Optical sensors in flow injection analysis

133

Journal of Molecular Structure, 292 (1993) 133-140 Elsevier Science Publishers B.V., Amsterdam

Optical Sensors in Flow Injection Analysis Otto S. Wolfbeis Institute of Organic Chemistry, Analytical Division, Karl-Franzens University, A - 8010 Graz, Austria.

Heinrich St. 28,

Representative examples are given on how optical chemical sensors (optrodes) can be coupled to flow injection systems to result in flow injection analyzers. These have served two main purposes so far, namely testing the performance of opt&es, and secondly as detectors in flow injection analysis (FIA). Specifically, the use of optrodes sensitive to pH, oxygen, ammonia and other chemical species as detectors in FIA will be described, all mainly in conjunction with enzymatic reaction schemes. Finally, optrodes are shown to be useful for determination of enzyme activity and enzyme inhibition by heavy metals. 1. INTRODUCTION Rarely has an analytical method so readily been accepted as has been flow injection analysis (FIA). Competent reviews and books are available and cover the subject from various points of view [1,2]. The great success results from the fact that it is compatible with almost any analytical detection scheme and lends almost any analytical method to serial assay. After optical sensor development has come to some maturity [3,4] it was recognized rather early that a combination of the optrode and FL4 technologies can overcome many of the difficulties encountered with optrodes, in particular with respect to recalibration, d.ri& varying sample pH and temperature, and automatization. A typical arrangement for a FIA system with fiber optic sensor detection is shown in Fig. 1. FL4 methods in combination with optrodes nowaday serve two main purposes. The one comprises the testing of the performance of sensors (not necessarily optrodes). Ruzicka and Hansen [l] refer to this as “an impulse-response technique, i. e., one that is based on repetitive action of a well defined zone of a selected chemical species (the reagent) on a target (the sensor) situated in a continuous, unsegmented carrier stream”. The second purpose is to perform conventional chemical analysis, using optrodes as detectors in FIA. Because optical sensors for oxygen, pH, 0022-2860/93/$06.00

0 1993 Elsevier Science Publishers B.V.

Fig. 1. Schematic of a FIA manifold with fiber optic detection. Lit from a source (IS) is guided through one bundle of the bifurcated fiber optic (BFB) to the optical sensor membrane contained in the flow-through cell (FTC). Reflected or fluorescent light is guided back to the detector (PMT) and the resulting electrical signal is recorded in R. The carrier stream (C) is propelled with a pump (P). The sample (S) is introduced by means of an injection valve (V). After passing a mixing chamber (MC), the stream passes an air damper (AD) and analyte concentrations are detected in the FTC. In case of biosensors, au enzyme bed reactor is frequently placed between V and MC, or MC and AD. All rights reserved.

134

carbon dioxide, ammonia, and certain monovalent cations are most advanced, they have been employed most frequently, both for these species, but also as transducers for monitoring biochemical reactions. In this report I want to give selected examples on the progress that has been made in the PIA-optrode hyphenation so far. 2. FLOW INJECTION SYSTEMS FOR TESTING THE PERFORMANCE OF OPTRODES The application of flow injection (FI) systems as a diagnostic tool for development and testing of chemical sensors has been suggested first by Ruzickas groups [5,6] in a need of frequent recalibration of a sensor, and this could easily be provided by an FI system. Specifically, sensors for pH and urea have been designed and characterized in terms of sensor stability (both dye and enzyme stability), response time, sensitivity, reproducibility, response to a typical interferent, and sensor lifetime.

Weigl [7] reports on a FI system specifically designed for testing the performance of optrodesusing the software package FLOWLAB. Inessence, it is a FI system run under an AT PC. Sensors can be calibrated with standard solutions. Fit routines generate functions used for conversion of light intensity raw data into analytically useful parameters such as PH. The system is most suitable for testing the performance of sensors. Figs. 2 and 3 show the response of a “good” and a “bad” sensor, respectively. The one in Fig. 3 displays a substantial drift that makes it almost useless for practical applications i

16

0

10

20

30

40 time (min)

Fig. 2. Typical recording of a PIA/ optrode combination for measurement of PH. Samples of varying pH (5.0, 5.5, 6.0, 6.5, 7.0, 7.5, 8.0, 8.5, 9.0, 10.0, 11.0 and 4.0) were injected into a 20 mM carrier buffer of pH 7.0 and the resulting change in pH detected using an absorbance based pH sensing membrane placed in the flow-through cell (Fig. 1).

