Anal. Chem. 1999, 71, 2700-2707
Whole-Column Radioactivity Detection: Simultaneous Separation and Enhanced Detectability Jeanne M. Link† and Robert E. Synovec*,‡
Department of Radiology, Box 356004, University of Washington, Seattle, Washington 98195-6004, and Department of Chemistry, Box 351700, University of Washington, Seattle, Washington 98195-1700
The development of a whole-column radiation detector for measurement and separation of low amounts of beta emitting analytes is described. The design of this detector is unique with all of the chromatography media located within the detector volume. This whole-column design provides the advantage of increased radiation signal without loss of chromatographic efficiency, which translates to increased detectability. This increase was compared theoretically with flow-through radiation detection, and the theory was tested experimentally. Using two analytes, carbon-11-labeled m-hydroxyephedrine and r-methylepinephrine, only 3 and 8 Bq (80 and 220 pCi), respectively, were needed to obtain a 10% coefficient of variation using whole-column detection. For [11C]-mhydroxyephedrine, 100 times more radioactivity was required to achieve the same coefficient of variation using flow-through detection. A limit of detection (LD) for the analytes of 2 Bq (54 pCi) was obtained for whole-column detection, an improvement of 50 in LD compared with flow-through detection. Signal improvement increased linearly with the chromatographic resolution. The wholecolumn detection method is robust and applicable to many chromatographic separations. Many flow-through radiation detectors have been developed for measuring low-level radioactive analytes following separation by chromatography or other techniques. Three general approaches have been taken to improve radioactivity detection: develop high sensitivity by increasing either the detector efficiency or the time the signal is viewed by the detector,1 improve the limit of detection by decreasing the background of the detector,2,3 and integrate the flow-through signal to increase signal-to-noise ratio.4 In this paper, we describe a new approach, whole-column radiation detection. * Corresponding author. E-mail:
[email protected]. † Department of Radiology. ‡ Department of Chemistry. (1) Pentoney, S. L., Jr.; Zare, R. N.; Quint, J. F. Anal. Chem. 1989, 61, 16421647. (2) Nickles, R. J.; de Jesus, O. T.; Solin, O.; Haaparanta-Solin, M. IEEE Trans. Nucl. Sci. 1992, 39, 2316-2321. (3) Shear, J. B.; Colon, L. A.; Zare, R. N. Anal. Chem. 1993, 65, 3708-3712. (4) Synovec, R. E.; Yeung, E. S. Anal. Chem. 1985, 57, 2162-2167.
2700 Analytical Chemistry, Vol. 71, No. 14, July 15, 1999
Whole-column detection has been previously described for refractive index5-7 and fluorescence detection in capillary chromatography8 and absorbance detectors.9-12 We propose that whole-column radiation detection is a method that can be applied in a novel way to increase radioactivity detectability while simultaneously separating the analytes of interest. Our approach to whole-column detection is unique. With our approach, the entire chromatography column is contained within the radiation detector volume. After injection, all of the radiolabeled compounds are on the column at one time and are detected together until each compound sequentially elutes, reducing the detected signal by the amount eluted, thus preserving the temporal information of the separation. The HPLC medium is contained in a planar geometry within the whole-column radiation detector to minimize the distance between the chromatography medium and the scintillation material. The positrons to be detected are emitted with an energy spectrum distributed from 0 to 0.96 MeV13 and a maximum range of 3.5 mm. The planar geometry is only 500 µm thick to minimize the energy loss of the positron within the chromatography phases before it reaches the scintillator. There are many radiotracers which could be used to test the whole-column radiation detector. We chose carbon-11-labeled m-hydroxyephedrine (MHED) and R-methylepinephrine as tracers because they are analogues of epinephrine that are useful for studying the sympathetic nervous system using positron emission tomography. In addition, they were convenient to use because we had previously characterized a chromatographic separation of related molecules.14 The increased radiation detectability with whole-column radiation detection is due primarily to continuous (5) Xi, X.; Yeung, E. S. Anal. Chem. 1990, 62, 1580-1585. (6) Foster, M. D.; Synovec, R. E. Anal. Chem. 1996, 68, 1456-1463. (7) Synovec, R. E.; Sulya, A. W.; Burgess, L. W.; Foster, M. D.; Bruckner, C. A. Anal. Chem. 1995, 67, 473-481. (8) Johansson, J.; Johansson, T.; Nilsson, S. Proc. SPIE-Int. Soc. Opt. Eng. 1996, 2629, 2-9. (9) Gelderloos, D. G.; Rowlen, K. L.; Birks, J. W.; Avery, J. P.; Enke, C. G. Anal. Chem. 1986, 58, 900-903. (10) Rowlen, K. L.; Duell, K. A.; Avery, J. P.; Birks, J. W. Anal. Chem. 1989, 61, 2624-2630. (11) Rowlen, K. L.; Duell, K. A.; Avery, J. P.; Birks, J. W. Anal. Chem. 1991, 63, 575-579. (12) Wang, T.; Hartwick, R. A. Anal. Chem. 1992, 64, 1745-1747. (13) Lederer, C. M., Shirley, V. S., Eds. Tables of Isotopes, 7th ed.; John Wiley & Sons: New York, 1978. (14) Link, J. M.; Synovec, R. E.; Krohn, K. A.; Caldwell, J. H. J. Chromatogr., B 1997, 693, 31-41. 10.1021/ac981401o CCC: $18.00
© 1999 American Chemical Society Published on Web 06/11/1999
Because of the constraint that the radioactivity be constant over the experiment, A(t) simplifies to A. The detector response, Eft, is the amount of electronic pulses detected for each nuclear decay event, or efficiency. The response depends on decay characteristics of the radionuclide and properties of the detector and is constant for any radionuclide. Thus, the flow-through signal simplifies to
Sft(t) ) AF(t) Eft
Figure 1. Simulated whole-column detector signal (- - -), compared with the flow-through detector signal (s) as a function of time postinjection for a single analyte. Detector response, S(t), is equal for the two detectors, Eft ) Ewc)1, and is plotted as a fraction of activity (A) injected where A is equal to 1. The flow-through detector residence is optimized as described in the text with τft ) 0.235σ. The times are relative; t′ is when the injectate is fully loaded onto the column, and t1 and t2 are the tr - 3σ and tr + 3σ integration limits. The increase in signal area with whole-column detection is readily apparent.
