You are on page 1of 9

951686

Chemiluminescent Ethanol Biosensor Development


James E. Atwater, James R. Akse, Jeffrey DeHart, and Richard R. Wheeler, Jr.
Umpqua Research Co.

Charles E. Verostko
NASA-Johnson Space Center

ABSTRACf The investigation and development of a chemiluminescence based ethanol detection concept into a biosensor system is descn1>ed. The biosensor uses alcohol oxidase to catalyze the reaction of short chain primary alcohols with elemental oxygen to produce hydrogen peroxide and the corresponding aldehyde. The reaction of hydrogen peroxide with an organic luminophore in the presence of a sufficient electric field results in emission of blue light with peak intensity at 42Snm. The chemiluminescent light intensity is directly proportional to the alcohol concentration of the sample. The aqueous phase chemistIy required for sensor operation is implemented using solid phase modules which adjust the pH of the influent stream, catalyze the oxidation of alcohol, provide the controlled addition of the luminophore to the flowing aqueous stream, and minimi7e the requirement for expendables. Precise control of the pH bas proven essential for the long-term sustained release of the luminophore. Electrocatalysis is achieved using a controlled potential across gold mesh and gold foil electrodes which undergo periodic polarity reversals. Ethanol concentrations (as C) in the range between 40 and 4.000 J1gILhave been determined. The performance features of this sensor are presented in this paper. INTRODUCTION For aqueous waste streams generated by a crew in regenerative life SUPPOrt. thanol is a major impurity. It is e poorly removed by physicochemical processes such as multifiltration (MF) and reverse osmosis (RO). and consequently forms a significant fraction of the efi1uent total organic carbon (fOC) from these processes. The inability of sorption and membrane based separations to remove ethanol and other low molecular weight water soluble organics bas led to the evaluation of alternative water reclamation methods such as electrochemical water recoveryl.2. and to the development of post-treatment technologies for RO and MF

systems such as low temperature aqueous phase catalytic oxidation3.4.and immobilized enzyme bioreactocs>7. The capability for quantitation of ethanol in waste water streams such as humidity condensate and in finished potable water can be used advantageously in the control of water processor operation. The initial chemiluminescent ethanol biosensor development effort was intended to demonstrate the feasi.oility of an accurate. precise. sensitive. and reliable means for the determination of ethanol and other low molecular weight straight chain alcohols within aqueous streams aboard spacecraft. A unique aspect of the ethanol biosensor operation is that all reagents are provided by solid state flowthrough modules. thus minimhing the storage and resupply logistic requirements for long term operation in earth orbit or on station within a Lunar or planetary outpost. Operation of the biosensor is based upon the oxidation of ethanol to acetaldehyde. and reduction of <>2to hydrogen peroxide ~Oi' which then reacts with a luminophore to produce light. The intensity of the light is directly proportional to the concentration of ethanol in the sample. Advantages inherent to the technique are the potential for detection of extremely low levels. the stability of the devices used to measure light intensity (photomultiplier tubes or charge coupled devices). and the use of solid state modules for the controlled release of reagents. This detection scheme may be conveniently embodied in either continuous on-line or flow injection analysis configurations. Ethanol is detected using the enzyme alcohol oxidase. Alcohol oxidase catalyzes the reaction of ethanol with molecular oxygen to form acetaldehyde and ~O2' H

I ~H

AIcchoI Oxida-.

'-./

~~

~c ~H
0

II

(1) The ~O2 containing aqueous stream flows through a solid phase basification module where the :l1hlinity is adjusted to pH II:: 0. The basified efi1uentflows into a bed of 1

luminol (3-aminophthalhydrazide) crystals which slowly dissolve at a controlled rate. The pre-treated solution containing all reagents necessary for the quantitative reaction, then flows into an Electrocatalyzed Chemiluminescence (ECL) cell. A controlled electric potential across the anode and cathode catalyzes the oxidation ofluminol by Hz2. 0 ~ Q

switching frequencies were synthesized using a Wavetek model 142 function generator.

ENZYME IMMOBJLlZATION Enzymes were immobilizedon titaniumactivateddiatomaceousearth using ethylene diamine and glutaraldehyde5-? s illustrated in a Figure 1.

