TY - JOUR
AU - Hafeman, Dean, G
AB - Pipettors and liquid handling systems should be checked regularly to verify their precision and accuracy. Professional organizations recommend performing 4-replicate tests monthly and 10-replicate tests at least quarterly (1)(2)(3)(4). Traditional gravimetric procedures are adequate to meet these testing requirements for single-channel pipettors but are tedious and impractical for multichannel devices. An alternative is to dispense a solution of colored dye into wells of a microplate and to measure the resulting absorbance values in a microplate reader. Volume calibration is performed with a calibration curve relating absorbance to volume (5). We recently reported a method for verifying multichannel pipettor performance by a spectrophotometric procedure that utilizes the near infrared absorbance of water and does not require addition of a dye (6). Water or other aqueous reagent is dispensed from the pipettor into microplate wells, and the optical pathlength in each well is determined in a microplate spectrophotometer. Water is essentially transparent from 200 to 900 nm but has a distinctive absorbance peak near 977 nm. As predicted by the Lambert law of light absorption, absorbance is proportional to the distance that light travels through the sample; thus the characteristic absorbance of water can be utilized to measure the pathlength of an aqueous sample. The maximum absorbance is affected by temperature; however, temperature dependency can be avoided by making the absorbance measurements at a temperature isosbestic point (near 1000 nm). Baseline absorbance is measured at a wavelength distant from the water absorbance peak, e.g., 900 nm, where again the absorbance is independent of temperature. The pathlength through an aqueous reagent in a microplate well is calculated from the difference between peak and baseline absorbance in the well and the value obtained by making the same measurements on the reagent in a standard 1-cm cuvette: \[\ \frac{(A_{1000}\ -\ A_{900})\mathrm{\ }_{\mathrm{Reagent\ in\ Well}}}{\mathrm{(A}_{\mathrm{1000}}\mathrm{\ -\ A}_{\mathrm{900}}\mathrm{)\ }_{\mathrm{Reagent\ in\ 1-cm\ Cuvette}}}\mathrm{{=}Pathlength\ in\ Well\ (cm)}\] In the published procedure, the lower end of the range of dispense volumes (30 μL) is limited by variable meniscus formation in nearly empty wells. We have now devised a modification of that method to extend the range to much lower volumes. By using half-area microplates and an incremental pipetting method, dispense volumes of 4 μL or less can be accommodated. The inset in Fig. 1 illustrates the principle of the method. The first step is to put sufficient aqueous reagent into a microplate well to cover the bottom and to establish a uniform meniscus. A measurement of the initial optical pathlength (P1) is made. Without delay, the aqueous dispense volume is pipetted, and a second measurement of the optical pathlength (P2) is made. The difference between P2 and P1 is the pathlength increment associated with the dispense volume pipetted. In the example presented below, the initial volume of water was 25–40 μL, and subsequent dispense volumes pipetted were 1.0–30 μL. Half-area microplates (Corning Costar Corp.) were used to maximize sensitivity. All absorbance measurements (at 900 nm and 1000 nm) for pathlength calculations were made in a microplate spectrophotometer (SPECTRAmax PLUS, Molecular Devices Corp.) as described previously (6). Figure 1. Open in new tabDownload slide Calibration curve of pathlength vs incremental volume obtained with 1.0–30 μL water. The regression equation for a linear curve fit was y = 0.06204 mm/μL (x) + 0.00054 mm; r2 = 0.9998; n = 40 wells per data point. Standard deviations ranged from 0.010 to 0.015 mm (error bars on plot are ± 0.015 mm). The inset shows a diagram of the optical pathlength measurements made in a microplate well (not drawn to scale). P1 represents the initial pathlength, and P2 represents the pathlength after dispensing an aliquot from a pipettor or liquid handling system. The wells typically are not cylindrical, but tapered slightly outward. Figure 1. Open in new tabDownload slide Calibration curve of pathlength vs incremental volume obtained with 1.0–30 μL water. The regression equation for a linear curve fit was y = 0.06204 mm/μL (x) + 0.00054 mm; r2 = 0.9998; n = 40 wells per data point. Standard deviations ranged from 0.010 to 0.015 mm (error bars on plot are ± 0.015 mm). The inset shows a diagram of the optical pathlength measurements made in a microplate well (not drawn to scale). P1 represents the initial pathlength, and P2 represents the pathlength after dispensing an aliquot from a pipettor or liquid handling system. The wells typically are not cylindrical, but tapered slightly outward. Pathlength/volume calibration curves were prepared using single-channel pipettors to dispense defined volumes of reagent into microplate wells. The calibration of the pipettors at each of the intended dispense volumes was verified by dispensing aliquots of water and weighing them with a certified analytical balance. The aliquots were typically dispensed and absorbance measurements made in