Article pubs.acs.org/Langmuir
Characterizing the Surface Quality and Droplet Interface Shape for Microarray Plates M. J. Schertzer,† M. J. Ahamed,†,§ R. Ben-Mrad,† P. Lea,‡ and P. E. Sullivan*,† †
Department of Mechanical and Industrial Engineering, University of Toronto, Toronto, Ontario, Canada M5S 3G8 SQI Diagnostics, 36 Meteor Dr, Toronto, Ontario, Canada M9W 1A4 § Mechanical and Aerospace Engineering, University of California at Irvine, Irvine, California, United States 92697-3975 ‡
ABSTRACT: The variation in the surface quality of microarray plates was examined by measuring the contact angles of 480 droplets on five microarray plates. It was found that the measured contact angle did not accurately predict the droplet shape for moderate Bond numbers (∼0.5 ≤ NB ≤ 5). By defining an apparent contact angle using the ratio of the contact radius to the height, the variance in the predicted interface shape decreased by greater than a factor of 3 for both local and globally averaged characteristics. The error in the predicted droplet height was also reduced by 3 orders of magnitude.
I. INTRODUCTION A deoxyribonucleic acid (DNA) microarray is an orderly arrangement of samples spotted or printed on a substrate at fixed locations. Microarray technology is often used to perform gene expression analysis [e.g., refs 1−7] and spotting [e.g., ref 8]. DNA microarrays deposit picoliter to microliter sized spots onto a substrate or plate using different actuation principles.9−14 Typical microarray substrates include glass, silicon, and plastic.2 Of the three, glass is the most commonly used substrate because of its chemical resistance, mechanical stability, and fluorescence properties, and its surface chemistry can be easily modified with suitable coatings.15 The homogeneity of droplets deposited on the substrate has been identified as one of the important performance parameters for microarrays in practical applications.16,17 Variation of droplet volume and size across the plate may influence the gene expression level measurement. A nonuniform surface coating causes nonuniformities in droplet shape and distribution of cDNA molecules, causing errors in signal processing. Characterization of microarray plates has been performed using fluorescence measurements, atomic force microscopy, and contact angle measurement12,18−21 to determine microarray density, dispensing accuracy, and spot to spot consistency. These tests do not fully characterize surface quality and spot shape distribution across the plate. Atomic force microscopy (AFM) combined with contact angle measurements does provide detailed surface characterization of surface quality over microarray plates,22 but the high cost of this process has limited it to the characterization of micro to nanoscale surface areas where practical applications may require producing uniform spots areas on the order of 6 cm2.23 Measurement of the contact angle, the angle between liquid and substrate as measured within the droplet, is often used as an important tool for characterization of the wettability of surfaces.24−26 This accurately characterizes the surface proper© 2012 American Chemical Society
ties of a DNA microarray [e.g., ref 27]. For example, a hydrophilic spot on the plate increases the rate at which a droplet spreads across the microarray surface resulting in more volume of fluid transferred from the pin. Conversely, a hydrophobic spot results in less volume transfer from the dispenser. The hydrophobicity of a microarray surface is a function of the choice of coating on the substrate and the local quality of that coating. Surface quality also influences the biological response of materials.28 Overhead images of droplet contact area have been used as a proxy for contact angle measurements in an automated high-throughput system to evaluate the wettability of various polymers.26 It is important to characterize the uniformity of a microarray surface to detect variations in surface quality. Contact angle measurements provide information about the relative local hydrophobicity on the surface and a complete characterization of the surface would characterize the surface uniformity. Consistent droplet size and shape on the plate is desirable to produce a high quality microarray.4,8 Conventional contact angle measurements are performed by capturing images of a sessile droplet viewing from its side [e.g., refs 29−31]. The shape of the liquid droplet on a solid surface is influenced by the contact angle.32 The relative importance of gravitational to surface tension forces on the shape of a droplet is characterized by the Bond number, NB NB =
2gρF R2 γ
(1)
where ρF is the fluid density, g is gravitational acceleration, R is the radius of curvature, and γ is the surface tension between the fluid and the surrounding medium. Taylor et al.33 conducted Received: March 3, 2012 Published: June 6, 2012 9961
dx.doi.org/10.1021/la302091t | Langmuir 2012, 28, 9961−9966
Langmuir
Article
Figure 1. Picture of the experimental facility used in this investigation. This facility includes (1) LED array, (2) microarray plate, (3) manual jack and Delrin block, (4) stereo microscope, (5) camera, and (6) image processing system. the microarray plate was fixed to the Delrin block using double sided tape. In all cases, physical contact between the underside of the microarray plate and the grid plate was avoided by fixing two small pieces of single sided tape to each corner of the grid plate and the midpoint of the rear edge to minimize surface impurities on the underside of the microarray plate. Calibration of this optical measurement system was performed by Schertzer et al.,34,35 who demonstrated that the average absolute error in both principle axes in the measurement area was always below 2 pixels. CPL imaging software was used to create bitmap images that were analyzed using the image processing toolbox in MatLab to determine the shape of the droplet interface. A typical droplet image is shown in Figure 2. This image was vertically divided into five equal sections which were independently converted into binary images to decrease errors in the measured interface shape due to nonuniformities in
contact angle measurements on polymers used in biological applications and showed that if the Bond number is less than one, the shape of the droplet interface is not influenced by gravitational force. This suggests that contact angle alone does not sufficiently characterize the variation of droplet shape on microarray plates with moderate NB (0.5 ≤ NB ≤ 5). The current investigation examines the uniformity of 2 μL droplets (NB ≈ 2) across the entire surface of five commercially available microarray plate surfaces. Although droplet shape is first characterized using the local and averaged values of the measured contact angle, it is found that the global and averaged ratio of droplet radius to height better characterizes droplet shape. In addition to providing a better prediction of interface shape, the radius to height ratio can be measured using lower resolution images decreasing the cost and time required to fully characterize the surface of a microarray plate.