0 Ni:! 0 (2) Light is emitted as a by-product of the reaction. The light is detected and quantified by a photomultiplier tube (PMT). The intensity of the light emission is directly proportional to the ethanol concentration in the sample. By using an electrical potential across the ECL cell to catalyze the reaction of hydrogen peroxide with luminol, the reaction is confined spatially, since the excited metastable state of the reaction product has a short halflife9-ll. Light can be gathered efficiently by coupling the PMT directly to the ECL cell, or by using fiber optics as a waveguide to carry the light to a remotely located PMT. The ECL apparatus consists of a three electrode electrolytic cell. A potentiostat is used to maintain the desired potentials at the working (anode) and counter (cathode) electrodes versus a Ag/AgCI reference. An end-on PMT is optically coupled to the cell. Overall PMr gains of 106 are typical. This feature gives the chemiluminescence biosensors the potential for extreme sensitivity. Application of this 'reagentIess' methodology to the determination of glucose and H2~ has been reported previously2.

~
~

"

r ~..
~O2

~o

c""""

II

OH-

~
.

1/

c-.OH c--OH

II

~
i

-Oi.r)

'~ " / a
H

H,NCCH,J.NI"\

'/ T~ /~~
~0izNH.

HC-(CH,h-CH , ~CH,ha-t .. 1'.

~
0

~a-I:4r

l~
1-0, /~~~~~

~1'1'.~~
Figure 1. EnzymeImmobilization y TitaniumLinkage b
to Diatomaceous Earth.

EXPERIMENTAL

CHEMICALS

oxidoreductase, EC 1.1.3.13) was purchased from Provesta (Bartlesville, OK). Titanium oxychloride, ethylene diamine, glutaraldehyde, ethanol, copper(ll) sulfate, dimethylsulfoxide, and luminol were obtained from Aldrich (Milwaukee). Celite Bio-Catalyst Carrier R-648 (diatomaceous earth) was purchased from Manville (Denver). APPARATUS - Electrode potentials were controlled using a Pine Instrument Company model AFRDE potentiostat Light detection was achieved using a ten dynode fIamamatsu R878 5.08cm (2") diameter head-on type photomultiplier tube (PMT) with optimal spectral response at 420nm, and a Nucleus TB-I photomultiplier tube base. High voltage was applied to the PMr using a Nucleus model 575 scalerratemeter power supply. Conditioned PMr output was monitored using a Linear model 2030 chart recorder. In-line pH was determined using Cole Parmer 05992-64 flow through pH electrodes. Flows were established using Masterflex model 7520-35 multichannel peristaltic pumps. Electrocatalyzed chemiluminescence cells were constructed using a Bioanalytical model 94332 gold mesh electrode, an Alfa model 14721 gold foil electrode, and a Microelectrodes model MI 402 Ag/AgCI reference microelectrode. Polarity

Alcohol oxidase (alcohol oxygen

SOLID PHASE BASIFICATION BEDS Two proprietaIy solid phase basification (SPB) materials were examined as candidates for the flow through pH conditioning bed. ELECTRONIC CIRCUITS - An overview of the system electronics is shown in Figure 2. The photomultiplier tube receives a positive bias of 1,280V from the high voltage power supply. The current through the PMr Base, with charge proportional to light intensity, is input to the Baseline Zeroing Circuit in the form of a voltage drop across a IOID 1% current sensing resistor. A baseline corrected voltage is then output to the strip chart recorder. The PMT base was modified to provide a low voltage output across an external current sensing resistor. In order to provide a signal that was proportional to small increases in current through the PMT base, a Baseline-Zeroing Circuit was designed and constructed as shown in Figure 3.A high input impedance DC Differential amplifier, built using LM324N operational amplifiers and lOill 1% resistors, is used to compare the signal to an adjustable offset voltage. The differential amplifier output is a voltage which is proportional to the difference between the input and the offset voltages. The offset voltage is supplied by a voltage divider network through a potential follower. A fifteen turn 20ill potentiometer allows the experimenter to reduce the voltage output, corresponding to the PMr dark current, to less than ImV. The differential amplifier also provides an overall gain of2. The chemiluminescent reaction was electrocatalyzed in the ECL cell by a potential of approximately 0.6V applied across a gold foil working electrode (anode) and a gold mesh counter electrode (cathode). Both voltages were maintained relative to a AgIAgCI reference electrode by the potentiostat.

