Home
Search
Collections
Journals
About
Contact us
My IOPscience
Self-driven filter-based blood plasma separator microfluidic chip for point-of-care testing
This content has been downloaded from IOPscience. Please scroll down to see the full text. 2015 Biofabrication 7 025007 (http://iopscience.iop.org/1758-5090/7/2/025007) View the table of contents for this issue, or go to the journal homepage for more
Download details: IP Address: 130.237.122.245 This content was downloaded on 22/05/2015 at 15:02
Please note that terms and conditions apply.
Biofabrication 7 (2015) 025007
doi:10.1088/1758-5090/7/2/025007
PAPER
RECEIVED
5 January 2015
Self-driven filter-based blood plasma separator microfluidic chip for point-of-care testing
REVISED
9 March 2015 ACCEPTED FOR PUBLICATION
Hojjat Madadi1,2, Jasmina Casals-Terré2 and Mahdi Mohammadi2
30 March 2015
1
PUBLISHED
22 May 2015
2
Center for Advanced Biomaterials for Health Care, Italian Institute of Technology, Naples, Italy Technical University of Catalonia, Mechanical Engineering Department, Terrassa, Spain
E-mail: [email protected] and [email protected] Keywords: lab on a chip, microchannel integrated micro-pillars (MIMP), point-of-care testing (POCT) Supplementary material for this article is available online
Abstract There is currently a growing need for lab-on-a-chip devices for use in clinical analysis and diagnostics, especially in the area of patient care. The first step in most blood assays is plasma extraction from whole blood. This paper presents a novel, self-driven blood plasma separation microfluidic chip, which can extract more than 0.1 μl plasma from a single droplet of undiluted fresh human blood (∼5 μl). This volume of blood plasma is extracted from whole blood with high purity (more than 98%) in a reasonable time frame (3 to 5 min), and without the need for any external force. This would be the first step towards the realization of a single-use, self-blood test that does not require any external force or power source to deliver and analyze a fresh whole-blood sample, in contrast to the existing time-consuming conventional blood analysis. The prototypes are manufactured in polydimethylsiloxane that has been modified with a strong nonionic surfactant (Silwet L-77) to achieve hydrophilic behavior. The main advantage of this microfluidic chip design is the clogging delay in the filtration area, which results in an increased amount of extracted plasma (0.1 μl). Moreover, the plasma can be collected in one or more 10 μm-deep channels to facilitate the detection and readout of multiple blood assays. This high volume of extracted plasma is achieved thanks to a novel design that combines maximum pumping efficiency without disturbing the red blood cells’ trajectory through the use of different hydrodynamic principles, such as a constriction effect and a symmetrical filtration mode. To demonstrate the microfluidic chip’s functionality, we designed and fabricated a novel hybrid microdevice that exhibits the benefits of both microfluidics and lateral flow immunochromatographic tests. The performance of the presented hybrid microdevice is validated using rapid detection of thyroid stimulating hormone within a single droplet of whole blood.
1. Introduction Lab-on-a-chip devices are an attractive technology for clinical analyses and diagnostics due to their outstanding virtues, which include rapid operation, low requirements for samples and reagents, high convenience, and low cost [1]. Many pathologies and physiological conditions can be diagnosed through the analysis of blood plasma, which provides crucial information on various internal organs [2, 3]. For instance, various proteins are used as biomarkers to identify multiple cancers [4], and circulating nucleic acids serve as biomarkers for diagnosing malaria [5], stroke [6], and © 2015 IOP Publishing Ltd
several cancers [7]. To reduce the time from blood collection to testing and provide faster, less expensive, comprehensive results, lab-on-a-chip devices are an attractive tool for blood plasma separation and analysis when compared to traditional centrifugation processes. Currently, microfluidic blood plasma separation techniques can be classified in three main formats: paper-based microfluidics, the CD format, and the chip format [7]. Paper-based microfluidics are widely used in lateral-flow assays, and recently the use of paper to build microfluidic circuits has shown great potential for creating low-cost miniaturized bloodanalysis devices [9–11]. But this technology is
Biofabrication 7 (2015) 025007
H Madadi et al
relatively new, which means that issues such as control over the flow rates, the mixing process, and the interaction time between sample and reagents have not yet been perfected. The CD-based format has the potential to become a standard tool for blood plasma separation and micro total analysis systems [12, 13]. However, this format has its own limitations, such as the need for numerous valves and the fact that a high-speed spinner and electrical power are required to manipulate the fluid. In addition to these two approaches, there are different microfluidic technologies, which are summarized in [14–16]. In the chip format, the separation methods can be divided into two main categories: passive and active. Active methods need an external force field such as an acoustic [17, 18], electric [19, 20], or magnetic field [21, 22] to achieve separation. In contrast, passive separation methods do not require external forces; therefore, they are more desirable for point-of-care (POC) devices. In this category, researchers exploited several techniques such as sedimentation [23, 24], dead-end filtration [25], cross-flow filtration [26, 27], deterministic lateral displacement [28, 29], pinchedflow fractionation [30], and biomimetic separation methods [31, 32]. Since 2005, attempts have focused on generating self-driven microfluidic designs with high manufacturing accuracy to provide more plasma output. For instance, Crowley et al reported a capillary-driven in-plane cross-flow plasma separation design from whole blood [33]. In 2009, Gervais et al highlighted the usefulness of a capillary-driven multiparametric microfluidic chip for one-step immunoassays [34]. In this microdevice, an embedded pad was used to filter the blood plasma. This pad required 20 μl of blood sample. In 2010, Kim et al [35] presented a surfactantadded polydimethylsiloxane (PDMS) design, (PDMS + Silwet L-77), but the volume of the extracted plasma was small, on the order of 20 nl. In 2011, Dimove et al [36] presented a stand-alone microfluidic chip that extracted plasma from a 5 μl droplet of blood for an enzyme-linked immunosorbent assay using sedimentation in a trench structure. Later, in 2012, Sakamoto et al proposed a capillary-driven design operating on the wetting properties of poly (methyl methacrylate) (PMMA) [37]. Three percent of the volume of plasma from 5 μl of blood was extracted, but the instability of the PMMA surface-modification process and red blood cells (RBCs) clogging the entrance of the microchannel array were challenges that still remain unmet in the use of this process for POC manufacturing. Recently, Liu et al proposed a pump-free blood plasma separator [38] based on gravitational sedimentation and filtration. However, this device needs a large sample volume (1.8 ml) and also requires further manipulation of the plasma after the separation process. The same year, 2013, Kang et al [39] developed a 2