!

i )

li 32

fI L 8

time(mitt)

Fig. 3. Response of a FIA system for assay of nitrate using a poorly performing nitrate-sensitive optrode as the detector. While the sensor shows a distinct response to injections of 0.3, 1.0, 10.0 and 100.0 mM solutions of potassium nitrate into the pH 3.0 carrier buffer, the dye (Nile Blue) obviously bleaches and, possibly, the nitrate carrier (an organic quaternary ammonium ion) leaches out. As a result, the signal drops from 0.0395 to 0.0313 within &I min. For a more stable nitrate sensor, see ref. [S].

135

3. OPTICAL INJECTION SPECIES

(BIO)SENSOBS IN FLOW OF CHEMICAL ANALYSIS

3.1. General Aspects This chapter gives selected examples on the use of optical @io)sensors in combinations with FIA. It does not cover classical optical detection in FIA when, for instance, a chromogenic or fluorogenic reagent is added to the carrier stream. Fluorescence detection in FIA has recently been reviewed [9]. Rather, a sensor material, usually placed at the end of an optical fiber, specifically recognizes the analyte that has been formed in a reaction coil or, in the simplest case, has been injected into the carrier stream. In other words, the optical properties of the sensor are being detected, not those of the analyte or those of any product of a chemical reaction with an injected reagent. This approach has several attractive features: a) No chromogenic or fluorogenic reagents have to be added; b) The effect of interferents, contained in the analyte solution and adversely affecting the accuracy of the determination, can be minimized, c) Ionic strength, buffer capacity, pH, oxygen partial pressure etc. (all of which are particularly critical in case of biosensing) can be kept constant; d) Samples can be diluted in order to adjust their concentration to the dynamic range of the sensor; e) Sensor probes may be re-activated by addition of other reagents, or re-conditioned by maintaining the sensor in a controlled environment for an acceptable period of time. Notwithstanding these advantages, there are certain drawbacks. These include the following: a) Good FL4 instrumentation is comparatively expensive and rather sophisticated; b) Real time measurements with FL4 devices are difficult to perform. Some FL4 systems give but quasi-online results provided sampling and sample transport are quick and the detector optrode has a rapid response. The big advantage of fiber optical sensors, namely direct contact between sample

solution and sensor, is lost when use is made of a FI system into which the sample has first to be injected. c) Unlike in case of injected reagents, the indicators used in optrodes are the same during the whole lifetime of a sensor; this can cause problems with respect to dye leaching, leaching of other sensor material components, photobleaching, membrane swelling, and hence may require frequent recalibration of the sensor. d) Some optrodes have a response time too long to be useful in FL4 with its sample frequencies of 30 per hour or even more. While this article is mainly concerned with the use of optical (bio)sensors in FIA, the reader is referred to two monographs [lO,ll] which, in addition to what can be found in refs. [l] and [2], give an excellent overview on current activities in the area of FIA using all kinds of sensors, in particular electrochemical ones. 32. pH Optrodes in FL4 FL4 systems for on-line measurement of pH have been presented by Yerian et al. [6] and by Weigl [7l (see Fig. 1). The pH optrodes also have been coupled to enzyme based FL4 of urea [6] in which the increase in pH as a result of enzymatic hydrolysis of urea according to urea + 2H20

+ Ht ==> 2NH4++

HCO,-

(1)

is exploited. In a later version [12], urease was co-immobilized with BSA onto the surface of a pH sensor. By applying the stopped-flow technique, a dynamic range up to 10 mM urea was achieved. Busch et al. [l3] report on the application of optrodes in F&based fermentation process control. A conventional pH optrode measures the change in pH caused by enzymatic hydrolysis of urea. They use a software package (FIACRE) which, like FLGWLAB (see section 2), governs sampling, injection, and transport, and also acquires and handles analytical data,

136

33. Qxygen Optrodes in FIA Oxygen as such has not been assayed via PIA with optrode detection OS far. Rather, oxygen optrodes have been applied as transducers for quantitation of biochemical reactions during which oxygen is consumed. Thus, a glucose assay has been described more recently [14] in which a fast responding fiber optic glucose biosensor was used. It has a response time of about 10 s and a dynamic range from 0.01 to 2.0 mM glucose. To increase the analytical range, a I?IA system allowing the performance of the zone sampling technique has been used. Variable dilution factors can be achieved simply by changing the time span between the first and the second injection. The schematic of the PIA arrangement is shown in Pii. 4, and the corresponding calibration curve in Pii. 5. Glucose determinations in wine and fruit juices have been performed. The results obtained with the PIA compared to results obtained by an official food laboratory were excellent.