Polarity switching of the electrodes was incorporated into ECL cell operation to prevent deterioration of performance due to formation of an oxide coating on the gold electrode surfaces. Under conditions of alternating polarity, each

photoactive area was approximately 3cm2. In operation, a flowing stream containing basified luminol and Cu+2catalyst is combined with a separate target analyte containing stream at the inlet to the cell. Photons emitted from the solution are transmitted through the front plate of the cell and detected by the PMr.

+sv

KI

K2

Potentiostat

KI K2 ref reagent flow Function Generator n.n..n...

no

DC

DC

BASE

Tocal

-;, i - --=

IOI:!U% Cumnt Sensing Resistor

IIigb ~ I~
Baseline Zeroing CiIaJit

To Recorder
v

Sv Zener
To

Gold Foil Electrode

. To

Gold Mesh Electrode

Figure 4. Polarity Switching Circuit The ECL cell is shown schematically in Figure 5. The ECL cell was also constructed from two 0.48cm (3/16") thick polycarbonate circular plates 5.08cm (2") in diameter. Attached to the inside of the front plate was a 2.54cm (1") square gold mesh electrode with a mesh density of 39 lines per em, and a gold contact wire. The PMr was optically coupled to the external surface of the front plate. A 25mm square by Imm thick gold foil electrode with gold contact wire was attached to the inside of the back plate. A 0.038cm (0.015") thick PTFE gasket with a tortuous flow path was used as a spacer to separate the plates. The tortuous path configuration was found to improve overall performance by ensuring that air bubbles were not trapped within the cell. The ECL cell functions by combining flowing streams of target analyte and basified luminol at the influent orifice and electrically catalyzing the luminol oxidation reaction within the cell. By maintaining a potential across the electrodes, photons are emitted from the flowing solution, some of which are transmitted through the gold mesh to the PMr.

Figure 2. Electrocatalyzed Chemiluminescence Detection. oxidation step is followed by a corresponding reduction. A polarity switcher circuit, driven by a function generator, was designed and constructed to provide this function. The Polarity Switcher circuit is shown schematically in Figure 4.
High Zin DC Differential amplifier

OIrrDfl Sensntlll46e

f=:~:=':---l i :
1__J

+12v _NT

1.86..< ojfHtvolJ"Ie < 1.73.. +12 v

4"zero

..
~ -12v
LM324N (Qu8d)

Figure 3. Baseline Zeroing Differential Amplification Circuit


CHEMILUMINESCENCE CELLS

Two

separate Figure 5.

chemiluminescentcells were designed and fabricated for chemically and electricallycatalyzed oxidation of luminol respectively. The chemically catalyzed luminescence detection cell was constructed using two polycarbonate circular plates 0.48cm (3/16") thick and 5.08cm (2") in diameter, separated by a silicone O-ring. The PMI' was optically coupled to the surface of the front plate. The internalvolumeof the cell was approximately500~ and the 3

- PI8te
-

G88k8I

Front late P

Electrocatalyzed Chemiluminescence (ECL) Cell.

INTEGRATED CHEMILUMINESCENCE BIOSENSOR PreliminaIy experiments were conducted using chemically

catalyzedchemiluminescence detectionof hydrogenperoxide. These experiments were performed using the test stand

illustrated in Figure 6. A peristaltic pump was used to feed the chemically catalyzed chemiluminescent cell with three flowing liquid streams. The alkaline luminol stream was prepared by sequential flow ofDI water through a SolidPhase Base (SPB) bed. a solid phase luminol bed. and a second SPB bed. The catalyst stream consisted of a 0.024mM copper sulfate in DI water. These two streams were combined to form the carrier stream. The third stream, the analyte stream. mixed with the carrier stream immediately prior to entry into the chemiluminescence detection cell. The analyte stream consisted variously of deionized water blanks and samples with differing levels of hydrogen peroxide dissolved in deionized water. Flow ratios of 1:1 were used for the carrier and analytical streams. The integrated test stand for determination of ethanol by the ECL and with all solid state reagent addition modules is show in Figures 7. Independently controlled peristaltic pumps were used to establish carrier stream and analyte stream delivery to the electrocatalyzed luminescent (ECL) cell. The carrier stream consisted of degassed deionized water which had passed sequentially through a 2.5cm3 SPB bed. a 0.5cm3 crystallized luminol bed and a 5.0cm3 SPB bed. The resultant carrier stream contained 50mgIL luminol at pH 10.3. The analytical stream consisted of varying levels of hydrogen peroxide produced after passage through a 1.2cm3 bed of immobilized alcohol oxidase. The carrier to analytical stream flow ratio was 1.5:1. The carrier stream and the analytical stream were mixed inside the ECL cell.

