Development of a Multipurpose Optical System for Use ... - CiteSeerX

3 downloads 0 Views 2MB Size Report
Oak Ridge National. Laboratory,'. Oak Ridge, Tenn. 37830. 1 Operated for the Energy Research and Development. Adminis- tration by the Union Carbide Corp.
CUN. CHEM. 21/9, 1225-1233 (1975)

Development of a Multipurpose Optical System for Use with a Centrifugal Fast Analyzer C. A. Burtis, W. D. Bostick,

and W. F. Johnson

A Centrifugal Fast Analyzer is basically a sophisticated photometric measuring device containing a multicuvet rotor as its major component. Several reactions are simultaneously initiated in the rotor, which is then rotated

through a stationary optical monitor and the resulting signals acquired and processed by an on-line data system. With the miniature version of this analyzer, one has the option of directing the incident optical beam, via a fiber optical bundle, into the cuvets of the spinning rotor in either a 900 or a 1800 orientation relative to the analyzer’s photodetector. The combination of newly designed rotors and a flexible optical system having multiple configurations has provided a versatile system in which one can measure the transmittance, fluorescence, chemiluminescence, or light-scattering (either turbidimetrically or nephelometrically) properties of the ensuing reaction species, all with a single analyzer. This flexibility in choosing the optical mode in which a particular set of reactions is to be monitored provides the analyst with a powerful and versatile analytical tool for developing new methods for use in various clinical laboratory applications, including chemistry, toxicology, immunology, and hematology.

A miniature version of the Centrifugal Fast Analyzer (CFA) is currently under development at the Oak Ridge National Laboratory (5-8). This system should prove useful for a variety of analytical uses, including general clinical analyses (routine as well as emergency), pediatric or small-animal applications, water-pollution analyses, and even for providing onboard analyses in an orbiting space laboratory. A versatile, multipurpose optical system has been developed for use with the miniature CFA. With the new optical system, one has the option of directing the incident optical beam, via fiber optics, into the cuvets of the spinning rotor in either a 90#{176} or a 180#{176} orientation relative to the analyzer’s photodetector. This combination of a flexible optical system and specially designed rotors provides a versatile and useful analytical tool for measuring either the transmittance, fluorescence, chemiluminescence, or light-scattering (either turbidimetrically or nephelometrically) properties of several simultaneously initiated reaction mixtures. This paper discusses the design of the optical system and presents some preliminary results

AddftlonalKeyphrases: multiple rotor designs #{149} analytical versatility #{149} spectrophotometer #{149} spectrofluorometer

obtained

nephelometer minescence

Materials

#{149} turbidimetric #{149} fiber optics

measurements

#{149} chemilu-

analyzer is a photometric instrument that incorporates a rotating multicuvet assembly with a stationary optical detector and online data processing facilities (1, 2). The unique combination of centrifugal force and a stable, drift-free optical system results in an instrument with several interesting and advantageous features (3, 4). Basically,

a centrifugal

Oak Ridge National Laboratory,’ Oak Ridge, Tenn. 37830. 1 Operated for the Energy Research and Development Administration by the Union Carbide Corp. Received April 10, 1975; accepted April 28, 1975.

with it. and Methods

instrumental The basic analytical system used in these studies has been previously described (8). The original optical module of this system consisted of a quartz iodine tungsten lamp whose output was directed downward through the rotating cuvets with the resulting transmitted light passed through an interference filter and then into a miniature photomultiplier tube. The multipurpose optical system was developed by adding additional light sources, fiber optics, focusing lens, and ancillary electronics to the original system. A schematic of the modified system is shown in Figure 1. CLINICAL CHEMISTRY, Vol. 21, No. 9, 1975

1225

i

>K

FIBER

OPTIC

“-‘-‘-“-#{149}

BUNDLE

SOURCE

TTY

PUT

l__-1ANPLIFIERI--_--__JfIPUTERl

Fig. 1. Schematic of the multipurpose optical system devel-

oped for use with a miniature Centrifugal Fast Analyzer

PUT

Fig. 3. Cross-sectional view of lens and filter assembly used to collect and focus emitted or transmitted light onto the active element of the analyzer photomultiplier tube Fig. 2. Modification of the optical head of the Centrifugal Fast Analyzer, which allows for either of two fiber optic configurations: (A) 180#{176} orientation; (B) 90#{176} orientation

Light sources.