cyclic olefin copolymer whole-blood plasma separator driven by asymmetric capillary forces. Although this chip separates approximately 102 nl from a single droplet of whole blood, it requires complicated post-processing to achieve filtration. Despite all these attempts, we still lack a simple device that addresses all the requirements for building a batch-manufacturable POC device, including the ability to separate plasma from untreated fresh blood, a long shelf life, and a reasonable separation time. In this work, we design and fabricate a novel selfdriven blood plasma separation microfluidic chip. To validate the quality and quantity of the separated plasma and to show its potential as a clinical tool, the microfluidic chip has been combined with lateral flow immunochromatography technology to perform a qualitative detection of thyroid-stimulating hormone (TSH). The clogging delay caused by RBCs in a hydrophilic PDMS constriction channel and the symmetric out-of-plane cross-flow filtration microchannel integrated micropillars (MIMPs) channel are used to maximize the amount of extracted plasma from the untreated blood. The proposed microdevice has several advantages over the previously mentioned devices for blood-plasma separation. This microdevice is selfdriven, which means that capillary forces are sufficient to deliver and separate more than 0.1 μl of blood plasma from a 5 μl droplet of real fresh blood (the required sample for blood testing in a POC device in terms of quantity and quality). Moreover, it does not require filters, tubes, or plasma manipulation, and due to its simple design, it can be easily mass produced. This paper is organized as follows. The principle of operation of the microfluidic device and the fabrication procedure are described in section 2. The proposed blood plasma separation microfluidic device is experimentally tested and validated in section 3, which also includes the experimental validation of the plasma quality with analyte detection.
2. Materials and methods 2.1. Principle of operation of the self-driven microfluidic device The proposed microfluidic design combines several aspects—including separation science, fluid dynamics, and blood rheology—to maximize the volume of extracted plasma. The phenomenon of RBCs clogging the entrance of the filtration channel is an inevitable and major issue for passive blood-plasma filtration. The proposed design addresses this issue by increasing the shear rate on the RBCs and using a symmetric arrangement of the capillary forces to extract the plasma from the whole blood. This technique moves RBCs away from the entrance of the filtration channel and consequently delays the RBC clogging, thereby increasing the extracted volume of blood plasma.
Biofabrication 7 (2015) 025007
H Madadi et al
Figure 1. Principle of operation of the self-driven microfluidic device. The yellow arrows show capillary forces on the extracted plasma in the MIMP filtration channel, while the red arrow shows the shear force on the RBCs due to the constriction design. The main channel includes a transport channel (300 μm × 10 μm) and two constrictions (100 μm × 10 μm), while two plasma-collected channels (100 μm × 10 μm) are placed on both sides of the main channel. (See video 1 in the supplementary information avaliable at stacks.iop.org/BF/7/025007/mmedia).
Figure 1 shows a schematic view of the microfluidic chip design, which is manufactured in two parts. The top part comprises the 10 μm-deep main channel (called the transport channel and used to deliver the blood drop sample) and two plasma-collected channels of the same depth. The bottom part of the microdevice is called the MIMP filtration channel and includes an arrangement of diamond microposts to decrease the flow resistance and increase the capillary force. The proposed MIMP filtration channel not only filters RBCs, but also paves the way for collecting the separated plasma in the top part of the microdevice using multiple collection channels. The MIMP design is the result of our previous work on developing a high-throughput microcapillary pump with efficient, integrated, low-aspect-ratio micropillars [40]. The obtained results indicate that the diamond-pillar array presents the minimum resistance flow compared to other pillar shapes. Furthermore, because of the deformability of human RBCs, the depth of the bottom part of the device, HFC , is designed to be less than 2 μm to successfully filter the RBCs (1.5 μm thickness). Two different transport-channel designs— straight and curved—have been investigated and compared, as seen in figure 2(a). In the straight transportchannel design, the blood flow in the transport channel is perpendicular to the extracted blood plasma flow, while in the curved design, both flows are parallel. The fluidic behavior of the whole device can be analyzed using an equivalent electrical network 3
(figure 2(b)). RT1 and ΔPT1 are the flow resistance and the pressure drop of the first part of the transport channel (before the constriction channel). RC and ΔPC are the flow resistance and the pressure drop of the constriction channel. RT 2 and ΔPT 2 are the flow resistance and the pressure drop of the second part of the transport channel (after the constriction channel). R F and ΔPF are the flow resistance and the pressure drop of the MIMP filtration channel, and R PC and ΔPPC are the flow resistance and the pressure drop of the plasma-collected channel, respectively. Once the hydrophilicity of the material is fixed (surface tension), the capillary pressure, ΔP, for each of the rectangular microchannels can be calculated using the Young–Laplace equation: ΔPi = σ cos θ
⎛1 1 ⎞⎟ ⎜ + ⎝h w ⎠
(1)
where σ is the surface tension, θ is the contact angle, h is the half depth of the microchannel, and w is the half width of the microchannel [41]. On the other hand, the flow rate, Q, of a liquid in a capillary microchannel depends on the viscosity of the liquid, the capillary pressure, and the microchannel flow resistance, as follows: Q=
1 ΔP μ R
(2)
where ΔP is the pressure difference between the area inside and in front of the liquid (equation (1)), μ is the viscosity of the liquid, and R is the microchannel flow resistance [41].