II. EXPERIMENTAL FACILITY The experimental facility used in this study consists of a microarray plate and the optical measurement system (Figure 1). Three commercial microarray plates with (3-glycidoxypropyl) trimethoxysilane, 98% epoxy-silane coatings were used in this investigation. Each plate measured 74 × 110 mm. Each surface was divided into eight rows (A−H) and 12 columns, or 96 equally sized positions. A single 2 μL droplet of deionized water was placed on the microarray surface for visualization using a single-channel Finnpipette 0.5−10 μL pipet. To minimize evaporative loses, the time between deposition and imaging was always below 1 min. This process was repeated at each of the 96 locations on each surface. Droplets were imaged on both sides of the first two plates and on one side of the third (a total of 480 droplets). All plates were stored in a nitrogen desiccator when not in use to minimize surface fouling. Although all droplets were imaged on a single surface type, the measured contact angle varied between 37° and 60° in the range of other common microarray surfaces observed in refs 26 and 36. The optical measurement system used here consisted of a Leica MZ16F stereo microscope fitted with a 1x objective lens and a MS5K black and white camera from Canadian Photonics Laboratories (CPL). The CPL camera has a CCD array of 1280 × 1020 pixels where each pixel is 12 μm × 12 μm. The resolution of the optical measurement system was approximately 4 μm/pixel. All images in this investigation were backlit using an F&V Lighting Technology Z96 LED array. The optical measurement system was mounted on an x−y stage, which was used to focus the image and to adjust the field of view in the horizontal direction. The microarray plate was mounted on a custom machined Delrin block, which was mounted on a Melles Griot manual jack. The jack allowed for vertical adjustment of the field of view. The alignment of the plate and the microscope was controlled using custom-made fixtures. A grid plate showing the locations of the 96 well positions on
Figure 2. Typical images of (a) the entire field of view with the area of interest (dashed line), (b) the area of interest with the reconstructed interface shape (solid line), and (c) a sketch of a sessile droplet showing relevant geometric parameters. 9962
dx.doi.org/10.1021/la302091t | Langmuir 2012, 28, 9961−9966
Langmuir
Article
Figure 3. Surface plot showing a typical spatial distribution of the measured contact angle on the surface of a microarray plate. The surface shown here interpolates between 96 measurements (closed circles). lighting. The curves from the region of interest in each binary image were used to reconstruct the shape of the interface (e.g., Figure 2b). The contact radius (r), droplet height (h), and contact angle (θM) were determined from this reconstructed curve (Figure 2c). Second degree polynomial fits of the last 100 data point on each end were used to find the experimentally observed contact angle. An apparent contact angle was also determined using measured values of the droplet height and contact radius. Global averages of these parameters were also determined to predict the shape of the droplet interface using only the measured contact radius.
III. RESULTS AND DISCUSSION The contact angle of the droplet was measured in all cases by fitting a second degree polynomial to the first 100 points on the left and right most edges of the droplet interface. The contact angle was found by taking the derivative of these curves at the contact line and averaging the two results. A typical spatial distribution across a single plate is shown in Figure 3. In this case, the average measured contact angle was 52.6° with a standard deviation of 3.2°. Across all cases, the average contact angle was 50.2° with a standard deviation of 3.8°, whereas individual surfaces had average contact angles between 45.9° and 52.6° with standard deviations between 2.3° and 3.3°. The distribution of measurements was similar across all plates examined here which shows that it is possible to develop a method that can quickly and easily determine the distribution of the surface wettability across a commercial grade plate. The measured contact angle was used to generate a predicted interface shape and compare it to the experimentally observed interface (Figure 4a,b). Despite good agreement in the contact angle between the measured and predicted shape, the measured contact angle does not give an accurate prediction of the interface shape. For the cases examined here, the average absolute error in the shape predicted by the measured contact angle was approximately 120 μm, or 13.4% of the droplet height. This method also over predicted the height of the droplet by an average of 250 μm (28.8% of the droplet height). When the average contact angle across all cases was used to
Figure 4. Typical result for (a) the predicted interface and (b) error in the predicted shape using θM and (c) the predicted interface and (d) error in the predicted shape using r/h. Predictions for local (solid line) and characteristic (dashed line) values of θM, and r/h are compared to the experimentally observed interface (thick solid line). The resolution in these cases is approximately 4 μm per pixel, and the average height and contact radius were 0.9 mm and 1.9 mm, respectively.
predict the interface shape, the errors in the interface shape was 13.4% and 28.7%, respectively. Using the contact angle alone does not provide good predictions of the droplet interface as expected. In many microscale applications, the contact angle can be used to provide an accurate prediction of the curvature of the droplet interface. In these cases, surface tension forces dominate over gravitational effects and the radius of curvature is constant. The transition to mass independent interfaces occurs at radii of 9963
dx.doi.org/10.1021/la302091t | Langmuir 2012, 28, 9961−9966
Langmuir
Article
the system, it cannot capture the effects of gravity on the interface shape. The shape of the droplet interface can also be characterized using the ratio of the contact radius to the droplet height. A parabolic approximation of the interface (y(x)) and apparent contact angle (θA) can be defined as y(x) = (h/r 2)x ,
−r