procedures and ~O2 production rates for alcohol oxidase were investigated sufficiently to demonstrate production of ~O2 proportional to influent ethanol. Solid phase basification beds were sized to consistently produce an alkaline efiluent. Solid phase modules for the controlled release of luminophore were investigated. Preliminary investigations were conducted using chemically catalyzed chemiluminescence. Several designs for electrocatalyzed luminescence cells were evaluated. Operational electrode potentials and means for polarity switching required for stable detection and quantification of ~O2 were studied. PMr high

voltage bias . and detection circuit requirements were


evaluated. This was followed by fabrication and testing of a fully integrated chemiluminescent ethanol biosensor system. Feasibility of the sensor system was ultimately demonstrated by the development of calibration curves for ~O2' and ethanol. Effluent

Cell
PMT

pH

Alcohol Oxidase Column Peristaltic Pumps Deionized Water In Tedlar Gas Bag

Ethanol In DIWater

Figure 7.
H202

Integrated ECL Chemiluminescent Sensor.

RESULTS AND DISCUSSION Prior to assembly and testing of the integrated ethanol chemiluminescent biosensor, each unit operation required for sensor performance was refined to ensure successful operation. These unit operations include: 1) solid phase basification ofDI water to provide an efiluent pH in the range of 10-11, using crystalline media in a packed flowthrough bed; 2) controlled dissolution of a sufficiently high concentration of luminol in basified DI water using a flowthrough module; 3) detection of luminescence; 4) electrocatalyzed chemiluminescent oxidation of luminol by hydrogen peroxide and; 5) enzyme catalyzed reaction of ethanol to produce hydrogen peroxide, using in-line beds of immobilized enzymes. Following the investigation and refinement of individual unit operations, the components were integrated into a fully operational biosensor system. SOLID PHASE BASIFICATION MODULES Solid phase basification (SPB) beds13,14containing a range of particle sizes of a proprietary crystalline material were tested. Each bed was challenged with DI water at varied flow rates, corresponding to a range of empty bed contact times between 1 15 minutes. The results, shown in Figure 8, clearly indicate that efiluent pH is directly proportional to contact

Hp

Figure 6. Chemically Catalyzed Chemiluminescence. Samples were analyzed by operating the carrier and analyte pumps for ten minutes prior to application of potential to the electrodes. After this equilibration period, a 0.6 volt potential difference, relative to the AgIAgCI reference electrode, was applied between the gold foil anode and the gold mesh cathode for a period of one minute. A voltage peak from the PMT circuit proportional to the ~O2 concentration in the analytical stream was measured relative to baseline. EXPERIMENTAL APPROACH Each of the unit operations required for fiber optic chemiluminescent biosensor detection of ethanol were investigated sequentially. These unit operations include: enzyme catalyzed substrate oxidation to produce ~O2' solid phase basification of the flowing aqueous stream to pH~10, controlled release of luminol into the aqueous stream. electrical catalysis of chemiluminescence, and quantification of chemiluminescent light intensity. Enzyme immobilization

time and inversely proportional to particle size. Based on these data, the 75-106~ particle size fraction was selected for testing in series with a luminol bed at a flow rate of 2mUmin. The resultant pH was -7.5 for the 2.5cm3 SPB0.5cm3 luminol configuration. When a second 5.Ocm3 SPB bed was placed downstream from the luminol bed, efi1uentpH increased to 10.3. This pH value was judged much better for promoting efficient luminol chemiluminescence. SOLID PHASE MODULES FOR CONTROLLED RELEASE OF LUMINOL The controlled release of luminol into the sample stream using an in-line solid phase module greatly simplifies the analytical process!5-!7. The module must release enough luminol so that the quantitative chemiluminescent reactions can proceed over the desired concentration range for the target analyte. The module must also exhibit favorable stability and life characteristics so that near constant aqueous luminol levels can be sustained over prolonged periods of continuous flow.

continually decreasing throughout the experiments. Evidently, the quantity of luminol that dissolves depends more on the influent pH than on the bed size (contact time). In the present application, the aqueous luminol concentration can be controlled by the pH adjustment of the SPB to produce a nearly constant level for prolonged periods of sustained flow. This behavior is ideal for application in reagentless chemiluminescent biosensors.
CHEMICALLY CATALYZED LUMINESCENCE

