum of 4-bromobiphenyl standard, @E = 270' 26'
+
400
500
600
WAVELENGTH (nm)
60 0
500
300
Figure 9. Frequency method resolved emission spectrum of a mixture of 2.2 X lO-5M benzophenone and 2.9 X 10-5M 4bromobiphenyl. Excitation wavelength is 275 nrn A = Emission spectrum of mixture, frequency = 37.5 Hz, @R = 180" i40' B = Emission spectrum of mixture, frequency = 14.7 Hz, @R = 180" -t 40" ( - -1 = Emission spectrum of benzophenone standard (- - ) = Emission spectrum of 4-bromobiphenyl standard
.
~
+
The results in Table I indicate good agreement within the experimental error between the phase and modulation data. Phase Resolved Phosphorescence of Synthetic Mixtures. The selection of optimum experimental conditions for phase resolution studies involved the selection of the optimum instrumental conditions. When one of the components of a mixture is nulled (phased) out, the resulting signal from the other component is proportional to the sine of the difference in phase angles for the two components, and so the phase characteristics of the two components of the mixture are important. The operating frequency was also important because it affects the maximum signal from the compounds and the value of the phase angles. Referring to Figure 7, it can be observed that the maximum difference in phase angle for the molecules studied occurs in the frequency range between approximately 10 and 100 Hz. The ease of measurement would favor the use of the higher frequency end of this scale. However, for some of the long-lived phosphors, the degree of modulation drops to 10% or less of its original value in this higher frequency range. As a compromise choice, the operating frequency was chosen as 25 Hz for the spectral and quantitative measurements involving phase resolved phosphorimetry. A number of binary mixtures of phosphors were spectrally resolved by means of the phase method of nulling out one of the components. One mixture studied consisted of benzophenone and 4-bromobiphenyl; there was approximately 25" difference in their phase angles and the spectral emission characteristics of each were quite different, whereas the excitation spectra overlapped severely. Thus, only the emission spectra are reported here (see Figure 8). Spectrum A of Figure 8 is the spectrum of the mixture measured a t the peak output phase setting which in this case lies between the phase angles for each of the two components. By setting the reference phase setting on the
lock-in amplifier so that the signal from 4-bromobiphenyl was 90" out of phase with the reference signal, the spectrum of the mixture was again determined and Curve B in Figure 8 results. On the other hand, if the reference phase angle is set such that the signal from the benzophenone is phased out, the spectrum of the 4-bromobiphenyl, Spectrum C of Figure 8 results. In certain situations, even if the excitation and emission spectra of the two components do not completely overlap, phase resolution may still be quite useful if one component emits a much stronger phosphorescence then the other component. A good example of this problem was found in the phosphorescence spectrum of a mixture of anthraquinone and 4-bromobiphenyl. By means of spectral and phase resolution, the strong phosphorescence signal of anthraquinone can be completely phased out, and the spectrum of the weaker phosphor, 4-bromobiphenyl, can be easily measured. Phase resolved excitation and emission phosphorescence spectra of the binary mixture of anthraquinone and 4-bromobiphenyl can be obtained by writing one of the authors (JDW). An alternate method of resolution is called the frequency method and the use of this technique is illustrated in Figure 9 for the mixture of benzophenone and 4-bromobiphenyl; in this experiment, however, the reference phase angle was set at a value 90" away from a selected phase angle &. This phase angle was selected by consideration of the modulation and phase curves given in Figures 6 and 7. From previous measurements, it was known that the instrumental phase angle of the exciting light was approximately 90" + 270" on the lock-in amplifier. Setting 270" would the reference phase angle constant a t 40" give maximum response to a molecule having its absolute phase shift angle a t around 50" + 270". This corresponds to a horizontal line a t 50" on the 8 us. frequency diagram, Figure 7. Observation of the intersection point with the phase curves for 4-bromobiphenyl and benzophenone shows that these points occur at approximately 10 Hz and
+
ANALYTICAL C H E M I S T R Y , VOL. 46, NO. 9, A U G U S T 1974
1203
I
I
0
id'
io-.