Biofabrication 7 (2015) 025007
H Madadi et al
Figure 2. (a) Transport microchannel design layouts. (b) The equivalent electrical network of the microdevice. (c) Diamond-shaped post-array in the filtration area.
On this basis, the minimization of the MIMP channel flow resistance, R F, will desirably increase the plasma flow rate in the MIMP filtration channel. To reach this goal, according to Madadi et al [40], a diamond-post shape with increased side distance between the pillars is the most efficient design, and it can be calculated through equation (3). RF =
12μL ⎛⎜ 1 ⎞⎟ wH 3 ⎝ 2ϵ − 1 ⎠
(3)
where μ is the fluid viscosity, L is the length of the microchannel, w is the half width of the microchannel, H is the height of microchannel, and ε is the porosity [40]. Since the fluid and the depth of the channel are fixed, we have maximized the porosity to achieve a manufacturable device. Figure 2(c) shows the dimensions of the selected MIMP filtration channel. 2.2. Fabrication procedure The microfluidic device is manufactured in two main parts—the top part and the bottom part, which are fabricated using different procedures. 2.2.1. Manufacturing process, top part SU-8, a negative-tone photoepoxy, is used to create a master mold via photolithography. PDMS, a common material in microfluidic chips, is used to manufacture the microdevice’s top part due to its outstanding characteristics—including its elastomeric properties, 4
biocompatibility, gas permeability, optical transparency, ease of bonding to glass, and low manufacturing cost [42, 43]—by means of single soft-lithography process. In spite of all the favorable physical and chemical properties, the hydrophobicity of PDMS is a handicap when pumping aqueous solutions through microchannels by capillary forces alone. To overcome PDMS’s hydrophobicity and to control PDMS’s hydrophobic recovery over a long period of time, a nonionic surfactant (Silwet L-77) is added to the PDMS before the curing process [44, 45]. Using the surfactant to modify PDMS’s hydrophobicity not only guarantees the shelf life of the final POC device for mass production (because of the negligible hydrophobic recovery [44]), it also eliminates the nonspecific absorption of proteins or biological material. Further details of the fabrication process are explained in the supplementary information avaliable at stacks.iop. org/BF/7/025007/mmedia. 2.2.2. Manufacturing process, bottom part The MIMP filtration channel is fabricated on a glass substrate using both photolithography and wet chemical methods. Figures 3(a)–(c) illustrate the fabrication steps of the MIMP filtration channel. Further details on the fabrication process are available in the supplementary information avaliable at stacks.iop.org/BF/7/ 025007/mmedia.
Biofabrication 7 (2015) 025007
H Madadi et al
Figure 3. (a)–(c) MIMP filtration channel manufacturing steps. The microdevice’s top and bottom parts (d) before and (e) after plasma bonding.
2.2.3. Microdevice assembly and test strip integration The PDMS microchannels (top part) and the etchedglass MIMP filtration channel (bottom part) are bonded via an oxygen plasma (Gambetti, Italy) treatment (see figures 3(d) and (e)). Oxygen at 200 Torr is introduced into the oxygen plasma chamber, and the plasma runs for 30 s. The oxygen plasma parameters are chosen to achieve maximum bonding strength [46]. Figure 4 shows a photograph of the completed blood plasma separation microfluidic device. To demonstrate and validate the potential diagnostic value of the microdevice and the quality of the separated plasma, the microfluidic chip was combined with a lateral flow immuno-chromatographic strip for TSH detection (NADAL® TSH Test, Germany). This test measures the levels of TSH, a hormone that is produced and released by the pituitary gland. The pituitary can sense whether there is enough thyroid hormone in the bloodstream, and it releases TSH when it detects insufficient thyroid hormone. The official normal range for TSH in blood runs from approximately 0.5 to 4.5/5.0 μIU ml−1. Therefore, a TSH level under 0.5 μIU ml−1 indicates hyperthyroidism (an overactive thyroid), and a TSH over 5.0 μIU ml−1 indicates hypothyroidism (an underactive thyroid). The separated plasma is driven to the entrance of half of a TSH test strip, as shown in figure 5. The filtration pad (fiberglass) and the absorption pad (at the end of test strip) are removed from the test strip. The test strip gold-conjugate pad is overlapped with the blood-plasma reservoir and sealed with tape. In the gold-conjugate pad, the plasma antigens are recognized and bound to the gold-conjugated 5
antibodies, as shown in figure 5(a). When the gold nanoparticle complex reaches the test line, a strong bond is formed with the immobilized antibodies embedded in the nitrocellulose membrane, and a colored line appears. Since the required volume of the blood plasma for wetting the conjugate pad is relatively high, 10 μl of phosphate-buffered saline (purchased from Life Technologies S A, Madrid, Spain) are added to the lateral-flow assay (LFA) strip after the plasma has reached the conjugate pad.