Two different

external

light sources

have been used with the modified system. One, which is used routinely in our clinical laboratory for the analysis of uric acid at 292 nm by a uricase method, is a deuterium source coupled to a 180- to 400-nm grating monochromator (Bausch & Lomb, Inc., Rochester, N. V. 14625). The second source, which is used primarily in the fluorescence and light-scattering ex-

periments, is a 200-W xenon mercury source coupled with a small monochromator and a stabilized dc power supply (Schoeffel Instrument Co., Westwood, N. J. 07675). Fiber optics. The monochromatic output from each light source can be directed into the cuvets of the spinning rotor in either a 90#{176} or a 180#{176} orientation (relative to the photomultiplier tube) via quartz optical fiber bundles, 4 mm in diameter and 20 cm long (Schott Optical Glass, Inc., Duryea, Pa. 18642).

Two such bundles are used, one for each of the orientations shown in Figure 2. When operating in either of the two orientations, is attached to the light adapter mounted on the tor. Focusing of emissions. provided for excitation

resulting

emitted

the appropriate fiber bundle source by insertion into an exit slit of the monochromaMonochromatic light is by the monochromator. The

signals are monochromatically

iso-

lated by either the appropriate interference or sharpcut filter (Ditric Optics, Inc., Marlboro, Mass. 01752).

In addition, the emitted light is collimated and focused on the photosensitive element of the photomultiplier tube by the lens assembly (Li = stock No. 94007, L2 = stock No. 94754; Edmund Scientific Co., 1226 CLINICALCHEMISTRY,Vol. 21, No. 9, 1975

LI

17-mm FL piano-convex lens; L2 = 15.4-mm FL double-convex lens

Barrington, N. J. 08007). A cross-section of the lens and filter assembly is shown in Figure 3. Photodetector. The miniature CFA contains a miniature photomultiplier tube (Type HTV R-300, Hamamatsu Corp., Lake Success, N. V. 11040), mounted internally in its cabinet and located directly below the rotating cuvets. The output from the photomultiplier tube is amplified by means of a variable gain operational amplifier (i.e., lOX, bOX, 1000x, 5000x) before being processed by the on-line data system (9). This amplification of its output allows the photomultiplier tube to be operated at 300 to 500 V, while still maintaining a 0- to 10-V signal across the analog-to-digital converter of the computer.

Rotor Design and Fabrication Techniques A basic distinction between the original and miniature versions of the Centrifugal Fast Analyzer is the difference in the rotor designs. In the original analyzer, the rotor and its cuvet assembly were part of the basic analyzer; aliquots of samples and reagents were introduced into it via a removable and reloadable transfer disk. With the miniature CFA, the cuvet assembly and transfer disk have been combined into a single integrated device, which can be externally

loaded, cleaned, and refilled, or discarded. The separation of the basic analyzer module the analytical

rotor

is a subtle,

but

and

important, conceptual design difference because it provides for much flexibility and versatility in the design of rotors for use with the miniature CFA. Consequently, various types of rotors have been designed and fabricated for use with the miniature CFA (8-15). Figure 4A shows one of the more useful designs,

Fig. 4. Family of rotors designed and fabricated for use with the miniature Centrifugal Fast Analyzer (A) Gsneral-ptxpose transmission rotor, body fabricated of 0.5-cm black acrylic plastic (chambers In serial array), windows fabricated of uftravlolet transmitting (UVT) acrylic plastic: (B) general transmission rotor having quartz windows; (C) clear-body rotor, body fabricated from 0.5-cm UVT acrylic plastic, chambers In serial array, UVT wIndows; (0) clear-body rotor, chambers in parallel array. LJVT windows; (5) clear-body rotor, black acrylic top window, UVT bottom window (this rotor has since been modified, as described in text); (F) black-body rotor with quartz cylinders sealed into the peripheral wall of each cuvet, UVT top and bottom windows

used primarily rotor

into

each

cuvet

with

a silastic

adhesive

(Dow

Cor-

ning, 3145 RTV; Dow Corning Corp., Midland, Mich. 48640), which was used to bond the windows to the rotor body. Although this rotor has the advantage of accepting energy at the lower wavelengths, it is more expensive and has a higher within-rotor, cuvet-tocuvet variability when compared with the clear-body rotor. Fabrication of a reflective rotor cover. In previous work, an acrylate resin (Plexiglas) rotor cover has

light. Each

been

device

con-

body and two cover windows.

