1246
Anal. Chem. 1987, 59, 1246-1248
LITERATURE CITED
number of Stern-Volmer determinations. In addition, with a small amount of added software and valve-selectable reagent bottles, it should be a simple matter to automate this system for a number of different fluorophores and quenchers during an overnight or weekend period. Hazardous or sensitive compounds are more readily handled in such a closed system than in open cuvettes. In principle, less expensive equipment such as low-pressure metering pumps and/or an inexpensive flow-through cell mounted in a conventional fluorometer could be used. Fragile biomolecules could also be examined with a peristaltic pump and an efficient gradient maker.
(1) Alwattar, Abbas H.; Lurnb, Michael D.; Birks, John B. I n Organic Molecular Photophysics; Birks, J. B., Ed.: Wiley: London, 1973; Vol. l , Chapter 8. (2) Turro, Nicholas J. Modern Molecular Photochemistry: Benjarnin/Cumrnings: Menlo Park, CA, 1978; p 248. (3) Wolfbeis, 0. S.; Urbano, E. Anal. Chem. 1983, 5 5 , 1904-1906. (4) Guilbault. George G. I n fluorescence : Theory, Instrumentation,and Practice; Guilbault, Georae S., Ed.: Marcel Dekker: New York, 1967: Chapter 8. (5) Cukier, R . I. J . Am. Chem. SOC. 1985, 107, 4115-4117. (6)Lakowicz, J. R. Principles of fluorescence Spectroscopy; Pienurn: New York, 1983; p 279. (7) Peak, D.; Werner, T. C.; Dennin, R. M.,Jr.; Baird, J. K. J. Chem. Phys . 1983, 7 9 , 3328-3335.
ACKNOWLEDGMENT The authors gratefully acknowledge Griff Freeman, Wes Jacobs, Terry Miller, Steve Cook, and Harry Morrison for helpful discussion.
RECEIVED for review August 1, 1986. Accepted December 15, 1986. This work was supported in part by the Indiana Elks and the National Science Foundation, Grant No. CHE8320158.
Applications of a Versatile Injection Valve for Flow Injection Analysis Brice C. Erickson and Bruce R. Kowalski* Department of Chemistry, University of Washington, Seattle, Washington 98195
Jaromir Ruzicka Chemistry Department A, T h e Technical University of Denmark, 2800 Lyngby, Denmark The concept of flow injection analysis (FIA) as originally developed by Ruzicka and Hansen ( 1 ) involves the injection of a sample plug into a flowing stream of reagent. The sample disperses into the reagent, and reacts to form a product which is observed a t the detector. Over the years many variations on this basic method have been developed (2). In the process, several valve and injection systems have been created to allow the various combinations of injections of sample and reagents necessary for these techniques. We wish to bring to the attention of researchers in the field several configurations of a very simple, inexpensive, and versatile valve which may be useful in several of these techniques. The most commonly used valve in FIA is the six-port rotory valve diagramed in Figure 1. The sample loop is filled in the “load” position and then placed in the reagent stream when turned to the “inject” position. The valve design used here is similar, having six ports on the front rotor. But rather than simply connecting adjacent pairs of these ports on the back plate, all six also are individually accessible giving a total of 12 ports which provide a wide range of options in its configuration. The valve shown in Figure 2 is constructed by bolting a solid Teflon cylinder (1.75 cm diameter and 1 cm thick) with six evenly spaced holes on a PVC plate 1 cm thick. Matching holes are made through the plate, and channels are imprinted on the bottom of the plate to additional holes away from the rotating disk. For convenience, additional channels may be imprinted on the same plate to be used for reaction coils, merging zones, sequential reagent additions, etc. ( 3 ) . Short pieces of tubing are inserted into the exposed holes on the top of the plate and in the disk. Connections are made by simply fitting with sleeves of slightly larger tubing. The bottom of the plate is sealed, leaving the impressed channels open for flow. A lever on the disk allows it to be turned from one position to the other, with stops on the plate assuring that the holes in the disk and plate are aligned in either the LOAD or INJECT position. The valve is normally operated manually