2.2.4. Sample preparation Fresh human whole blood collected from a healthy volunteer using the finger-prick method is used for experiments. The sample is collected using the lancet of a POC device in every experiment. To ensure the accuracy of the input volume, the blood sample is pipetted. The micropipette is gently held in a horizontal or slightly downward position near the channel inlet to introduce the blood sample. In this way, the blood sample loading does not affect in the blood flow behavior inside the device. For the TSH blood test, two different types of blood samples are prepared. The first type is a 5 μl drop of healthy fresh human whole blood, and the second type is a 5 μl drop of blood with a high concentration of TSH, which is obtained from reconstituted blood by mixing high-TSH dried plasma (provided by GRIFOLS) with RBCs from regular blood. Then, the collected blood samples are mixed with 0.5 μl heparin (purchased from Pfizer Co., Madrid, Spain), a standard anticoagulant.
Biofabrication 7 (2015) 025007
H Madadi et al
Figure 4. Photograph of (I) PDMS part (top part), (II) glass part (bottom part), and (III) the final microdevice.
Figure 5. (a) Schematic view of the microfluidic chip, including LFA strip. (b) The steps for embedding the TSH test strip inside the microfluidic device.
6
Biofabrication 7 (2015) 025007
H Madadi et al
Table 1. Comparison between the straight and curved transport channel geometries. Microchip
Blood sample volume
Straight transport channel Curved transport channel
∼5 μl ∼5 μl
Average volume of extracted plasma
Average time (min)
0.09 μl ± 0.01 0.07 μl ± 0.01
4±1 6±1
2.2.5. Experimental setup To record images and movies of the blood plasma separation process and analyze the results, a digital camera (Tucsen ISH500, 5.0 M pixel,) is connected to a microinspection lens system (Optem zoom 125C with a broad 12.5:1 zoom range and a 20× objective). All quantitative data are analyzed and expressed as the mean ±the standard deviation.
3. Results and discussion To study the effect of the transport channel geometry on the blood-plasma separation, two different channel geometries are considered: a straight transport channel and a curved transport channel. The aim of the curved design is to maximize the first-instance filling rate of the MIMP channel before the RBCs clog the MIMP entrance due to the fact that the blood and extracted plasma flow in the same direction. The experiments are performed three times for each channel geometry. The volume of the extracted plasma is calculated by adding the volume of the MIMP channel, the plasma-collected channels, and the collected plasma in the outlet reservoir. The results are summarized in table 1, which shows the average of the recorded data. (See the supplementary information avaliable at stacks.iop.org/BF/7/025007/mmedia for the whole data). 3.1. Curved transport channel A 5 μl drop of fresh human whole blood is introduced into the channel inlet of the transport channel. The blood sample is drawn into the microdevice by capillary forces alone, due to the hydrophilicity of the surfactant-added PDMS (Silwet L-77). The separation process starts as soon as the blood flows into the transport channel and arrives at the MIMP filtration channel. Since the MIMP filtration channel depth is less than 1.5 μm, RBCs cannot penetrate into the plasma channel. In the separation zone, the RBCs’ velocity in the transport channel is slowed down, which leads to an accumulation of RBCs in the entrance of the separation area. The purpose of the constriction in the middle of the transport channel is to generate a local flow acceleration and delay the RBCs clogging of the filtration area. In the curved transport-channel design, although the constriction postpones the RBC-related clogging, the centrifugal force due to the curvature and the nonsymmetrical arrangement of the separation zone leads to precocious RBC clogging in the entrance of the 7
Purity % >98 80–85
MIMP filtration channel. Figure 6 illustrates the filtration performance for the plasma separation from whole blood in the curved transport channel. Using a curved transport channel, an average of 0.07 μl of blood plasma are extracted from a 5 μl droplet of fresh whole blood. In the separation process, the MIMP filtration channel is filled within 1 min due to the fact that the plasma flow direction is parallel to the blood flow in the transport channel (see figure 6(b)), while the plasma-collected channel is filled in 5 to 7 min (see figures 6(g) and (h)). Figures 6(b) and (c) show that the plasma-collected channel is initially filled from the top surface of the pillars in the collected channel due to the higher capillary force compared to the other zones. It is noticeable that although some air bubbles form inside the separation area and the plasma-collected channel during the first few minutes, they disappear in the following steps of the process due to the permeability of the PDMS. Obviously, these trapped air bubbles reduce the separation efficiency because they create a resistance to the capillary flow, but the formation of air bubbles can be controlled by unifying the wetting properties of the modified PDMS and the glass [47]. The purity of the extracted blood plasma is defined as: Purity = 1 −
CP CM
(4)
where C M is the number of RBCs in the main channel and C P is the number of RBCs in the plasma-collected channel [8]. Image processing is used to identify the purity of the extracted blood plasma from whole blood, which is around 80 to 85% for the curved transport-channel topology, as seen in figure 7(a). The measurement is done based on 90 images taken from three different regions of the filled plasma-collected channel using the described experimental setup. ImageJ software is used to count the number of RBCs in the same specified area (the considered volume is 5 × 105 μm3) of every image region. The average of the obtained results (the number of RBCs in the plasma-collected channel) compared to the hematocrit concentration of the initial sample (fresh whole blood) is 45% in our experiments. The curved transport channel fills the separation MIMP channel in less than 1 min. However, the centrifugal force on the RBCs attributed to this geometry and the nonsymmetrical capillary forces reinforce the RBCs’ accumulation in the entrance of the MIMP filtration channel and drive them into the MIMP
Biofabrication 7 (2015) 025007
H Madadi et al
Figure 6. Blood plasma separation process using the curved transport channel geometry. (a) The hollow transport channel (blood channel) and the plasma-collected channel before the introduction of the whole fresh blood in the microchip inlet. (b) The blood channel filled with fresh blood and the filled MIMP filtration channel with the separated blood plasma. (c) The starting point of the blood plasma filling process in the plasma-collecting channel. (d)–(f) The blood plasma filling process in the plasma-collected channel. (g) and (h) The plasma-collected channel completely filled with the extracted blood plasma after 5 to 6 min (see video 2 in the supplementary information avaliable at stacks.iop.org/BF/7/025007/mmedia).