The

aliquots of samples and reagents from a loaded rotor (8). Therefore, a rotor evaporation cover that has a reflective surface on its bottom surface was fabricated in an effort to increase the analyzer’s sensitivity for performing either fluorescent or nephelometric light scattering measurements. When this cover is placed on the clear-body rotors (Figure 4C and D) or the insert rotor (Figure 4F), its reflective surface serves to reflect emitted or scattered light downward into the photomultiplier tube of the analyzer.

in measuring

of this design

clear, transparent body, or (b) a cylinder of quartz could be inserted into the peripheral wall of each individual cuvet. Rotors based on these options have been fabricated and were designated “clear-body” (Figure 4C, D, and E) and “insert” (Figure 4F) types, respectively. Fabrication of the clear-body rotor. The clearbody rotor is the most desirable from the standpoint of economy in the fabrication process. The central body of the rotor could be fabricated from the same ultraviolet-transmitting acrylic plastic that is used in the cover windows and then assembled into the final rotor as previously described (10). However, it must be pointed out that the resulting rotor will not accept excitation energy below 300 nm because of the optical absorbance characteristics of the ultraviolet-transmitting plastic. Fabrication of the insert rotor. In the insert rotor, one of the black acrylic bodies usually used in the fabrication of rotors to be operated in the 180#{176} orientation was modified by drilling a hole in the peripheral wall of each cuvet. A cylinder of quartz was sealed

transmitted

is a three-component

sisting

of a center

center acrylic

body is made from a 0.5-cm-thick plastic, into which various

disk of black liquid-holding

chambers, transfer channels, and 17 cuvets have been machined. The two cover windows are fabricated from ultraviolet-transmitting acrylic plastic, which transmits light over the range 300 to 700 nm. In applications where it is desirable to perform measurements at lower wavelengths, annuli of quartz can be used as cover windows (Figure 4B). In the development of rotors for use with the multipurpose optical system, it was necessary to design

routinely

used

to minimize

evaporation

of the

Optical Configurations

rotors that would function in either the 90#{176} or the 180#{176} orientation. The choice of fabrication material and the technique for fabrication were also important from the standpoint of providing versatility. Design. To ensure capability for operating in either a 90#{176} or 180#{176} orientation, it was necessary to design several different types of rotors. The 180#{176} orientation posed no problems, because previously designed rotors have already been used in this mode

The components of the multipurpose optical system can be geometrically positioned in a variety of orientations. In Figure 5, three of the more useful ones are schematically demonstrated. The 180#{176} orientation, in which the optical beam from the light source passes downward through the cuvets and directly into the photomultiplier tube, is illustrated in Figure 5A; transmission and turbidimetric light scattering measurements are made in this orientation. The 90#{176} orientation, in which the optical beam enters

(6-8).

the

In the 90#{176} orientation, it was necessary to design a rotor having an end window in the peripheral wall of each cuvet. Two design options were possible: (a) the

multiplier tube, is illustrated in Figure 5B; fluorescence and nephelometric light scattering measurements are made in this orientation. No light source is required in the orientation shown in Figure SC, be-

center

body

of the

rotor

could

be machined

from

a

cuvets

at a right

angle

in relation

CLINICAL CHEMISTRY.

to the

photo-

Vol. 21, No. 9, 1975

1227

LIGHT SOURCE -

B

The calibration procedure consists of dynamically injecting (9, 12) a standard solution of the fluorescent compound to be analyzed into the spinning rotor. After the cuvets have been filled, the fluorescence intensity is measured, recorded, and stored for each cuvet (Ii), and the mean intensity of all the cuvets (1) is computed and stored. A factor (Fe) is computed for each individual cuvet as follows: F1 =

COOQ

Subsequent determinations of standards or unknowns corrected as follows:

(1)

of fluorescent intensity may then be automatically

I, corrected

=

11.F1

(2)

A

Referencing of Fluorescence Intensity Signals Fig. 5. Optical configurations of the multipurpose optical system

cause the chemical reactions inherently produce optically measurable energy; this orientation is used for chemiluminescence measurements.

Data Processing Because centrifugal analyzers produce data in a form and at a rate compatible with on-line data processing, they have been successfully interfaced with minicomputers (2, 3, 8). A data system has been previously described (8) for use with the miniature Centrifugal Fast Analyzer. Its basic hardware consists of a PDP-8/E computer (Digital Equipment Corp., Maynard, Mass. 01754) with 8192 words of core memory, a 1200-Hz line-frequency-based clock, an ASR-33 teletype, a Sykes cassette-tape unit (Sykes Datatronics, Inc., Rochester, N. V.), and the necessary analog and digital interfaces for coupling the analyzer to the computer hardware. Several analytical programs have been developed and used (3, 8, 16, 17) to process data obtained in the transmission mode. Two programs have been developed to process the intensity data acquired from fluorescence measurements. One of these is used to obtain cuvet calibration factors, while the other is used to convert relative intensity measurements directly to units of concentration.