W I

c2

AD

S

57 -

so

100

150

glucose concentration

200

250

300

(mM)

Fig. 5. Calibration graph of the glucose mA system of I?@. 4, using an oxygen optrode as the detector. Quite obviously, the slope of the calibration graph can be adjusted by changing the zone sampling time.

W

FTCtW

Cl ,p

oxidase, biosensors (with oxygen transducers) were obtained for determination of lactate [l5] and ascorbic acid [16]. In both cases analyses in were performed with “real” samples, namely dairy products and fruit juices, respectively.

LS PM1 W

Fig. 4. PIA arrangement for zone sampling. (a) Optical part: IS, light source; BPB, bifurcated fiber bundle; FTC, flow through cell with the fiber optic glucose biosensor; PMT, photodetector, and R, recorder. (b) mA part: Cl, and C2, carrier streams; S, sample; Vl, and V2, valves; MC, mixing chamber; AD, air damper, P, pumps. By immobiig lactate oxygenase and ascorbic acid oxidase, respectively, in place of glucose

3.4. Ammonium Sensors in FIA An ammonium-sensitive optical sensor membrane based on the use of a neutral ammonium carrier (nonactin) has been employed [lq as the detector in FIA (Fig. 6). The linear response is from 0.1 to 4.0 mM urea. The sensor has been utilized for FIA assay of urea in serum. Urea is enzymatically hydrolyzed according to (1) when passing an urease reactor. The liberated ammonium ion rather than the change in pH (see above) is detected. An assay requires 200 1 of a serum sample solution only, and 20 sample per hour can be run. Some thirty serum sample were analyzed with this system, and results were compared with those obtained with a commercial clinical analyzer. The maximal deviations found were in the order of lo%, the correlation coefficient is 0.993.

137

3.5. Other Sensors in FIA A penicillin-sensitive optrode sensor not based on the use of an enzyme has been described recently [18] and was used, in conjunction with a flow system, for determination of penicillin G in pharmaceutical formulations. Recognition of penicillin relies on the use of an anion carrier which transports the penicillin anion from the aqueous sample into the membrane where it causes an optical change. The sensor fully reversibly responds to penicillin V over the 0.01 to 10 mM concentration range, and to penicillin G from 0.03 to 10 mM. Potential interferences by about 20 other anions have been investigated. Only nitrate, salicylate, and ascorbate were found to interfere significantly. These species are, however, usually not present in penicillme bioreactors or drug formulations where penicillme sensing is most important.

two kinds of gIutamate assays. In the first, an oxygen optrode was covered with a layer of immobii L-glutamate decarboxylase, and the decrease in oxygen partial pressure measured as a result of enzymatic oxidation of glutamate. In the second, a carbon dioxide optrode was covered with L-glutamate decarboxylase. It was concluded that the oxygen optrode showed a much better performance and even that, in general, oxygen transducers are to be preferred over carbon dioxide transducers because of their sensitivity, quick response, and inertness to interfering organic acids or acidic gases. a

__---______-_______ ____-____________ ______-_________

I

III

“11111 / _-_________

_-_---

_ __________

IA--w

_-____-_-_______

Id*

1

ulloyl

__-_---_-~---__-~

\

l@lryl

1

-log

0-c 4 3 -loglNh+l WI

2

Fig. 6. PIA assay of ammonium ion using, as the detector, an ammonium-sensitive optrode composed of PVC, plasticizer (dioctyl sebacate), nonactin (an ammonium carrier) and di-octadecyl rhodamine as the fluorescent proton carrier. From ref. [17]. Similar systems have been designed for assay of thiamine (vitamin Bl) [19] and aspirin and salicylic acid [20]. A typical recorder output along with the repective calibration graph is shown in Fii. 7. The I%% of salicylate using an aspirin/salicylate sensor has been applied for analysis of aspirin in various tablets. Dremel et al. [21] have compared the results of

Isa&ate] (mol/l)

ii

io

do

do

rib 160

time (mm)

Fii 7. Response and calibration curve of the salicylate-sensitive membrane in a flow injection analyzer toward triple injections of varying concentrations of salicylate at pH 5.30. From ref.