Figure 7. (a) A comparison between the concentration of RBCs in the main blood channel and their concentration in the plasmacollected channel. (b) The effect of the curved geometry and the resulting centrifugal force on the RBCs (see video 2, supplementary information avaliable at stacks.iop.org/BF/7/025007/mmedia).
channel, and thus into the plasma-collected channel (see figure 7(b)). A video clip of the performance of the blood plasma separation microdevice using the curved channel is provided in video 2 in the supplemental information avaliable at stacks.iop.org/BF/7/ 025007/mmedia. 3.2. Straight transport channel To increase the separation efficiency and the purity of the extracted plasma and to avoid the effect of the 8
centrifugal force due to the curvature design, the straight transport channel is tested. An example of the microdevice’s performance can be seen in figure 8. The mechanism of separation is similar to the curved transport channel, but there are two main differences. First, the direction of the blood flow in the transport channel is perpendicular to the extracted plasma flow in the MIMP channel. Second, the capillary forces due to the plasma extraction are applied symmetrically on both sides of the RBCs that flow in the transport
Biofabrication 7 (2015) 025007
H Madadi et al
Figure 8. Blood plasma separation process using the straight main channel geometry. (a) The hollow transport channel (blood channel) and the plasma-collected channel before the introduction of the whole fresh blood droplet in the microchip inlet. (b) The blood channel filled with fresh blood and the MIMP filtration channel filled with the separated blood plasma. (c) The starting point of the blood plasma filling process in the plasma-collected channel. (d)–(f) The blood plasma separation and filling process in the plasma-collected channel. (g) and (h) The plasma-collected channel completely filled with the extracted blood plasma after 3 min. (see video 3 in the supplementary information avaliable at stacks.iop.org/BF/7/025007/mmedia).
channel. This symmetric plasma extraction does not disturb the RBCs’ trajectory in the transport channel, which leads to a delay in the RBC-related clogging of the filtration area. In the case of the straight transport channel, an average of 0.09 μl of blood plasma is successfully separated from a 5 μl droplet of whole fresh blood. According to the results, the plasmacollected channel is filled in 3 to 5 min with a purity of more than 98%. The obtained results confirm that the perpendicularity of the plasma flow decreases the amount of time required to fill the plasma-collected channel [48]. A video clip of the performance of the blood plasma separation microdevice using the straight channel is provided in video 3 in the supplementary information. Table 1 shows a comparison of the results between the straight and curved transport channel geometries. To investigate the effect of the passive capillary filling flow rate on the microdevice’s performance, three types of microdevices with different surfactant concentrations (Silwet L-77 concentration 1%, 2%, and 3% in PDMS) are fabricated. More details about the experiments and the obtained results are explained in the supplementary information. According to the obtained results, the whole 9
microdevice performance time is decreased by more than 25% to less than 3 min by increasing the surfactant (Silwet L-77) concentration from 1% to 3%. This reduction in filtration time addresses the bottleneck in the sample preparation procedure for some blood analyses in order to achieve the appropriate performance of the POC device. 3.3. Double-embedded plasma-collected channel To further investigate the ability of the proposed microdevice to perform different assays simultaneously on the same blood sample, another plasmacollected channel is added to the other side of the main blood-flow channel (1 mm apart). The filling of this second plasma-collected channel not only opens the possibility of performing multifunctional blood analysis, but also increases the volume of extracted plasma up to 0.1 μl from a 5 μl droplet of whole fresh blood with the same purity (more than 98%). Since the presented blood-plasma separation device is driven by capillary forces alone, no hemolysis is observed. More details about the hemolysis study can be found in the supplementary information. Moreover, a video clip of the performance of the double-embedded plasma-
Biofabrication 7 (2015) 025007
H Madadi et al
Figure 9. The results of the TSH test. (a) Negative TSH result using a hybrid microfluidic chip. (b) Positive TSH result using a hybrid microfluidic chip. (c) Negative TSH result using an LFA strip.