Cuvet Calibration Fluorescence signal and background intensities are dependent on cell geometry. With a multicuvet rotor, it is desirable to develop a calibration procedure that will correct for within-rotor cuvet-to-cuvet geometric differences. This is especially important when the insert rotor design is used; for this design, a cuvet-tocuvet variability of up to 6% (in terms of signal observed from a single concentration of fluorescent compound) may occur. The cuvet-to-cuvet variability for a clear-body rotor is about 3%. 1228 CLINICALCHEMISTRY.Vol. 21, No. 9, 1975

A basic technique in spectrofluorometry is to convert relative fluorescence intensity measurement to units of concentration. Typically, this is done by comparing the relative fluorescence intensity of the unknown sample to that of a known concentration of a fluorescent compound. This process is quite simple with a CFA (18), because a cuvet can be dedicated to contain the standard solution whose relative fluorescence intensity then becomes the standard measurement to which the intensities from the other cuvets are compared and normalized. A program previously developed for the “fluorometric” GeMSAEC Fast Analyzer (18) has been adapted for use with the miniature version of the CFA. In this routine, cuvet 1 serves as a reference cuvet and contains a solution of the standard reference compound with the remaining cuvets (2 through 17) serving as analytical cuvets. With this routine, the relative fluorescence intensity of the solution contained in each cuvet is measured and digitized during each revolution of the rotor. For signal-averaging purposes, the intensities may also be measured for several successive revolutions, averaged, and then stored. This process is repeated at several successive observation intervals as predetermined by the values entered by the operator for the data acquisition variables. At the end of the analysis, the stored values are recalled and correlated for cuvet-to-cuvet variation (see Cuvet Calibration above); the normalized intensity for each cuvet at each observation interval is subsequently cal#{232}ulated by obtaining the ratio of sample intensity to the reference intensity. Because the concentration of the solution producing the reference intensity is accurately known, the normalized intensities of the analytical cuvets can be calculated directly in units of concentration. The resulting data are then used in various analytical calculations. Thus this dynamic

referencing

technique

serves

as a means

for rapidly converting fluorescence data into units of concentration, and also provides for improved instrumental performance as a result of its continuous referencing capabilities, which minimize the effect of long-term instrumental drift and short-term intensi-

ty changes in the light source tion energy.

supplying

the excita-

1,4

12

Chemicals Fluorescein was purchased in the form of the sodium salt (VWR Scientific, San Francisco, Calif. 94119). Standard solutions were prepared by dissolving various quantities of the compound in appropriate volumes of NaOH (10 mmol/liter). Practicalgrade 4-methylumbelliferone (Sigma Chemical Co., St. Louis, Mo. 63178) was further purified by recrystallizing twice from ethanol. Standard solutions were prepared in 0.1 mol/liter glycine/carbonate buffer, pH 10.0. Luminol (5-amino-2,3-dihydro-1,4-phthalazinedione) was purchased from the Aldrich Chemical Co., Milwaukee, Wis. 53233. Cholesterol oxidase (EC 1.1.3.6) and cholesterol standards were provided by Boehringer Mannheim GmbH, Mannheim, West Germany, as part of their enzymatic cholesterol kit. TEKIT (Searle Diagnostic, Inc., Columbus, Ohio 43216) was used as a diluent for the cholesterol standards.

and DIscussion Transmission Measurements

I0

008 C

0,6

0.4

02

50

-

40

POTASSIUM

60

50

CONCENTRATION

00 I.gI.I)

FIg. 6. Turbidlmetric determination of potassium

sodium

with

tetraphenylboron Reaction condItions: 25 p1of sample, 100 p1of H20 (dlluent), 25 pI of reagent (30 g/llter) sodium tetraphenyl borate In 0.6 mol/llter borate buffer, pH 9.0

Results

Table 1. Centrifugal Force vs. Rotation Speed for the Miniature Centrifugal

The performance of CFA’s in the spectrophotometric mode is well documented (1-8); a recent review (19) lists some 30 different colorimetric clinical laboratory analyses that have been adapted for use with such instruments. These include kinetic as well as nonkinetic analyses (16). The use of an external light source coupled to a grating monochromator, as in the present module, has certain advantages over the original optical module (6) in which the cuvet was irradiated with white light and monochromatic light produced by an interference filter positioned between the cuvet and detector. With a contbiuous source and grating monochromator, any wavelength within the continuum may be rapidly selected; with a motor-driven grating monochromator, automatic scanning or multiple wavelength monitoring is feasible. In addition, the decomposition of photosensitive chromogens may be

decreased

by using light that has been made mono-

chromatic.

The parallel

analyzer

concept

turbidimetric (180#{176} excitation) yses (13, 20, 21). The inherent turbidimetric analysis include

has also been used in light-scattering

anal-

advantages of parallel the virtual instantaneous and simultaneous treatment and observation of samples and standards (20). Hence many of the variables that affect variable particle growth-e.g., mixing, temperature, and time-are controlled within a run and, in turn, lead to an improved analytical precision. In this laboratory, we have used turbidimetric measurements for blood grouping (14), prothrombin-time determination (22), and chemical analysis for potassium (via sodium tetraphenylboron,

Figure 6).