id '
id'
Concentrotion (moles/i~ferl
0 1 ' 200
400
300
Figure 12. Analytical curves for 4-bromobiphenyl and 4-iodobiphenyl at the peak phase setting for each molecule and at the phase setting where one component is phased o u t
600
500
WAVELENGTH (nm)
Figure 10. Excitation and emission spectra of 2 1 bromobtphenyl at 50 Hz and @R = 270' 68"
+
X
lO-3M 2-
A = Excitation spectrum emission wavelength = 465 nm B = Fluorescence emission peak, excitation wavelength = 275 nm C = Phosphorescence emission peak, excitation wavelength = 275 nm ( - - -) = Phosphorescence emission spectrum of 2-bromobiphenyl standard
70 60
t1
0 200
500
400
300
600
WAVELENGTH (nm)
Figure 1 1 . Phase resolved
excitation
and emission spectra
2 1 X 1 0 - 3 M 2-bromobiphenyl at 50 Hz and 95 3"
C$R
Of
= 180' 4-
A = Excitation spectrum, emission wavelength = 465 nm B = Residual fluorescence emission, excitation wavelength = 270 nm C = Phosphorescence emission spectrum, excitation wavelength = 270 nm
35 Hz. Therefore, if the reference phase angle is set at 40" + 180" and the frequency of modulation adjusted to near 10 Hz, the signal from the 4-bromobiphenyl would be phased out. Similarly, if the frequency is adjusted to near 35 Hz, the signal from the benzophenone would be phased out. The spectrum of the mixture (see Figure 9) determined at 13.1 Hz shows only the spectrum of the benzophenone, B. At 37.5 Hz, the spectrum, A, of the biphenyl results. As mentioned earlier, the fluorescence emission from a sample may also be phased out by proper selection of the 1204
A1 = Analytical curve for 4-iodobiphenyl. $R = 270' f 55" A2 = Analytical curve for 4-iodobiphenyl, @R = 0" f 12.5" 61 = Analytical curvefor 4-bromobiphenyl, $R = 270" f 12.5" 82 = Analytical curve for 4-bromobiphenyl, $R = 180" = 55'
reference phase angle. An example of this ability is given in Figures 10 and 11. In Figure 10, the excitation and emission spectra are given of a solution of 2-bromobiphenyl determined at the reference phase angle which gave the peak response for the phosphorescence emission. Curves A and C of Figure 10 are the phosphorescence excitation and emission spectrum for this solution. Note that an off-scale intense fluorescence emission (Curve B of Figure 10) is also present. Because in the frequency range of operation in this technique, the fluorescence should have the same phase and frequency characteristics as the source radiation, measuring the scattered light phase angle, &K, and setting the reference phase angle at 90" k @SK should phase out this fluorescence emission. The results are illustrated in Figure 11. Note here the residual fluorescence signal (Curve B of Figure 11) and the undistorted phosphorescence excitation and emission spectra (Curves A and C, respectively, of Figure 11). In the preceding examples of spectral resolution, the phase angles of the individual components were determined through the use of pure standard solutions. However, in some cases where the excitation and emission spectra do not completely overlap, these individual phase measurements can be made directly on the mixture through wavelength selection, if either the excitation or emission spectra are widely different, or by measurements on the wavelength edges of the spectrum where the signal from one component predominates. Quantitative Analysis of Binary Mixtures. Most quantitative measurements were carried out using only the phase method because it required measurements at only one frequency. In the frequency method, measurements had to be made at two frequencies which necessitated accurate resetting of the frequency values; this was more difficult to achieve with the experimental apparatus used here than the resetting of reference phase angle values in the phase method. Also, the selection of an appropriate phase angle setting in the frequency method was critical because the null frequencies established by the chosen phase angle setting had a great effect upon the signal magnitudes obtained from each phosphor. Referring to Equations 23 and 24, the signal obtained from the remaining component of a binary mixture when
ANALYTICAL CHEMISTRY, VOL. 46, NO. 9 . AUGUST 1974
~
.~
~~
Table IV. Phase Resolution of B i n a r y Mixture of Benzophenone and 4-Bromobiphenyl b y the Phase Methodalb.c
Table 11. Phase Resolution of Binary Mixture of 4-Iodobiphenyl and 4-Bromobiphenyl b y the Phase Methoda.bjC Mixture
Lifetimes, msec
A: 4-Iodobiphenyl B: 4-Bromobiphenyl
3.5 17.