detection of the radioactive analytes as they pass through the column. This advantage is especially important for short-lived radioactivity such as 11C. Because detection occurs on-column, the efficiency of the separation is maintained without the decrease in resolution caused by band broadening postcolumn within a flowthrough detector. Postcolumn loss of resolution decreases the signal within a flow-through detector and thus decreases the signal-to-noise ratio (S/N). Also, because whole-column detection happens earlier in time than flow-through detection, there is less decay of the analyte and thus more signal for whole-column detection. THEORY In this section, the signals from flow-through and whole-column radiation detectors are simulated using typical parameter values. Many of the variables which contribute to the detector signals are constrained in order to focus on the differences between the two methods of detection. The constraints are that the detection of only a single analyte be considered in separations for two cases: retention of a single analyte and a separation with two analytes. No background is included initially, but it is added to the theoretical discussion later. The final constraint is that the total radioactive decay rate be constant over the time of the experiment. Within these constraints, the signals from each detector are compared theoretically. The significance of the difference between the signals is discussed, as well as the effect of the constraints. Flow-Through Radioactivity Detection Model. A flowthrough radiation detector, located at the outflow of a column, detects only those decay events which occur as the analyte elutes from the column and passes through the detector (Figure 1). The signal from a flow-through (ft) radiation detector, Sft(t), is in units of counts per sampling interval and sums the decay of the analyte within the detector and background counts, Bft, not due to the analyte. For a single analyte and no background, the signal is the product of three terms: the analyte radioactivity, A(t); the fraction of the analyte within the detector during the sampling time, F(t); and the response of the detector for the analyte radionuclide, Eft.
(1)
In chromatography, analytes are injected simultaneously, retained for varying times, and then sequentially eluted from the column. During the separation, the concentration of the analyte changes, due to on- and off-column band broadening, and the analyte elutes as an approximately Gaussian peak. The fraction of the analyte which passes through the detector during any sampling interval, Fft(t), can be modeled as a Gaussian peak with an area of unity times the residence time of the analyte within the detector:
Fft(t) ) τ
[
(
)]
2
(t - tr) 1 exp 1/2 σ(2π) 2σ2
(2)
where τ is the residence time of the analyte within the detector, which is equal to the flow-through detector volume divided by the flow rate through the detector and is given in units of detector sampling frequency for this work, t is time since injection, tr is the median residence of the analyte peak, and σ is the standard deviation of the peak width, also in units of sampling time. The upper limit of the residence time of the analyte in a flowthrough detector can be defined on the basis of chromatography theory. A criterion for flow-through detectors is that the time constant for sampling should result in at least 10 points for peak width at half-height, or at least 0.235σ per sampling point.15 Thus, the residence time in a flow-through detector is τ e 0.235σ. For the rest of this discussion, 0.235σ will be used for τ in order to have the maximum flow-through counts for comparison with whole-column detection. If F(t) is integrated from tr - 3σ to tr + 3σ, the result of eq 2 will be 0.9974τ = 0.235σ, essentially all of the analyte times the residence time. The total analyte counts from a flow-through detector, Cft, is the sum of the counts detected (eq 1) from the beginning of peak elution, tr - 3σ, (t1 in Figure 1), to the end of peak elution, tr + 3σ, (t2 in Figure 1): tr+3σ
Cft =
∑
t)tr-3σ
tr+3σ
Sft(t) )
∑
AEftF(t) ) AEft(0.235σ)
(3)
t)tr-3σ
Radioactivity is a random event, and the statistical uncertainty in measured counts is described by Poisson statistics.16 The standard deviation is the square root of the sum of the counts detected from the analyte plus the background counts during the (15) Braithwaite, A.; Smith, F. J. Chromatographic Methods, 5th ed.; Chapman and Hall: New York, 1996. (16) Knoll, G. F. Radiation Detection and Measurement, 2nd ed.; John Wiley & Sons: New York, 1989.