collected channel is provided in video 4 in the supplementary information avaliable at stacks.iop. org/BF/7/025007/mmedia. 3.4. Integration of plasma separation with analyte detection for TSH blood test To demonstrate the utility of the quantity and quality of the extracted plasma, the plasma from the collected channel is driven to the entrance of an LFA to qualitatively detect the TSH level. Two different types of tests are performed based on the 5 μl prepared samples of healthy fresh blood and reconstituted blood. Figure 9 shows an example of the results of the TSH tests using both the test-strip device and the hybrid microfluidic chip (the implemented test strip in the presented microfluidic device). Note that in figure 9, the hybrid microfluidic chip is shown from the back view due to the rotation of the test strip in the implementation process (see figure 5(b)). Three test devices are tested for every sample type, and the colored line appeared in the control section of the test strip 18 ± 2 min after adding the blood droplet. Figure 9 shows a picture of the obtained results from the hybrid microfluidic chip compared to the standard NADAL® TSH test. The TSH test result is negative and the colored line that appears in the control ‘C’ section confirms the correct operation of the TSH test in the presented hybrid microfluidic chip. Figure 9(b) also shows the results when high-TSH blood is processed. The results are positive, since a result line appears in addition to the control line. According to the results, the proposed design shows its potential as a clinical tool compared to previous works. First, the initial volume of the required sample, 5 μl, is considerably reduced compared to other works [34, 38]. Second, the volume of the extracted blood plasma, 0.1 μl, is enough to implement the TSH test and also is more than the obtained volume in previous filter designs [33, 35]. 10
Importantly, the use of a simple design, which is made of only two parts without post-processing (such as in [37, 39]) or embedding additional parts for filtration [34], makes it suitable for batch manufacturing. Furthermore, the capability of the proposed microdevice for separating and gathering blood plasma within more than one plasma-collected channel paves the way for blood panels in a single POC test.
4. Conclusion In this work, a novel, self-driven blood plasma separation microfluidic chip is designed, built, and validated through integration with a TSH qualitative test. The test works with a single droplet of whole blood (∼5 μl), and 0.1 μl of blood plasma is successfully separated and collected in 10 μm-deep channels in 3 to 5 min, with a high purity level of more than 98%. The long-term hydrophilicity of the employed material, the straightforward manufacturing process, and the achieved plasma volume (0.1 μl) pave the way for doing multiplex blood analyses simultaneously in the same device, as is promised in portable medical POC testing. Supporting information Additional information as noted in text. Ethics statement The whole human blood sample for the experiments was collected from a healthy volunteer using the finger-prick method. The protocol was approved by the committee of the ethics of the Technical University of Catalonia, Barcelona-Tech.
References [1] [2] [3] [4] [5]
Whitesides G 2014 Lab Chip 14 3125–6 Psychogios N et al 2011 PLoS One 6 e16957 Wagner J 2012 Biochemia Medica 22 24–38 Polanski M and Anderson N L 2007 Biomarker Insights 1 1−48 Franklin B S, Vitorino B L F, Coelho H C, Menezesneto A, Santos M L S, Campos F M F, Brito C F, Fontes C J, Lacerda M V and Carvalho L H 2011 PLoS One 6 e19842
Biofabrication 7 (2015) 025007
H Madadi et al
[6] Rainer T H, Wong L K S, Lam W and Yuen E 2003 Clin. Chem. 49 562−9 [7] Schwarzenbach H, Hoon D S B and Pantel K 2011 Nat. Rev. Cancer 11 426−37 [8] Kersaudy-Kerhoas M and Sollier E 2013 Lab Chip 13 3323−46 [9] Vella S J, Beattie P, Cademartiri R, Laromaine A, Martinez A W, Phillips S T, Mirica K A and Whitesides G M 2012 Anal. Chem. 84 2883–91 [10] Yang X, Forouzan O, Brown T P and Shevkoplyas S S 2012 Lab Chip 12 274−80 [11] Songjaroen T, Dungchai W, Chailapakul O, Henryand C S and Laiwattanapaisal W 2012 Lab Chip 12 3392−8 [12] Gorkin R, Park J, Siegrist J, Amasia M, Lee B S, Park J M, Kim J, Kim H, Madou M and Cho Y K 2010 Lab Chip 10 1758−73 [13] Amasia M and Madou M 2010 Bioanalysis 2 1701−10 [14] Sollier E, Rostaing H, Pouteau P, Fouillet Y and Achard J 2009 Sensors Actuators B 141 617−24 [15] Pamme N 2007 Lab Chip 7 1644–59 [16] Mukherjee S, Kang T G, Chen Y and Kim S 2009 Crit. Rev. Biomed. Eng. 37 517−29 [17] Petersson F, Aberg L, Sward-Nilson A M and Laurell T 2007 Anal. Chem. 79 5117−23 [18] Doria A, Patel M and Lee A P, μ TAS 2011 15th Int. Conf. on Miniaturized Systems for Chemistry and Life Sciences (Seattle, WA, 2–6 October 2011) [19] Kralj J G, Lis M T, Schmidt M A and Jensen K F 2006 Anal. Chem. 78 5019−25 [20] Liao S H, Chang C Y and Chang H C 2013 Biomicrofluidics 7 024110 [21] Furlani E 2007 J. Phys. D: Appl. Phys. 40 1313−19 [22] Han K H and Frazier A B 2006 Lab Chip 6 265−73 [23] Tachi T, Kaji N, Tokeshi M and Baba Y 2009 Anal. Chem. 81 3194−8 [24] Zhang X B, Wu Z Q, Wang K, Zhu J, Xu J J, Xia X H and Chen H Y 2012 Anal. Chem. 84 3780−6 [25] Li C, Liu C, Xu Z and Li J 2012 J. Biomed. Microdevices 14 565−72 [26] Aota A, Takashi S, Mawatari K, Tanaka Y, Sugii Y and Kitamori T 2011 Anal. Sci. 27 1173−8 [27] Sollier E, Rostaing H, Pouteau P, Fouillet Y and Achard J L 2009 Sensors Actuators B 141 617−24