Fast Analyzer

Rotational speed

Centrifugal force”

500 1000 1500

11.5 46.1 103.6

2000 2500

287.9

3000 3500 4000

414.6 564.3 734.0

(rpm) 200

(xg)

1.8

184.3

Calculated from the following (N1(R), where N = revolutions radius

(cm)

4.1 for

the rotor

formula (23): g = (1.118 X 10’) per minute (rpm), R = rotating used (I the miniature CFA.

When a CFA is used for light scattering experiments, one must be concerned with the possible sedimentation of particles. The small size (89 mm diameter) of the rotors used in the miniature CFA, together with the wide range of rotation speeds that may be used, yields a dynamic range of attainable centrifugal fields (Table 1). Typically, rotor speeds of 500 to 1000 rpm (equivalent to a force of 11 to 46 X g within the cuvet) are used during photometric measurement; the maximum speed available is approximately 4000 rpm (a force of 734 X g).

Chemiluminescence Measurements Chemiluminescence

(CL)

is the production

of light

via a chemical reaction. Because there is no excitation radiation source to be stabilized and no reflection of exciting radiation to contribute to the signal CLINICAL CHEMISTRY.

Vol. 21,

No. 9, 1975

1229

+

H202

METAL CATALYST

2 I-

+ hi’ (425

FIg.

7.

nm)

U) 2 Ui I-

z

Luminol reaction ChOLESTEROL CHOLESTEROL CHOLESTEROL

CHOLESTEROL CHOLESTEROl.. TIME IN SECONDS

Fig. 9. Chemlluminescence intensity as a function of the lumifbI reaction when coupled to the H202-generating cholesterol oxidase:cholesterol reaction Reaction conditIons: 5 p1 of cholesterol standard. 30 p1 of TEKIT, 40 p1 of 1 X 102 mouNter K3Fe(CN). and 10 p1 of cholesterol oxldase (2.5 U/mI) were loaded Into the chambers of a general-purpose transmission rotor and transtarred into and mixed within thee respective cuvets. After an 8-nile Incubation, the rotor was stopped and 30 p1of kiminol solution (11 mmol of luminol per NterIn 0.1 mol/lIter potassium borate buffer, pH 10.5)was added to each reagent chamber. The luminol solution wasthen transferred; aftermixIng, the resultant chemulumlnescencewas monitored at 425 nni for several seconds at 1-s intervals

D1202J

Fig. 8. Chemiluminescence ide as a function of time

zlO

determination of hydrogen perox-

Reaction condItions: 50 p1 of H202 standard, 50 pP of 1 X iO mol/Ilter K3Fe(CP’l)6,50 pPof iumlnol (2 g/llter). and monitoring wavelength. 425 nm

background, many CL systems have remarkable sensitivity. One of the most efficient of these systems is based on the reaction of luminol (5-amino-2,3-dihydro-1,4-phthalazinedione) with hydrogen peroxide (Figure 7). Figure 8 shows the response of the miniature CFA to this reaction (as “catalyzed” by potassium hexaferricyanide) as a function of time and peroxide concentration The signal decays with time, so rapid mixing and observation are desirable. With use of the conventional sequential transfer rotor design, transfer and mixing of samples and reagents require a total of 5 to 6 s. A new design with parallel sampleand reagent-transfer chambers (15) should be capable of shortening this time to 3 a or less. The luminol/hydrogen peroxide reaction has been shown to have several useful analytical applications, primarily in the general area of trace metals analysis (24). This reaction also has clinical applications in that it can be coupled with reactions in which hydrogen peroxide is used or generated. Thus, assays for 1230 CLINICALCHEMISTRY,Vol. 21, No. 9, 1975

specific substrates may be developed. For example, a CL method for glucose in which glucose oxidase (EC 1.1.3.4) was coupled with the luminol reaction has been previously reported (25, 26). Other enzymes which potentially may be coupled to the luminol reaction include galactose oxidase (EC 1.1.3.9), various amino acid-specific oxidases, and purine oxidases such as xanthine oxidase (EC 1.2.3.2) and uricase (EC 1.7.3.3). Another enzyme that has potential as an analytical reagent is cholesterol oxidase (27). This enzyme catalyzes the oxidation of cholesterol to cholestenone with a stoichiometric production of hydrogen peroxide. Therefore, hydrogen peroxide may be quantitatively determined and related to cholesterol concentration (27, 28). In Figure 9, the CL response of various solutions of cholesterol of known concentration to the cholesterol oxidase/luminol reaction is shown as a function

of time.