11.0
10.8
10.6 10.4 :I: 2" 10.2 c: .. ::J IE 10.0 w 9.8 9.6
9.4 9.2 0 1 2 3 4 5 6 7 8 9 v 0 0 10 7S.I06J1111 300-600 JIIII 6OO-IOOOJIIII

I!J. ISO. 300 JIIII

Empty Bed Contact Time (minutes) V/Q

11

12

13

14

15

Figure 8. pH versus Contact Time for SPB Size Fractions. Luminol is sparingly soluble in aqueous solutions, and for this reason is compatI1>le with use in a solid phase module. Luminol is commercially available as a fine powder which is not usable for an in-line module due to the inordinately high inlet pressures required to sustain flow. Crystalline luminol with a more favorable particle size distribution was obtained by recrystallization using a proprietary process. A 0.5cm3 bed of crystalline luminol was prepared and used to determine the washout characteristics of this material with a DI water influent at flow rates between 1.6 - 2.Ocm3/min. The results obtained by UV spectrophotometry are shown in Figure 9. The same bed of crystalline luminol was subsequently challenged with the effiuent from an SPBbed. The efi1uentluminol concentration ranged between 42-59mgIL. The solubility of luminol is known to be strongly pH dependent. The drop in pH from 10.3 to 7.5 after passage through the luminol bed indicates pH control of the dissolution process. Under continuous flow conditions, small variations in the pH of the water produced by the SPB bed produced only minor variations in aqueous luminol concentration even though the bed size was 5

Chemically catalyzed chemiluminescence was first investigated using prepared aqueQJlSreagents. DI water was basified by passage through a SPB bed. Luminol dissolved in an ethanol-dimethyl sulfoxide mixture, and aqueous eu+2 catalyst was added to the basified solution. Hydrogen peroxide was also mixed with the luminol-catalyst solution just prior to injection into the chemiluminescent chamber adjacent to the PMI'. This ensured that the chemiluminescence was confined within a small volume in close proximity to the PMI'. A cah1>rationcurve for ~O2 in the concentration range between 0 - 30mgIL obtained using this method is given in Figure 10. This was followed by similar experiments using SPB and crystalline luminol beds, with aqueous eu+2 introduced into the basified luminol containing stream. The cah1>ratioD curve for hydrogen peroxide over the concentration spaD between 0 6OOmgIL generated using the solid phase modules is shown in Figure II. The two fold increase in th( effective range for the solid phase reagent system over th( liquid reagent system was most probably due the increased pF. of the solid phase system effiuents. Differing amounts OJ Cu+2 between the two systems might also have beer responsible, since the formation of the insoluble hydroxide if favored at these high pHs. The chemically catalyze< luminescence experiments demonstrated the quantitativI utility of the chemiluminescent oxidation reaction betweer ~O2 and luminol, and also confirmed the adequacy of the PMI' and the optical coupling between the cell and the detector. The results of these experiments also prompted th( development of baseline compensation and amplificatiOl circuitry for the PMI' output so that much lower levels of the target analytes could be detected and quantified accurately. ECL DETECTION - The first ECL experiment provided ; greater than 30 fold improvement in the minimum detectio! limit of hydrogen peroxide over the previously use< chemically catalyzed system. After three hours of operation however, ECL system response decreased by a factor 0 approximately one third. Upon disassembly of the ECL cell a discoloration of the foil electrode surface was noted. Th, formation of an oxide coating of the gold surface wa suspected. The surface contamination was removed and th, cell reassembled. Cyclic voltammetry (CV) was performed using th, potentiostat and ECL cell electrodes to determine preciseI: the sign and magnitude of electrode potentials which resulte in oxidation and reduction of the gold electrodes. The C1 experiments were also expected to indicate if other oxidatior reduction reactions were 0 ccurring dwing ECL determin-

25

~
CI

715 7.0 '8:6.5 E 6.0 !6.5 ';a 5.0 -E 415 10


::> 4.0 ~3.S

..5.20 c .2 !
'E 15
c 0 0

..

::> ..J 'E

g E

e w

.
0 0 5 10 15 20 25 30 35 40 45 Throughput (UIersIcm')

.
50

~u 55

~ 3.0 f! 2.0
-2.5
0.!5 0.0 051015202530354045505511085
HydIOgen

l51.0

eoo
Peroxide (mgL)

850

Figure

9. Washout

Curve

CIYSfalline Luminol Module.