Analytical Chemistry, Vol. 71, No. 14, July 15, 1999
2701
sampling interval:
σSt ) (C + B)1/2
(4)
Thus, the greater the number of counts, the lower the percent standard deviation. For this reason, the peak area, not peak height, is used for quantitation in chromatographic radioactivity detection. Whole-Column Radioactivity Detection Model. With wholecolumn detection, all of the analyte is detected simultaneously from the time it is loaded onto the column until it elutes (Figure 1). The signal from whole-column detection is the total amount injected on the column at time t past injection minus the total amount which has eluted up to that time. This is different from flow-through detection, where only a portion of analyte is detected at any sampling time. The whole-column signal increases from the start of injection up to t′ (Figure 1). At t > t′ the analyte signal is constant until analyte begins to elute. The signal then decreases until the analyte has completely eluted, at which time the wholecolumn signal is due only to background radiation. We can define the beginning of the whole-column detector signal as time, t′, when the analyte has been fully loaded onto the column (Figure 1). At time t′, the signal from the whole-column detector, Swc(t′ ), is the total radioactivity injected:
Swc(t′ ) ) AFwc(t′ ) Ewc ) AEwc
(5)
The terms are as defined for flow-through detection except wc denotes the whole-column detector. At t′, all of the analyte is within the detector volume and Fwc(t) )1; eq 5 shows that Ewc is the only factor which reduces the efficiency of detection of the decay events from the theoretical maximum, unlike the case of flow-through detection where only part of the analyte is within the detector at any sampling interval, i.e., Fft(t) < 1. At some time after t′, the analyte begins to elute and Fwc(t) becomes equal to 1 minus the total amount of the analyte which has eluted up to that time. The amount which has eluted can be described by a Gaussian function: t
Fwc(t) ) 1 -
∑ t)t0
[
1
σ(2π)1/2
(
exp -
)]
(t - tr)2 2σ2
(6)
At any time t after t′, Swc(t > t′ ) can be determined by substituting eq 6 into eq 5, and the summed total analyte signal during the interval from t′ to tr + 3σ is tr+3σ
Cwc )
∑ t)t′
tr+3σ
Swc(t) ) AEwc
∑F
wc(t)
(7)
t)t′
Comparison of Whole-Column and Flow-Through Radiation Detections for a Single Analyte. A comparison of the signals, S(t), from the two radiation detectors, with equal response factors and peak widths (σ) is depicted in Figure 1. The improvement in signal with whole-column detection is apparent. Equation 7 can be divided by eq 3 and integrated over the time that the analyte is on the whole-column detector to define an area improvement factor, AIF (eq 8). The AIF is the ratio of integral 2702 Analytical Chemistry, Vol. 71, No. 14, July 15, 1999
Figure 2. Simulated whole-column (- - -) and flow-through (s) detector signals for a two-analyte separation. Detector response is equal for the two detectors, Eft ) Ewc ) 0.46 and counts/s is denoted as cps. Analyte retention times are 300 and 660 s. The standard deviation of the peak width (σC) ) 30. The flow-through detector residence time (τft) ) 0.235σ ) 7.0 s. The vertical dashed lines are the lower integration times for the whole-column detector for Rs ) 1, 2, or 3.
analyte counts of the whole-column detector to those of the flowthrough detector. t2
AIF )
Cwc Cft
∑F
wc(t)
AEwc )
t1
(8)
AEft(0.235σ)
Using eq 8 as depicted in Figure 1, the AIF is 34 when k′ ) 4. If k′ is reduced to 1.5, an AIF of 13 is obtained. Thus, chromatographic conditions can be chosen to preferentially increase the whole-column signal and the AIF by increasing analyte retention. Whole-Column Radiation Detection for Separations of More Than One Analyte. Chromatographic separation theory can be used to determine an upper limit for the ratio of the integral count areas from the two types of detectors based on chromatographic resolution rather than a retention factor alone. Chromatographic resolution, Rs, describes the separation between two analyte peaks and is defined in eq 9, where w1 and w2 are the 4σ
Rs )
tr2 - tr1 2d ) w1 + w2 2(σ1 + σ2)
(9)
widths of two analyte peaks, d is the difference in mean retention time of the two peaks, and tr1 and tr2 are the retention times of analytes 1 and 2, respectively. We will consider the signal from analyte 2. If the peaks are baseline resolved, then the wholecolumn signal can be integrated from the time the first analyte has completely eluted to the end of elution of the second analyte, i.e., tr1 + 3σ to tr2 + 3σ. If the chromatographic resolution is less than 1.5, the peaks are not baseline resolved. When Rs ) 1, tr1 + 2σ ) tr2 + 2σ. If one is willing to accept