but has also been automated. Sample and/or reagent loops may be filled manually by syringe, by aspiration, or by peristaltic pump (usually by drawing on the waste line); if the pump used to fill the loop(s) is separate from the carrier pump (which runs constantly), sample or reagent may be conserved by running it only when the valve is in the LOAD position and stopping it in the INJECT position. The standard six-port arrangement shown in Figure 1 can be formed by simply connecting pairs of ports on the rotor (or on the plate). Alternatively for a simple injection the valve can be configured as shown in Figure 3. While FIA methods greatly reduced reagent consumption in comparison to most previous methods, it was discovered that a further reduction could be accomplished by simultaneously injecting reagent and sample and then merging the two streams as shown by Mindegaard ( 4 ) and Bergamin et al. ( 5 ) . Toei and Baba (6) have also designed a valve that is capable of either simultaneous sample and reagent injection or simple sample injection. The valve configured as in Figure 4 allows simultaneous sample/reagent injection after filling the two loops. Note that the two carrier inlets may be propelled by a single pump tube by splitting the line from the pump as in the Toei design or by separate pump tubes, which is preferable since flow rates through both loops do not need to be hydrodynamically balanced in order to assure reproducible injections-as is the case in Toei’s design. Rios and co-workers (7) recently proposed a two-valve system that allows injection of a plug of reagent (at a different pH than the reagent stream) in the center of the sample plug. This allows more sensitive multiple analyte detection of the pH-sensitive products formed in the resulting pH gradient by the method originally proposed by Betteridge and Fields (8). We propose that this type of injection can also be accomplished with a single valve configured as shown in Figure 5 . A second application of this configuration involves reversing sample and reagent lines: a plug of sample can be
0003-2700/87/0359-1246$01.50/0 1987 American Chemical Society
ANALYTICAL CHEMISTRY, VOL. 59. NO. 8, APRIL 15, 1987
LOAD
1247
INJECT
Si-@ injectwn valve. The v a b configurations are shown in LOAD and INJECT positions. Lower diagram shows the resuning injection in the reagent stream. Figure 1.
B
S
Canier
Reanent
Carrier
Flgure 5. Gradient injection. (A) Reagent of different pH than reagent stream is Injected between two plugs of sample. This assures adequate reagent supply throughout the resulting pH gradient. L1and b2 are sample loops. (E) With the sample and reagent loops reversed. sample c a n be injectad between two plugs of reagent in an unreactive carrier, minimizing reagent consumption in a single-line system. Reagent LOOP
Reagent Loop
To Detector,
INJECT
LOAD
LOAD
Reagent
I
~~
Flgure 2. Valve construction. See description in text To Detector,
I Reagent I Sample I
INJECT
Sample
A
Reagent (highpH)
B
Carrier
c
Carrier
Reagent
I
Reagenl (lowpH)
Standard
I0mf-t
I
I
Rage", (highpH)
Sample
Carrier
SmQIC
Carrier
LOAD
Flgun 4. Simultaneous injection. Sample and reagent are injected simunaneovsly into separate streams which are merged downsheam. L, is the sample loop.
metrical double gradient when the reagent is injected in the center of the sample zone as in Figure 5A. The same configuration might also he used for the gradient scanning standards addition method proposed by Fang et al. (9) hy replacing the reagent in the reagent loop with a standard solution of the analyte. By injection of a sample plug immediately followed by a plug of standard into the carrier
1248
ANALYTICAL CHEMISTRY, VOL. 59, NO. 8, APRIL 15, 1987 Reagent Loop
Reagent Loop
LOAD
INJECT
Reagent
A
Sample
Reagent
Sample
Reagent
Figure 7. Double sample injection. Two plugs of the same sample
in LS1 and L,, are injected either (A) with different sample volumes or (B) with plugs separated by a reaction column.