Figure 11. Cu(ll) Catalyzed ~O2 CL (Luminol-Bed)


anons. The voltage sweep rate was set at 20m. Is, with the V gold foil as the working electrode, the gold mesh as the counter electrode, and the AgI AgCI reference electrode. The CV trace illustrated in Figure 12 shows symmetrical oxidation-reduction reactions at +O.45V and +O.175V, with peaks at higher and lower voltages representing the oxidation and reduction of water. These results indicated that the anode potential of +O.8V applied to the foil was too high, promoting oxidation of the gold electrode. Continuous oxidation of the working electrode surface was judged responsible for degradation of the capacity to catalyze luminescence between luminol and hydrogen peroxide. A significant problem in the determination of suitable non-oxidizing electrode potentials originates from the low conductivity of the supporting electrolyte in the chemiluminescent biosensor configuration. This results in a significant (but unknown) potential drop between the reference and working electrode(s). As a result, the optimal voltage could only be determined empirically using the ECL cell and reference electrode. This was accomplished, and in subsequent runs, the anode voltage was reduced to +O.6V.
9
'(;;'8 ... 210

~
~19)

::> 0 11)

50

Q8

1.0

\tI;ge

.11)

.19)

.a

Figure 12. ECL Cell

-Cyclic Voltammogram.

7 a. <i6 .
'E

Periodic switching of electrode parity was initially considered a more elegant solution to the gold oxidation problem. This was attempted when ECL cell performance degradation was first observed and abandoned because of instability in the output signal. The switching interval used varied between 0.2 to 30 seconds. Potential of both the working and counter electrodes was controlled relative to the reference. The lack of success for this approach most
probably lies in the asymmetry between the electrodes.

a 1-4
~
~
~3 'E 122

~s

c 1 Ou 0 5
10 15 20 25 30 35

Hydrogen Peroxide (mgIl)

Figure 10. Cu(ll) Catalyzed~O2 CL (AqueousLuminol) 6

During DC operation of the ECL cell, the PMr voltage reached its peak value within one minute of sample introduction. The primary delays in the detector response were due to the residence times within the solid phase beds and interconnecting tubing. Based upon these results it was concluded that continuous powering of the electrodes was not only unnecessary but also detrimental to long term sensor stability. Subsequently, the ECL electrodes were powered for one minute intervals at the end of which peak voltage was measured. The combined effects of the new timing format and the reduced anode potential resulted in an increase in the

viable operation time periods of over eight hours with only a 10% decrease in sensor response. Although the overall stability was much improved using the new operating procedures, this level of degradation was still considered unacceptable. For this reason development of suitable electrode regeneration methods were undertaken. This resulted in the adoption of a new polarity switching strategy in which the system was operated normally for six to eight hours, powering the electrodes intermittently for one minute intervalsto perform analyses. The operational period was followed by one hour of continuous reversed polarity which was found to fully return the ECL electrodes to their original performance level. The system was operated in this manner for approximately 100 hours over a six week period with no discernible degradation in performance. Using this slow switching operational protocol, the ECL for the luminolhydrogen peroxide reaction proved stable and reproducible. IMMOBILIZED ALCOHOL OXIDASE BEDS Alcohol oxidase (AO) was assayed for enzyme activity by the disappearance of ethanol and the appearance of acetaldehyde as determined by gas chromatography. AO on DES was found to produce stoichiometric conversion of 15.8mg/L ethanol to acetaldehyde in a IOcm3 bed with a flow rate of 2cm3/min. Unfortunately, ~O2 is readily decomposed by transition metal impurities on the diatomaceous earth support. This resulted in poor recovery of the hydrogen peroxide from the immobilized enzyme bed. It was found that H202 recovery could be improved dramatically when flow rates and bed sizes were adjusted to give a better match between challenge ethanol concentrations and enzymatic activity. For application in the integrated ethanol chemiluminescent biosensor, a 1.2cm3 AO bed was used at a flow rate of 3cm3/min, corresponding to a carrier stream to analytical stream flow ratio of 1.5:1. In this configuration a challenge stream containing 395J.1g/Lof ethanol produced 33 J.1g/Lof hydrogen peroxide. This corresponds to a H202 production efficiency of 11.3%. INTEGRATED BIOSENSOR TEST The components required to perform each unit operation were assembled into an integrated systems for quantitation of ethanol. Aqueous ethanol solutions ranging in concentration between 40 and 4114J.1g/L(ppb) as C were determined using the integrated chemiluminescent ethanol biosensor. The resulting data were used to construct the smooth quadratic calibration curve over this range shown in Figure 13. Because the ability to detect very low levels is an important feature of the ethanol biosensor system, an expanded scale version of this curve between 0-250J.1g/L(Ppb) is presented in Figure 14. These data clearly demonstrate the sensitivity of this technique and the capacity to extend the lower limits of detection to even lower values with the current sensor configuration. The ethanol sensor was challenged with three replicates each of standards containing 40, 200, 410, and 4,100J.1g/L(ppb), resulting in standard deviations (lcr) of 3.6, 17.0, 25.4, and 98.4J.1g/L (ppb) respectively.
.