sets of eight (one microplate column at a time). To control evaporation error, <1 min was allowed to elapse between pipetting and pathlength measurement. Fig. 1 shows the relationship between optical pathlength and incremental volumes between 1.0 and 30 μL. The slope of the regression equation was 0.06204 mm/μL, and standard deviations were typically 0.015 mm throughout the range. The standard deviation for volume determination (0.015 mm ÷ 0.06204 mm/μL) thus is ∼0.24 μL. Additional precision and confidence may be achieved by conducting multiple dispense and measurement cycles and averaging the results. An important determinant of the precision of the method is the reproducibility of the absorbance measurements used to calculate pathlengths. To assess this instrumental precision, microplates filled with 40 μL water/well were read 10 times. Mean A1000 and A900 values were 0.0763 and 0.0407, respectively. Average single-well standard deviations of A1000 and A900 were 0.00015 and 0.00004, respectively. Determination of pairwise A1000–A900 values gave a standard deviation of 0.00015, which corresponds to an optical pathlength difference of 0.010 mm (or 0.16 mL). This instrumental imprecision is only slightly less than the imprecision of the entire method. Thus, improved instrumental precision or averaging of replicate measurements promises to improve the precision of the method still further. As a word of caution, we note that microplate wells are not perfectly cylindrical, but are slightly conical (Fig. 1 , inset). Thus, the pathlength increment for a given dispense volume decreases as the total volume in the well increases. According to Corning product literature, the internal diameter of a half-area microplate well is 4.5 mm at the bottom and 5.0 mm at the top. Thus, the pathlength/volume ratio differs by ∼20% between the top and bottom of the well of Corning-Costar half-area microplates. Pathlength may be related to volume by construction of a nonlinear calibration curve. Alternatively, we have found it convenient to use a simple approximation to linearize the calibration curve. Preliminary dispense reagent volume (VP) is adjusted such that the midpoint of the dispense volume (VD) is the same for all dispense volumes, that is VP VD/2 = Constant (we chose 40 μL). For example, with 1, 10, 20, and 30 μL dispense volumes, 39.5, 35, 30, and 25 μL volumes of preliminary dispense reagent are used, respectively. The resulting calibration curve (Fig. 1) is linear to within the limits of precision of the method. To check relative precision of a multichannel pipettor or a liquid dispensing system, a calibration curve is not necessary. One simply dispenses the fluid into the wells and makes the pathlength measurements. The CV of volume pipetting will equal or exceed the CV of measured pathlength values. To check accuracy, a calibration curve such as that shown in Fig. 1 is required to relate pathlength to volume. The curve, once obtained, can be stored for subsequent use, provided that the aqueous dispense reagent, the lot of microplates, and values of VP are the same. The method described above offers a fast, convenient, and reliable procedure to verify performance of multichannel pipettors and automated microplate liquid dispensers down to 4 μL or less. The method avoids the need to add a dye to the dispense reagent. The performance check can be done with the same buffers or reagents that are routinely dispensed by the pipettor. Thus, discrepancies in pipettor performance between calibrator solution and routine dispense reagent are eliminated. Molecular Devices Corporation, 1311 Orleans Drive, Sunnyvale, CA 94089 References 1 . National Committee for Clinical Laboratory Standards. Determining performance of volumetric equipment 1984 : 1 -177 NCCLS NCCLS guideline. Villanova, PA. . 2 . College of American Pathologists. Laboratory instrument evaluation, verification and maintenance manual 4th ed. 1989 : 115 -116 College of American Pathologists Northfield, IL. . 3 . American Society for Testing and Materials. Standard specifications for piston or plunger operated volumetric apparatus designation 1993 : E1 -E9 ASTM Philadelphia, PA. . 4 Curtis RH. Performance verification of manual action pipets. Am Clin Lab 1994 ; 12 : 8 -9. 5 Kaufman D, Wobig GH. Colorimetric calibration of multichannel pipettes [Letter]. Clin Chem 1984 ; 30 : 1885 -1886. Crossref Search ADS PubMed 6 McGown E, Hafeman D. Multichannel pipettor performance verified by measuring pathlength of reagent dispensed into a microplate. Anal Biochem 1998 ; 258 : 155 -157. Crossref Search ADS PubMed © 1998 The American Association for Clinical Chemistry This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/open_access/funder_policies/chorus/standard_publication_model)
TI - Verification of Multichannel Liquid Dispenser Performance in the 4–30 μL Range by Using Optical Pathlength Measurements in Microplates
JF - Clinical Chemistry
DO - 10.1093/clinchem/44.10.2206
DA - 1998-10-01
UR - https://www.deepdyve.com/lp/oxford-university-press/verification-of-multichannel-liquid-dispenser-performance-in-the-4-30-5Z24yyhwXu
SP - 2206
EP - 2208
VL - 44
IS - 10
DP - DeepDyve
ER -