Concn added, M
X X X 10-5 X
Concn found, M
+ 12.5 + 55.0 0 + 12.5 180 + 5 5 . 0 0 + 12.5 180 + 5 5 . 0
A: 8 . 0 X 10-7 0 B: 1 . 8 6 X 10-6 180 A: 8 . 0 B: 1 . 8 6 A: 1 . 6 0 B: 1 . 8 6
270 270
Phase angle setting,
Mixture
Lifetimes, msec
A: Benzophenone B: 4-Bromobiphenyl
6.0 17.0
Peak phase angle,
8.3 X
1.80 X 8.0 X 1.77 X
+ 55.0 + 12.5
Concn added, M
Error,
7a
lo-’ lo-‘ lo-‘
1.60 X 1.20 X
A: B: A: B:
+3.8 -3.3 0 -4.8 0 -36
Phase angle
0 180 0 180
6.52 X 8.38 X 1.30 X
270 270
Concn found, M
setting,
4.19 X
Peak phase angle,
+ 27.5 + 53.0 + 27.5 + 53.0
4.0 6.9 7.6 1.4
+ 53.0 + 27.5 Error, 5%
-4.6 +5.8 -9.3 +7.7
X
X X X
a All measurements performed at 275 nm excitation wavelength and 475 nm emission wavelength, All measurements performed at 25 Hz. Degree of source modulation = 537,.
a All measurements performed at 275 nm excitation wavelength and 475 nm emission wavelength. All measurements performed at 25 Hz. Degree of source modulation 537;.
Table V. Phase Resolution of a Binary Mixture of Benzophenone and 4-Bromobiphenyl by the
Frequency Methodalb Table 111. Phase Resolution of B i n a r y Mixture 4,4’-Dibromobiphenyl and 4-Iodobiphenyl by the Phase Methoda9b.c Lifetimes, Mixture
mi3ec
A: 4,4’-Dibromobiphenyl B: 4-Iodobiphenyl
12. 3.5
Concn added, M
A: B: A: B: A: B:
1.83 X 1.89 X 3.66 x 3.78 X 1.83 X 3.78 X A: 3 . 6 6 X B: 1.89 X
Phase angle setting,
10-5
10-6 10-6 10-5
10-6 10-6 10-5
+ 67.0 + 29.8 + 67.0 + 29.8 180 + 6 7 . 0 0 + 29.8 180 + 6 7 . 0 0 + 29.8 180 0 180 0
Peak phase angle, *
270 270
Concn found, M
1.80 X 1.65 X 3.90 X 3.55 X 1.90 X 3.60 X 3.60 X 1.95 X
lo-‘
+ 29.8 + 67.0
Mixture
Lifetimes, msec
A: Benzophenone B: 4-Bromobiphenyl
6.0 17.
Concn added, M
Frequency,
Hz
Phase angle setting,
A: 4.50 x 10-6 1 3 . 1 0 B: 5 . 8 4 X 3 7 . 5 180
Error,
70
+ 40.0 + 40.0
Concn found, M
Error,
‘A
+8.9 -9.3
4.9 X 5.3 X
All measurements performed at 275 nm excitation wavelength and 475 nrn emission wavelength. Degree of source modulation = 75%.