This

response

can be integrated

over a selected time interval and, when plotted as a function of cholesterol concentration (Figure 10), indicates a linear response concentration relationship. Such determinations can then serve as a basis for a cholesterol method. Another efficient CL system, which is based on the luciferin/ATP/Mg(II) reaction, may be used to assay ATP and to determine bacteriological biomass (29).

Fluorescence and Nephelometric Light Scattering Measurements Optical constraints

necessitated light

scatter

performance. To date, physical or other on the original Fast Analyzer design have

the measurement by front-surface

of fluorescence

and

(18, 30) or near-forward

angle (31, 32). The greatest sensitivity, however, is expected from right-angle excitation, which decreases interference from reflected incident radiation. Be-

1.0

0.8

z a /

0.6

0.6

//

O.4

a 0 0 U. > C

a 0.2

0

50

00

ISO

200

220

300 4

[METHYL

CHOLESTEROL

Fig. 10. Cholesterol concentration (g/liter) as a function of

chemiluminescenceintensity

$

UMBELLIFERONE]

12 ,1O6

Fig. 12. Fluorescence signal enhancement by use of an evaporation cover with a reflective surface on the under side of the cover Excitation wavelength, 360 nm:emissionwavelength, 440 nm

U 2

in Figure 3, which serves to collimate and focus the emitted radiation into the photosensitive element of the photomultiplier. Figure 11 illustrates the en-

‘I)

hancement

a

in sensitivity

(about

10-fold)

effected

by

use of the lens assembly. A further enhancement of the observed signal in luminescence and light scattering is achieved by placing an evaporation cover with a reflective surface over the (transparent) top window of the rotor, nor-

0

> C

IL

mal to the photomultiplier

4

[4-METHYL

8

12

LJMBELLIFERONE]

IO

Fig. 11. Fluorescence signal enhancement by use of the lens assembly shown in Figure 3 Excitation wavelength, 360 nm; emission wavelength. 440 nm

cause of the unique design features of the miniature CFA (6, 7), with its portable, self-contained rotors, it is possible to achieve right-angle excitation. In an earlier report (13), fluorescence emission was detected by an external photomultiplier tube (with separate power supply) and focusing lens assembly. With this configuration, the sensitivity of the system was found to be about 2.5 X 10-8 mol/liter of sodium fluorescein in NaOH (10 mmolfliter). The further improvements made in the multipurpose optical system described in this report have now increased the sensitivity of the analyzer to about 3.0 X 10b0 mole per liter of sodium fluorescein in 10 mmol/liter NaOH. Several factors are responsible for the increased sensitivity, including use of the lens assembly illustrated

tube.

In the case of dilute

samples, where rescattering of radiation (nephelometry) or self-absorption (luminescence) is negligible, signals are intensified because of the effectively larger solid angle of observation (Figure 12). In Figure 13, the performance of the miniature CFA is compared with that of the Aminco-Bowman spectrofluorometer relative to the fluorescence of the substrate 4-methylumbelliferone. At lower substrate concentrations the performance characteristics of the two instruments are comparable, but at higher concentrations there is evidence of “self-quenching” of the substrate in the Aminco-Bowman spectrofluorometer. The favorable geometry of the multi-purpose optical system apparently prevents such quenching, because it has not been observed with the miniature CFA. Evaluation of rotor designs. Studies in which the clear-body (Figure 4C, D, and E) and insert (Figure 4F) rotors were compared and evaluated showed that the two designs had cuvet-to-cuvet variations of ±3% and ±6 to 8%, respectively. However, this variation is not an analytical problem, because it can be decreased to ±0.6% by use of the calibration procedure described in Methods. CLINICAL CHEMISTRY. Vol. 21, No. 9, 1975

1231

IC

00

06

02

I [4-METHYL

0

2

uMMELLIFE000E]

.101 4

0

Fig. 13. Comparison of the fluorometric performance of the miniature Centrifugal Fast Analyzer and the Aminco-Bowman spectrofluorometer for the substrate 4-methylumbelliferone Excitation wavelength, 360 nm; emission wavelength, 440 nm

0

IS

12 [4-METHYL

20

4

UMBELLIFEOSNE]

6

12

IS

20

.101

Fig. 14. improvement in precision of fluorescence ments with the use of a cuvet calibration protocol

measure-

‘I-bars” represent plus or minus two times the standard deviation for four

replicate determinations at each concentration of 4-methylumbelllferone. All measurementswere made within a single rotor (Insert rotor, Figure 4F). Excitation wavelength. 380 nm; emission wavelength, 440 nm