4500 4000 ~3500 as ~300) c: 0 2500

82000

8 1500 "0
~ 1000 W
500 0 0
10

20

30

40 PMT Response

50 (mY)

60

70

80

90

Figure 13. Ethanol Calibration Curve.


250

6 :::200

~ c
,8150 ,g c

8100 "0 c ffi 50

0 2 3 4 5
6 7 8 9 10

11

12

PMT Response (mV)

Figure 14. Expanded Scale Ethanol Calibration Curve

CONCLUSIONS Amperometric biosensors suffer from several problems which render their use impractical for deployment aboard spacecraft. Among these are short active lives, high rates of signal drift, and the consequent need for frequent recalibration. A novel chemiluminescence based approach to biosensor development, incorporating solid phase media in replacement of aqueous phase reagents has shown potential as a means to overcome these difficulties. Improvements in sensitivity are also achievable using this technique. Feasi.1>ility of the Ethanol Chemiluminescent Biosensor concept has been rigorously demonstrated. Calibration curves have been generated using fully integrated reagentless test systems. Ethanol concentrations as low as 40 J.1g/L have been detected. The height of the signal above the baseline for this concentration suggests that even lower levels could be detected using the current relatively crude embodiment of the innovative technology. Further

development is expected to lower the ethanol detection limit by an order of magnitude. Detection limits are primarily determined by enzyme conversion efficiencies, dark current limitations, light leakage, detection circuitry, and background chemiluminescence. Improvements in enzyme immobilization procedures and hardware design can certainly provide means to significantly overcome these limitations. Unit operations required for sensor operation have been developed sufficiently for demonstration of the concept These include pH adjustment using solid phase flow-through modules, immobilized enzyme catalyzed oxidation of ethanol to hydrogen peroxide, controlled release of luminol using a solid phase flow-through module, electrocatalyzed luminescence using a potentiostat and gold electrodes, and quantification of light intensity using a photomultiplier tube. The solid phase luminol beds were shown capable of sustained controlled release of the luminophore at a sufficient concentration for the biosensor systems to function over a broad span of target analyte concentrations. The ECL cell fabricated from transparent polycarbonate, and using a gold foil working electrode, and a gold mesh counter electrode, was found to effectively initiate the reaction between luminol and hydrogen peroxide at an applied potential of 0.6 volts. The ECL cell design was also found to adequately confine the reaction spatially for efficient collection of the light either directly by an optically coupled photomultiplier tube. Alcohol oxidase effectively catalyzes the oxidation of several low molecular weight primary alcohols. Work to date has examined the response of the chemiluminescence based biosensor to ethanol, although the sensor can be expected to respond to other alcohols. When applied to the examination of reclaimed waters and wastewater streams associated with regenerative life support systems, the sensor response will be proportional to the sum of methanol and ethanol present If quantitation of individual alcohols is required, this technique is applicable as an alcohol selective detector which can be used in conjunction with liquid chromatographic separations. NASA's current requirements for reclaimed potable water specifies a maximum of SOOJ.1g/L OC, of which a T maximum of lOOJ,Lg/L may be uncharacterized, i.e., of unknown composition. Current technologies such as aqueous phase catalytic oxidation systems (APCOS) are certainly capable of meeting the 500J,Lg/L limit, but may have difficulty with the lOOJ,Lg/Llimit depending upon the magnitude and composition of the influent organics. The problem of uncharacterized TOC can be diminished by using chemiluminescent alcohol biosensors to determine the concentrations of low molecular weight straight chain
alcohols in the product water.