‘
-13. -1.6 +6.6 -6.4 +3.8 -4.8 -1.6 +3.2
Table VI. Precision M e a s u r e m e n t s on Two B i n a r y Mixturesa,b.c
All measurements performed at 275 nmexcitation wavelength and 475 nm emission wavelength. I, All measurements performed at 25 Hz. Degree of source modulation = 53%. (I
Mixture 1
A: 1 . 3 X 10-hM
Mixture 2
B: 3.8 X 10+M A: 2 . 6 X 10-6M B: 2 . 1 X 10-EM
Mixture
1
one component is nulled out is reduced proportionally to the sine of the phase angle difference (see Figure 12). Figure 12 shows t h e analytical curves obtained for standard solutions of two compounds of a binary mixture when determined a t their peak signal phase angle and at the phase angle where the signal from one of the components is phased out. Calculating the ratio of the values for the relative signal at the same phase setting for each pair of curves, yields an average value of 0.68 for the 4-bromobiphenyl curves and 0.62 for the 4-iodobiphenyl. Taking the sine of the difference phase angles, in this case 4 2 3 , yields a value of 0.676. These values agree within experimental error. The results of analyses performed upon several binary mixtures are given in Tables I1 through IV. These results indicate that the best data are obtained when the concentration of each of the compounds in the mixture is less than 10-5M. This seems to suggest that an inner filter effect could be operative a t the higher concentrations. Also, depending upon the pair of compounds selected for the mixture, a ten times greater concentration of a strongly phosphorescent species in the presence of a weaker phosphorescent species introduces large errors in the determination of the less phosphorescent species. This
2
Phase setting,
A: 180 B: 0 A: 180 B: 0
+ 69.3 + 29.4 + 52.3 + 24.3
4-Bromobiphenyl 4-Iodobiphenyl 4-Bromobiphenyl Benzophenone No. of detns
Re1 std dev, 470
10 10
5.9 10. 7.4 8.9
10 10
a All measurements performed at 275 nm excitation wavelength and 475 nm e m k i o n wavelength. All measurements performed at 25 Hz. Degree of source modulation = 53%.
is evident in Table I1 in which the strongly phosphorescent 4-iodobiphenyl is present along with the weaker phosphorescent 4-bromobiphenyl. Increased errors are probably due to the inability to completely phase out a strong signal and to detect a weak signal in the presence of a large noise component. It should be noted here that the noise is dependent upon the sum of the noise components from both phosphors and the phasing out of one component signal does not phase out the noise associated with it. An example of the use of the frequency method for the quantitative analysis of a binary mixture is given in Table V. Compared with the results (Table IV) obtained for a comparable mixture by the phase method, the per cent error in the frequency method is greater. It should be
A N A L Y T I C A L C H E M I S T R Y , V O L . 46,
NO. 9,
A U G U S T 1974
1205
stressed that the signal level for the 4-bromobiphenyl is lower here because the measurements are made at 37.5 Hz instead of 25 Hz as in the phase method. The precision of quantitative measurements by the phase resolution method are given in Table VI. The per cent relative standard deviation expected is of the same order as the per cent error in the analysis of the binary mixtures. The principal sources of random noise in this experiment are bubbling noise in the liquid nitrogen cool-
ant and electronic noise from the amplifier's. Of these two, the bubbling noise is especially bad at the low frequencies of operation. Received for review December 28, 1973. Accepted April 8, 1974. Research was carried out as a part of a study on the phosphorimetric analysis of drugs in blood and urine, supported by U S . Public Health Service Grant No. GM11373-11.
Pulsed Source Time Resolved Phosphorimetry for the Quantitative and Qualitative Analysis of Drugs K. F. Harbaugh,' C. M. O'Donnell,2 and J. D. Winefordner3 D e p a r t m e n t of C h e m i s t r y . U n i v e r s i t y of Florida, Gainesville, Fla. 326 7 7
Pulsed source time resolved phosphorimetry is used to qualitatively identify and quantitatively measure synthetic mixtures of drugs. No physical separations are needed but rather tempciral resolution is sufficient to resolve mixtures of phosphors. The time resolved phosphorimeter consisted of a pulsed xenon flash lamp source, an emission monochromator with photomultiplier tube detector, and a signal averager-recorder readout system. Because phosphorescence lifetimes of organic molecules vary considerably with structure and with environment, whereas the phosphorescence emission spectra of organic molecules are quite similar, temporal resolution is a far more selective method of measurement than spectral resolution. The drugs studied in this work were morphine, ethylmorphine, codeine, quinine, procaine, phenobarbital, amobarbital, cocaine, amphetamine, and rnethamphetamine.