Figure 14 shows the effectiveness of the calibration procedure in correcting for the large cuvet-to-cuvet variation seen with the insert rotor. In this figure, each data point represents the mean of four replicate

or in cases where the excitation and emission lengths approach one another (40 nm or less).

samples

tively

at each

umbelliferone; deviation

of four

concentrations

of 4-methyl-

plus or minus two times the standard

is shown

for each

of these

means

for both

uncorrected and corrected data. The average coefficient of variation for the groups of values is 7.4% for the uncorrected data, and 6.0% for the corrected data,

although

the mean

intensities

are nearly

identi-

cal. To test the reproducibility of the individual cuvet factors, four calibrations of an insert rotor were performed, with a complete cleaning of the cuvets between runs (8). The run-to-run coefficient of variation of the individual cuvet factors averged about 1%. A significant problem encountered with the clearbody rotors is loss of sensitivity as the result of reflected light within individual cuvets or of reflected or scattered light that is transmitted through the transparent body of the rotor. (This loss of sensitivity has not been observed with the insert rotor because of its nonreflective, opaque black body). To minimize this problem, the rotor shown in Figure 4E has been modified by machining a slot into the rotor between each individual cuvet. When this rotor is placed into the analyzer, these slots then mate with black fins located in the rotor holder of the analyzer. In this way, each cuvet is separated by a black partition and each has a black surface for its top cover windows. This modification has decreased the cuvet-to-cuvet reflectance and scattering problems of the clear-body rotor; however, within-cuvet reflectance is still a problem. It should be noted that the loss of sensitivity with the clear-body rotors is only significant when small nephelometric measurements are being made 1232 CLINICAL CHEMISTRY, Vol. 21, No. 9, 1975

In summary, simple

the clear-body and

inexpensive

design,

wave-

which is rela-

to fabricate,

is pre-

ferred for fluorescence measurements and for many light-scattering experiments where the excitation wavelength exceeds 300 nm. The insert design is necessary where ultraviolet excitation is required, and may

be desirable

for relatively

small

nephelometric

measurements. The clear-body design is inappropriate for chemiluminescence, where no excitation radiation is necessary and the signal lifetime is large relative to cuvet rotation speed; the conventional black-body rotor or the slot-partition clear-body rotor is preferred in such cases. The technical assistance of J. B. Overton and R. C. Lovelace in performing many of the experiments is greatly appreciated. The indispensable technical assistance of M. L. Bauer, W. A. Walker, and R. A. Mathis in designing, fabricating, and assembling the system is also acknowledged. We are indebted to Dr. August Wahienfeld (Boehringer Mannheim GmbH) for supplying the cholesterol assay kit. Work supported by the NIGMS, NIH; NASA; the National Center for Toxicological Research; and the Energy Research and Development Administration.

References 1. Anderson,

N. G., Basic

principles

of Fast

Analyzers.

Am. J.

Clin. Pat hol. 52, 778 (1970). 2. Anderson, N. G., The development of Fast Analyzers. Z. Aria 1. Chem. 261, 257 (1972). 3. Burtis, C. A., Johnson, W. F., Attrill, J. E., et al., Increased rate of analysis by use of a 42-cuvet GeMSAEC Fast Analyzer. Clin. Chem. 17, 686 (1971). 4. Tiffany, T. 0., Johnson, G. F., and Chilcote, M. E., Feasibility of multiple simultaneous enzyme assays for diagnostic purposes with the GeMSAEC Fast Analyzer. Clin. Chem. 17, 715 (1971). 5. Anderson, N. G., Burtis, C. A., Mailen, J. C., et al., The feasibility of miniaturizing a Fast Analyzer. Anal. Lett. 5, 153 (1972).

6. Burtis, C. A., Mailen, J. C., Johnson, W. F., et al., Development of a miniature Fast Analyzer. Clin. Chein. 18,753 (1972).

19. Tiffany, ry analysis.

7. Scott, C. D., and Burtis, C. A., A miniature tem. Anal. Chem. 45, 327A (1973).

20. Coleman, Turbidimetry 1031 (1972).

Fast Analyzer

sys-

8. Burtis, C. A., Johnson, W. F., Mailen, J. C., et al., Development of an analytical system based around a miniature Fast Analyzer. Clin. Chem. 19,895 (1973). 9. Scott, C. D., and Mailen, J. C.,Dynamic introduction of whole blood samples into Fast Analyzers. Clin. Chem. 18,749(1972). 10. Mailen,

J. C., Overton,

Techniques

for fabrication

iature

Fast Analyzer.

Anal.