PI* Interferences by organic acids in CO, sensor are most obvious from the data obtained in the FIA/optrode analysis of glutamate in various samples (Table 1). The CO2 sensor is composed of buffer/dye system entrapped in a silicone polymer which is permeable to uncharged species only, but not to ions such as the proton or carboxylates. As the pH is lowered, a substantial fraction of the organic fatty acids (with their pKas around 4 - 5) will become protonated and, thus, can penetrate the silicone. The acids lower the pH of the internal buffer of the sensor and falsely indicate a high CO, partial pressure.

138

T46&I.CompclriponofthcnsuuSfotdctaminationofLgi[llrcuncztcuring3different m&o&;allwncenrma in mhf. Fnnn trj 1211. Method

Sample

Ma&Q+ flavor Mushroomsoup

GlChP’

GlDCb)

HPL@

340.0 1.8

10m.o

310.0 L7 1.4 05 24

Mexicansauce

1.6

soup cube

05

NCuroglutamatC~

2.2

4.7

38 4.0 52

a) glutamate oxid24scimmobilized on an oxygen optrode; b) glutamate decarboxylase immobilized on a -bon dioxide optrode, c) detection via tluorescence derkakation of glutamate with o-phthalaldehyde.

4.DETERMINATION

OF ENZYME ACTIVITY AND ENZYME INHIBITION USING FIA 4.1. Determination of Enzyme Activity A combination of a RI system with immobiid urease and an ammonium optrode was used to determine the activity of urease. Enzymatic hydrolysis of urea at pH 7.14 yields ammonium ions which is quantitated using an optical fibre ammonium sensor [22]. Urea is contained in the carrier stream into which samples of varying urease activity are injected. The concentration of the ammonium ions, formed as a result of enzymatic hydrolysis of urea according to (l), is measured with an ammonium fluorosensor, and is a direct indication of enzyme activity. The system measures activities over the 1 to 30 units/mg range. Fig. 8 shows the decrease in relative signal change with increasing quantities of injected urease activity. Recently, a related paper has appeared in which the activity of certain dehydrogenases is determined using FL4 [23].

4.2. Determination of Heavy Metals Via Inhibition Urease is sensitive to traces of heavy Silver(I), mercury(II), copper( cadm cobalt(H), nickel(H), manganese(I1) and have all been quantified by using data fi inhibition of the urea/urease reaction by th The urease inhibition is relatively selec mercury(H) [24], with a selectivity of greater) over the other. Yeriatl et al. [6] report on two b approaches for determination of heavy m the first, immobilized urease was exp various levels of mercury(I1) contaminatic carrier stream (without EDTA), and the I to a 10mM urea standard was monitored levels of mercury(I1) (0.2 - 2.0 M), injectic mM urea showed an exponential decrease height. The other mode of sample mc would be alternate injections of urea stanc mercury samples. This, however, \k investigated further by the authors. Con1 both methods are suitable for continuous mercury in environmental samples.

139

5. CONCLUSION

REFERENCES

We think to have demonstrated that FIA with optrode detection can result in a very powerful analytical technique that has a broad range of applications, some of which are not fully exploited yet. The optrode detector can be specific, reliable and can simplify the experimental arrangement. The FI arrangement, in turn, provides the benign environment that is required in case of enzymatic analysis where constant pH, ionic strength, and other parameters are essential for proper enzyme function. The FI system, in addition, provides the possibility of frequent sensor calibration. Therefore, the FIA/optrode hyphenation indeed looks iike an almost ideal combination.

1 16

64

time (min)

Fig. 8. Determination of enzyme activity by FIA using an ammonium optrode as the transducer. Urease solutions (with specific activities of 0.3,1.0, 3.0, 10.0 and 30.0 units per ml) were injected (in triplicate) into a carrier stream (10 mM urea solution of pH 7.14), and the formation of ammonium ion was measured with an ammonium sensor membrane.

ACKNOWLEDGEMENT This work was supported by the Austrian Science Foundation (FWF) within project S5702-TEC which is gratefully acknowledged.