Temporal resolution in phosphorimetry was first used by Kiers, Britt, and Wentworth ( I ) in 1957, when they showed that a binary mixture of structurally similar compounds could be resolved by judicious selection of excitation and emission wavelengths and by choice of the delay before the observation time and after termination of the exciting radiation. Initial time resolution studies made use of a spectral continuum source with mechanical shutters for termination of exciting radiation. Kiers et ul. ( I ) used a modified Becquerel type rotating disk, St. John and Winefordner ( 2 ) a manual guillotine shutter, and Hollifield and Winefordner ( 3 ) a single disk mechanical phosphoroscope. Winefordner ( 4 ) was first to suggest the advantages of pulsed source, gated detector instrumentation for time resolved analytical phosphorimetry, and O'Haver and Wine-
fordner (5) put pulsed source phosphorimetry on a sound theoretical basis. Fisher and Winefordner (6) described a pulsed source, gated detector phosphorimeter and compared the system to conventional phosphorimeters. Later, O'Donnell, Harbaugh, Fisher, and Winefordner (7) modified this system and demonstrated the utility of time resolved phosphorimetry for the quantitative measurement of synthetic mixtures of halogenated biphenyls. They also suggested the use of time resolved phosphorimetry for qualitative identification of phosphorescent molecules. Since that time, however, no other analytical uses of pulsed source time resolved phosphorimetry have been shown. The use of drugs for both legitimate and illicit purposes is extremely widespread. The qualitative identification of drugs and the estimation of drug levels in biological materials for therapeutic purposes is needed for the effective use and/or control of drugs. Because conventional phosphorimetry as well as fluorimetry have already proved to be sensitive analytical methods for measurement of drugs which phosphoresce ( 8 - E ) , the application of time resolved phosphorimetry to the selective determination of drugs should greatly extend the usefulness of luminescence spectrometry to drug assay. In the present study, a pulsed source, gated detector, time resolved phosphorimeter is used for the qualitative and quantitative measurement of selected drugs in serial solvents. The advantages and limitations of this technique are illustrated for this application. The feasibility of drug analysis in real samples is indicated.
EXPERIMENTAL Apparatus. The p u l s e d source phosphorimeter used in these studies was i d e n t i c a l t o t h e one described p r e v i o u s l y b y O'Donnell, H a r b a u g h , a n d W i n e f o r d n e r (7). A l l phosphorescence m e a -
Present address, Celanese F i b e r Co., Charlotte, N.C. Present address, D e p a r t m e n t o f C h e m i s t r y , Colorado S t a t e U n i v e r s i t y , F o r t Collins, Colo. A u t h o r t o w h o m r e p r i n t requests s h o u l d be sent. (1) R . J. Kiers. R . D. Britt, Jr., and W. E. Wentworth, Anal. Chem.. 29, 202 (1957) ( 2 ) P.A. St. John and J. D. Winefordner, Anal. Chem.. 39, 500 (1967). ( 3 ) H . C. Hollifield and J . D. Wlnefordner. Chem. Insfrum.. 1 , 341 (1969). ( 4 ) J. D. Winefordner, Accounts Chem. Res . 2, 361 (1969)
1206
A N A L Y T I C A L C H E M I S T R Y , VOL. 46, NO. 9, AUGUST 1974
T. C. O'Haver and J. D. Winefordner, Anal. Chem.. 3 8 , 6 0 2 ( 1 9 6 6 ) . R. P. Fisher and J. D.Winefordner, Anal. Chem. 44, 948 ( 1 9 7 2 ) . C. M. O'Donnell, K . F. Harbaugh, R. P. Fisher, and J. D. Winefordner. Anal. Chem.. 45. 609 (1973) J D Winefordner and M. Tin, Anal C h m Acta 32, 64 (1965) /bid 31. 239 11964) W. J. McCarthy, P. A.St. John, and J. D. Winefordner, "Phosphorimetry as a Means of Chemical Analysis," in "Fluorescence Assay in Biology and Medicine," S. Udenfriend, Ed.. Academic Press, New York, N.Y., 1970. H. C. Hollifield and J. D. Winefordner. Talanta. 12, 860 (1965) C. J Miles and G . H. Schenk, Ana/. Chem , 45, 130 (1973)