C. A., and Scott, C. D., and assembly of rotors for use in a minLett. 6, 245 (1973). J. B., Burtis,

11. Burtis, C. A., Johnson, W. F., and Overton, J. B., Automated loading of discrete, microliter volumes of liquids into a miniature Fast Analyzer. Anal. Chem. 46, 786 (1974).

T. 0., Centrifugal

Fast

Analyzers

in clinical

laborato-

Crit. Rev. Clin. Lab. Sci. 5, 129 (1974). R. L., Shulta, W. D., Kelley, M. T., and Dean, J. A., via parallel photometric analysis. Anal. Chem. 44,

21. Bloom, M., and Hjorne, N., Immunochemical determination of serum albumin with a centrifugal analyzer. Clin. Chem. 21, 195 (1975). 22. Bostick, W. D., Bauer, M. L., Morton, J. M., and Burtis, C. A., Coagulation time determination with automatic, multivariable analysis, by use of a miniature Centrifugal Fast Analyzer. Clin. Chem. 21, 1288 (1975). 23. Erskine, A. G., The Principles and Practice ing, C. V. Mosby Co., St. Louis, Mo., 1973, p 192. 24. Cormier,

nescence

M. J., Hercules,

and Bioluminescence,

of Blood Group-

D. M., and Lee, J., Eds., ChemilumiPlenum Press, New York, N. Y.,

12. Burtis, C. A., Johnson, W. F., Overton, J. B., et al., Optimization and analytical application of the technique of dynamic introduction of liquids into centrifugal analyzers. Clin. Chem. 20, 932 (1974).

25. Bostick, D. T., and Hercules, D. M., Enzyme-induced chemiluminescence-determination of blood glucose using luminol. Anal. Lett. 1, 347 (1974).

13. Tiffany, T. 0., Parella, J. M., Johnson, W. F., and Burtis, C.

26. Auses, J. P., Cook, S. L., and Maloy, J. T., Chemiluminescent

A., Specific protein analysis by light scatter measurement with a miniature Centrifugal Fast Analyzer. Clin. Chem. 20, 1055 (1974).

enzyme method for glucose. Anal. Chem. 47, 244 (1975). 27. Richmond, W., Preparation and properties of a cholesterol oxidase from Nocardia sp. and its application to the enzymatic assay of total cholesterol in serum. Clin. Chem. 19, 1350(1973). 28. Allain, C. C., Poon, L., Chan, S. G., et al., The enzymatic determination of total serum cholesterol. Clin. Chem. 20, 470’(1974).

14. Tiffany, T. 0., Parella, J. M., Burtis, ing with a miniature Centrifugal Fast 1043 (1974).

C. A., et al., Blood

group20,

Analyzer. Clin. Chem.

15. Burtis, C. A., Johnson, W. F., and Tiffany, T. 0., The development of rotors having separate sample and reagent transfer channels for use with Centrifugal Fast Analyzers. Anal. Lett. 7, 591 (1974). 16. Tiffany, T. 0., Jansen, J. M., Burtis, C. A., et al., Enzymatic kinetic rate and end-point analyses of substrate using a GeMSAEC Fast Analyzer. Clin. Chem. 18, 829 (1972). 17. Tiffany, T. 0., Chilcote, D. D., and Burtis, C. A., Evaluation of kinetic enzyme parameters by use of a small computer interfaced “Fast Analyzer”-an addition to automated clinical enzymology. Clin. Chem. 19,908 (1973). 18. Tiffany, T. 0., Watsky, M. B., Burtis, C. A., and Thacker, L. H., Fluorometric Fast Analyzer: Some applications to fluorescence measurements in clinical chemistry. Clin. Chem. 19,871 (1973).

1973, p 427,

29. Seliger, minescence. Cormier, D. York, N. Y.,

H. H., Applications of bioluminescence and chemiluIn Chemiluminescence and Bioluminescence, M. J. M. Hercules, and J. Lee, Eds., Plenum Press, New 1973, p461.

30. Tiffany, T. 0., Burtis, C. A., Mallen, J. C., and Thacker, L. H., Dynamic multicuvette fluorometer-spectrophotometer based on the GeMSAEC Fast Analyzer principle. Anal. Chem. 45, 1716 (1973). 31. Buffone, G. J., Cross, R. E., Savory, J., and Soodak, C., Measurement of laser induced front surface light scattering with a parallel Fast Analyzer system. Anal. Chem. 46, 2047 (1974). 32. Buffone, G. J., Savory, J., and Cross, R. E., Use of a laserequipped centrifugal analyzer for kinetic measurement of serum IgG. Clin. Chem. 20, 1328 (1974).

CLINICALCHEMISTRY, Vol. 21, No. 9, 1975 1233