11
[28] Huang L R, Cox E C, Austin R H and Sturm J C 2004 Science 304 987−90 [29] Inglis D W, Morton K J, Davis J A, Zieziulewicz T J, Lawrence D A, Austin R H and Sturm J C 2008 Lab Chip 8 925−31 [30] Rodriguez-Villarreal A I, Arundell M, Carmona M and Samitier J 2010 Lab Chip 10 211−19 [31] Yang S, Undar A and Zahn J D 2006 Lab Chip 6 871–80 [32] Kersaudy-Kerhoas M, Kavanagh D M, Dhariwal R S, Campbell C J and Desmulliez M P Y 2010 Lab Chip 10 1587–95 [33] Crowley T A and Pizziconi V 2005 Lab Chip 5 922–9 [34] Gervais L and Delamarche E 2009 Lab Chip 9 3330−7 [35] Kim Y C, Kim S H, Kim D, Park S J and Park J K 2010 Sensors Actuators B 145 861−8 [36] Dimov I K, Basabe-Desmonts L, Garcia-Cordero J L, Ross B M, Ricco A J and Lee L P 2011 Lab Chip 11 845–50 [37] Sakamoto H, Hatsuda R, Miyamura K and Sugiyama S 2012 Micro Nano Lett. 7 64−7 [38] Liu C, Mauk M, Gross R, Bushman F D, Edelstein P H, Collman R G and Bau H H 2013 Anal. Chem. 85 10463−70 [39] Lee K K and Ahn C H 2013 Lab Chip 13 3261−7 [40] Madadi H, Casals-Terre J, Lopez R C and Anfrez M S 2014 Microfluid Nanofluidics 17 115−30 [41] Bruus H 2007 Theoretical Microfluidics (Oxford: Oxford University Press) [42] Mata A, Fleischman A J and Roy S 2005 Biomed. Microdevices 7 281–93 [43] Van Poll M L, Zhou F, Ramstedt M, Hu L and Huck W T S 2007 Angew. Chem. Int. Ed. Engl. 46 6634–7 [44] Madadi H and Casals-Terre J 2013 J. Microsyst. Technol. 19 143–50 [45] Kim H T, Kim J K and Jeong O C 2011 Japan. J. Appl. Phys. 50 06GL04 [46] Bhattacharya S, Datta A, Berg J M and Gangopadhyay S J 2005 Microelectromech. Syst. 14 590−7 [47] Le N B and Hsu Y C 2010 Adv. Nat. Sci.: Nanosci. Nanotechnol. 1 015006 [48] Kim D, Yun J Y, Park S J and Lee S S 2009 Microsyst. Technol. 15 227–33
Search
Collections
Journals
About
Contact us
My IOPscience
Self-driven filter-based blood plasma separator microfluidic chip for point-of-care testing
This content has been downloaded from IOPscience. Please scroll down to see the full text. 2015 Biofabrication 7 025007 (http://iopscience.iop.org/1758-5090/7/2/025007) View the table of contents for this issue, or go to the journal homepage for more
Download details: IP Address: 130.237.122.245 This content was downloaded on 22/05/2015 at 15:02
Please note that terms and conditions apply.
Biofabrication 7 (2015) 025007
doi:10.1088/1758-5090/7/2/025007
PAPER
RECEIVED
5 January 2015
Self-driven filter-based blood plasma separator microfluidic chip for point-of-care testing
REVISED
9 March 2015 ACCEPTED FOR PUBLICATION
Hojjat Madadi1,2, Jasmina Casals-Terré2 and Mahdi Mohammadi2
30 March 2015
1
PUBLISHED
22 May 2015
2
Center for Advanced Biomaterials for Health Care, Italian Institute of Technology, Naples, Italy Technical University of Catalonia, Mechanical Engineering Department, Terrassa, Spain
E-mail: [email protected] and [email protected] Keywords: lab on a chip, microchannel integrated micro-pillars (MIMP), point-of-care testing (POCT) Supplementary material for this article is available online
Abstract There is currently a growing need for lab-on-a-chip devices for use in clinical analysis and diagnostics, especially in the area of patient care. The first step in most blood assays is plasma extraction from whole blood. This paper presents a novel, self-driven blood plasma separation microfluidic chip, which can extract more than 0.1 μl plasma from a single droplet of undiluted fresh human blood (∼5 μl). This volume of blood plasma is extracted from whole blood with high purity (more than 98%) in a reasonable time frame (3 to 5 min), and without the need for any external force. This would be the first step towards the realization of a single-use, self-blood test that does not require any external force or power source to deliver and analyze a fresh whole-blood sample, in contrast to the existing time-consuming conventional blood analysis. The prototypes are manufactured in polydimethylsiloxane that has been modified with a strong nonionic surfactant (Silwet L-77) to achieve hydrophilic behavior. The main advantage of this microfluidic chip design is the clogging delay in the filtration area, which results in an increased amount of extracted plasma (0.1 μl). Moreover, the plasma can be collected in one or more 10 μm-deep channels to facilitate the detection and readout of multiple blood assays. This high volume of extracted plasma is achieved thanks to a novel design that combines maximum pumping efficiency without disturbing the red blood cells’ trajectory through the use of different hydrodynamic principles, such as a constriction effect and a symmetrical filtration mode. To demonstrate the microfluidic chip’s functionality, we designed and fabricated a novel hybrid microdevice that exhibits the benefits of both microfluidics and lateral flow immunochromatographic tests. The performance of the presented hybrid microdevice is validated using rapid detection of thyroid stimulating hormone within a single droplet of whole blood.