1. J. Ruzicka and E. H. Hansen. Flow Injection Anaiysis. 2”* ed., Wiley, New York, 1989. 2. M. Vaicarcel and M. D. Luque de Castro. Flow Injection Analysis: Principles and Applications. EIlis Harwood, Chichester, UK, 1990. 3. W. R. Seitz, Chemical sensors based on immobilized indicators and fiber optics, CRC Crit. Rev. Anal. Chem. 19 (1988) 135. 4. 0. S. WoIfbeis (ed.). Fiber Optic Chemical Sensors and Biosensors. CRC Press, Boca Raton, Florida, 1991. 5. R. Singer. Applications of Immobilized Indicators for Optosensing by Flow Injection Anaiysis. MSc Thesis, TU Denmark, 1986. 6. T. D. Yerian, G. D. Christian and J. Ruzicka. Fiow injection analysis as a diagnostic technique for development and testing of chemical sensors. Anal. Chim. Acta. 204 (1988) 7. 7. B. Weigl. A simple and versatile method for testing the performance of optical chemical sensors. In: ChemicaI, Biochemical, and Environmental Fiber Sensors III, R. A. Lieberman (ed.), Proc. SPIE 1587,288 - 295 (1991). 8.0. S. Wohbeis, Optical sensing based on analyte recognition by enzymes, carriers and molecular interactions”, Anal. Chim. Acta 250 (1991) 181. 9. M. Valcarcel and M. D. Luque de Castro. Fluorescence Detection in Flow Injection Analysis. In: Fhrorescence Spectroscopy: New Methods and Applications, 0. S. Wohbeis (ed.), Springer, Berlin, 1992. 10. R. D. S&mid (ed.). FIow Injection Analysis (FIA) Based on Enzymes or Antibodies. GBF Monographs, ~0114, VCH Publ., Weinheim, FRG, 1991. 11. R. D. S&mid and F. Scheiier (eds.). Technologies and Biosensors: Fundamentals, Applications. GBF Monographs, vol. 17, VCH Publ., Weinheim, FRG, 1992. 12. T. D. Yerian, G. D. Christian and J. Ruzicka. Enzymatic determination of urea in water and serum by optosensing flow injection analysis.

140

Analyst 111 (1988) 865. 13. M. Busch, F. Gutberlet, W. H”obel, J. Polster, H. L. Schmidt and M. Schwenk. The application of optodes in FIA-based fermentation process control using the software package FIACRE. Sens. Actuators, part B (1993) in press. 14. B. A. A. Dremel, B. P. H. Schaffar and R. D. Schmid. Determination of glucose in wine and fruit juice based on a fibre-optic glucose biosensor and flow injection analysis. Anal. Chim. Acta 225 (1989) 293. 15. B. A. A. Dremel, G. Trott-Kriegeskorte, B. P. H. Schaffar and R. D. S&mid. Llactic acid determination in milk products based on a fibre optic biosensor and flow injection analysis (FJA). Biosensors, Application in Medicine, In: Environmental Protection and Process Control, GBF Monographs, vol. 13, R. D. S&mid and F. Scheller (eds.) VCH Publ., Weinheim, FRG, 1989, p. 225. 16. B. P. H. Schaffar, B. A. A. Dremel and R. D. S&mid (1989). Ascorbic acid determination in fruit juices based on a fibre optic ascorbic acid and flow injection biosensor analysis. In: Application in Medicine, Biosensors, Environmental Protection and Process Control, GBF Monographs, vol. 13, R. D. Schmid and F. Scheller (eds.) VCH Publ., Weinheim, FRG, 1989, p. 229. 17. H. Li. PhD thesis. University Graz, 1992. 18. H. He, H. Li, G. Uray and 0. S. Wolfbeis. Non-enzymatic optical sensor for penicillins, Talanta (1992) in press. 19. H. He, G. Uray, and 0. S. Wolfbeis. A thiamin-selective optrode. Anal. Lctt. 25, (1992) 405. 20. H. He, G. Uray, and 0. S. Wolfbeis. Optical sensor for salicylic acid and aspirin based on a new lipophilic carrier for aromatic carboxylic acids. Fresenius J. Anal. Chem. 343 (1992) 313. 21. B. A. A. Dremel, R. D. Schmid and 0. S. Wolfbeis. Comparison of two fibre optic L-glutamate biosensors based on the detection of oxygen or carbon dioxide, and their application in

combination with flow-injection analysis to the determination of glutamate. Anal. Chim. Acta 248 (1991) 351. 22. H. Li and 0. S. Wolfbeii. Unpublished results, 1992. 23. M. D. Luque de Castro and J. M. Fernandez-Romero. Total and individual determination of creatinine lcinase iso-enzyme by flow injection and liquid activities chrom$ography, Anal. Chim. Acta 263 (1992) 43. 24. L. Ogren and G. Johansson. Determination of traces of mercury(E) by inhibition of an enzyme reactor electrode loaded with immobilized urease. Anal. Chim. Acta 96 (1978) 1.