,@, L
Salnple Reagent Loop
To Detector LOAD
Reagent
which may merge with reagent downstream, again allowing observation of response a t times corresponding to various concentration ratios (IO). Another configuration is shown in Figure 7 , which results in a double injection of sample separated by a plug of the reagent stream which can result in two separate peaks. By use of sample loops of different volumes, the dynamic range of the system can be increased by calibrating high concentrations from the smaller injection volume and low concentrations by the larger injection volume. A fiagram identical to that obtained with a split stream manifold (11) will be obtained, but since the present method does not rely on hydrodynamic balance of split streams, reproducible results over extended dynamic range and time periods will be observed. The reagent loop can also be replaced by a packed reactor. An example of the use of this configuration is the analysis of nitrate and nitrite. The reagent loop is replaced by a copper/cadmium reaction column BS used by Anderson (12). Two loops are filled with the sample. When injected, the first sample passes directly to the reaction manifold where it is merged with reagent and nitrite is detected. The second, however, first passes through the copperized cadmium reduction column converting nitrate to nitrite, allowing measurement of total nitrate plus nitrite. Nitrate concentration can then be calculated by difference. Note that each peak position must be calibrated separately, since the second peak travels a greater distance causing additional dispersion. Figure 8 shows a similar arrangement which allows injection of two separate samples. It may also be used for injections similar to those shown in Figure 6 but where a small plug of the carrier is placed between the sample and reagent, standard, or interferent.
LITERATURE CITED
INJECT
1 Sample 1 I Reagent I Sample 2 I
(1) (2) (3) (4) (5)
Reagent
Figure 8. Double injection. Two separate loops LS1 and L,, are filled
with distinct samples (similar to Figure 7). Alternatively, sample in Ls2 may be injected with reagent, standard, or interferent (similar to Figure 6) in L,, separated by a small plug of the carrier. stream (and then merging with a reagent stream if necessary), response can be observed at times corresponding to various sample/standard concentration ratios. This valve configuration can also be used for interference studies where the sample plug is immediately followed by a plug of potential interferent. The interferent partially disperses into the sample plug while both mutually disperse into the carrier stream,
(6) (7)
(8) (9) (10) (11) (12)
Ruzicka, J.; Hansen. E. H. Anal. Chlm. Acfa 1975, 78, 145-157. Ruzicka, J.; Hansen, E. H. Anal. Chlm. Acta 1986, 779, 1-58. Ruzicka, J. Anal. Chem. 1983, 55, 1040A-1053A. Mindegaard, J.; Anal. Chim. Act8 1979, 104, 185-189. Bergamin F., H.: Zagatto. E. A. G.; Krug, F. J.; Reis, B. F Anal. Chim. Acta 1878, 101, 17-23. Toei, J.; Baba, N. Anal. Chem. 1986, 58, 2348-2350. Rios, A.; Luque de Castro, M D.; Valcarcel, M. Anal Chem 1986, 58,663-664. Betteridge, D ; Fields, B. Anal. Chem. 1978, 5 0 , 854-656. Zhaolun, F.: Harris, J. M.; Ruzicka, J.; Hansen, E. H. Anal. Chem. 1985, 57, 1457-1461. Hansen, E. H.; Ruzicka, J.; Krug, F. J.; Zagatto, E. A. G. Anal. Chim. Acta 1983, 148, 111-125. Ruzicka, J.; Stewart, J. W. B.:Zagatto, E A Anal. Chim. Acta 1976, 81, 387-396. Anderson, L. Anal. Chim. Acta 1979, 770, 123-128.
RECEIVED for review December 1, 1986. Accepted January 13, 1987. This research was supported in part by the Office of Naval Research.
CORRECTION Sampling and Excitation of Refractory Solids with a Theta Pinch Discharge Designed as aa Atomic Emission Source Jeffrey S. White and Alexander Scheeline (Anal. Chem. 1987,59, 305-309). On page 309, ref 2 should read as follows:
Carr, J. W.; Horlick, G. Spectrochim. Acta, Part B 1982, 37B(1), 1-15.