REFERENCES 1. Akse, J.R, Atwater, J.E., Thompson, 1.0., and Wheeler, RR, Jr., A Breadboard Electrochemical Water Recovery System for Producing Potable Water from Composite Wastewater Generated in Enclosed Habitats, in Water Purification by Photocatalytic, Photochemical, and Electrochemical Processes, Rose, T.L., Conway, B.E., Mwphy, O.J., and Rudd, E.J, Eds., Electrochemical Society, Hooksett, NIl, 1994. 2. Akse, 1.R, Atwater, JE., Schussel, L.J., and Verostko, CE., Development and Fabrication of a Breadboard Electrochemical Water Recovery System,SAE Trans J. Aerospace, 102, (1), 513-529, 1993. 3. Akse, J.R, and Atwater, J.E., Advanced Catalytic Methods for Destruction of Environmental CoDt~min:mts, AIAA-95-LS-70, presented at AIAA/NASA Life Sciences and Space Medicine Conference, Houston, April 3-5, 1995. 4. Akse, J.R, Carter, D.L., Jolly., C, Thompson, J., and Scott, B., Catalytic Methods Using Molecular Oxygen for Treatment of PMMS & ECLSS Waste Streams, presented at the 22nd International Conference on Environmental Systems, Seattle, Washington, July 1992. 5. Schussel, L.J., and Atwater, J.E., A Continuous Alcohol Oxidase Bioreactor for Regenerative Life Support, submitted to Enzyme Microb. Technol., 1995. 6. Schussel, L.J., and Atwater, J.E., A Urease Bioreactor for Water Reclamation Aboard Spacecraft, Chemosphere, 30 (5), 985-994, 1995. 7. Jolly, C.D., Schussel, L.J., and Carter, L., Advanced Development of Immobilized Enzyme Reactors, in Regenerative Lift Support: Systems & Processes, SP-873" Behrend, AF., MacElroy, RD., and Reysa, RP., Eds., Society of Automotive Engineers, Warrendale, FA, 1991. 8. Nieman, T.A, Detection Based on Solution-Phase Chemiluminescence Systems, in Chemiluminescence and Photochemical Reaction Detection in Chromatography, J.W. Birks, Ed., VCR, New York, 1989. 9. Van Dyke, D.A, Applications ofElectrocatalyzed Chemiluminescence ofLuminol, Doctoral Dissertation, University oflllinois, Urbana, 1986. 10. Faulkner, L.R and A1. Bard, Techniques of Electrocatalyzed Chemiluminescence, in Electroanalytical Chemistry, A Series of Advances, 10, 1, 1977. 11. Van Dyke, D.A and H.-Y. Cheng, Electrochemical Manipulation of Fluorescence and Chemiluminescence Signals at Fiber-Optic Probes, Anal. Chem. 61, 633, 1989. 12. Atwater, J.E., Akse, J.R, DeHart, J., and Wheeler, RR, Jr., Reagentless Chemiluminescence Based Fiber Optic Sensors for Regenerative Life Support in Space, SPIE Proceedings Series, 2574, 78-84, 1995. 13.Ieffers, E.L., Atwater, I.E., and Dehart, 1., New Technologies for On-Line Water Quality Monitoring, SAE Technical Paper Series No. 932181, presented 23rd International Conference on Environmental Systems, Colorado Springs, July 1993.

ACKNOWLEDGMENT This work was supported by the National Aeronautics and Space Administration's Lyndon B. Johnson Space Center, Houston, Texas, under contract NAS9-19021.

-Colorado Springs, July 1993. 14. Jeffers, E.L., Dougherty, D.R, Paxton, T.A, and Atwater, J.E., An On-Line Water Quality Monitor for Space Station Freedorn, paper 79c, presented at AIChE Spring National Meeting, Houston, March 1993. 15. Hool, K. and T.A Nieman, Chemiluminescence Analysis in Flowing Streams with Luminol Immobilized on Silica and Controlled-Pore Glass, Anal. Chern., 59, 869, 1987. 16. Hool, K. and T.A Nieman, Immobilized Luminol Chemiluminescence Reagent System for Hydrogen Peroxide Determinations in Flowing Streams, Anal. Chern., 60, 834, 1988. 17. Nieman, T.A, Immobilized and Solid-State Reagent Systems for Luminol Chemiluminescence in Flow Systems,Mikrochirn. Acta {Wien} lII, 239, 1988.