1. Introduction Lab-on-a-chip devices are an attractive technology for clinical analyses and diagnostics due to their outstanding virtues, which include rapid operation, low requirements for samples and reagents, high convenience, and low cost [1]. Many pathologies and physiological conditions can be diagnosed through the analysis of blood plasma, which provides crucial information on various internal organs [2, 3]. For instance, various proteins are used as biomarkers to identify multiple cancers [4], and circulating nucleic acids serve as biomarkers for diagnosing malaria [5], stroke [6], and © 2015 IOP Publishing Ltd
several cancers [7]. To reduce the time from blood collection to testing and provide faster, less expensive, comprehensive results, lab-on-a-chip devices are an attractive tool for blood plasma separation and analysis when compared to traditional centrifugation processes. Currently, microfluidic blood plasma separation techniques can be classified in three main formats: paper-based microfluidics, the CD format, and the chip format [7]. Paper-based microfluidics are widely used in lateral-flow assays, and recently the use of paper to build microfluidic circuits has shown great potential for creating low-cost miniaturized bloodanalysis devices [9–11]. But this technology is
Biofabrication 7 (2015) 025007
H Madadi et al
relatively new, which means that issues such as control over the flow rates, the mixing process, and the interaction time between sample and reagents have not yet been perfected. The CD-based format has the potential to become a standard tool for blood plasma separation and micro total analysis systems [12, 13]. However, this format has its own limitations, such as the need for numerous valves and the fact that a high-speed spinner and electrical power are required to manipulate the fluid. In addition to these two approaches, there are different microfluidic technologies, which are summarized in [14–16]. In the chip format, the separation methods can be divided into two main categories: passive and active. Active methods need an external force field such as an acoustic [17, 18], electric [19, 20], or magnetic field [21, 22] to achieve separation. In contrast, passive separation methods do not require external forces; therefore, they are more desirable for point-of-care (POC) devices. In this category, researchers exploited several techniques such as sedimentation [23, 24], dead-end filtration [25], cross-flow filtration [26, 27], deterministic lateral displacement [28, 29], pinchedflow fractionation [30], and biomimetic separation methods [31, 32]. Since 2005, attempts have focused on generating self-driven microfluidic designs with high manufacturing accuracy to provide more plasma output. For instance, Crowley et al reported a capillary-driven in-plane cross-flow plasma separation design from whole blood [33]. In 2009, Gervais et al highlighted the usefulness of a capillary-driven multiparametric microfluidic chip for one-step immunoassays [34]. In this microdevice, an embedded pad was used to filter the blood plasma. This pad required 20 μl of blood sample. In 2010, Kim et al [35] presented a surfactantadded polydimethylsiloxane (PDMS) design, (PDMS + Silwet L-77), but the volume of the extracted plasma was small, on the order of 20 nl. In 2011, Dimove et al [36] presented a stand-alone microfluidic chip that extracted plasma from a 5 μl droplet of blood for an enzyme-linked immunosorbent assay using sedimentation in a trench structure. Later, in 2012, Sakamoto et al proposed a capillary-driven design operating on the wetting properties of poly (methyl methacrylate) (PMMA) [37]. Three percent of the volume of plasma from 5 μl of blood was extracted, but the instability of the PMMA surface-modification process and red blood cells (RBCs) clogging the entrance of the microchannel array were challenges that still remain unmet in the use of this process for POC manufacturing. Recently, Liu et al proposed a pump-free blood plasma separator [38] based on gravitational sedimentation and filtration. However, this device needs a large sample volume (1.8 ml) and also requires further manipulation of the plasma after the separation process. The same year, 2013, Kang et al [39] developed a 2
cyclic olefin copolymer whole-blood plasma separator driven by asymmetric capillary forces. Although this chip separates approximately 102 nl from a single droplet of whole blood, it requires complicated post-processing to achieve filtration. Despite all these attempts, we still lack a simple device that addresses all the requirements for building a batch-manufacturable POC device, including the ability to separate plasma from untreated fresh blood, a long shelf life, and a reasonable separation time. In this work, we design and fabricate a novel selfdriven blood plasma separation microfluidic chip. To validate the quality and quantity of the separated plasma and to show its potential as a clinical tool, the microfluidic chip has been combined with lateral flow immunochromatography technology to perform a qualitative detection of thyroid-stimulating hormone (TSH). The clogging delay caused by RBCs in a hydrophilic PDMS constriction channel and the symmetric out-of-plane cross-flow filtration microchannel integrated micropillars (MIMPs) channel are used to maximize the amount of extracted plasma from the untreated blood. The proposed microdevice has several advantages over the previously mentioned devices for blood-plasma separation. This microdevice is selfdriven, which means that capillary forces are sufficient to deliver and separate more than 0.1 μl of blood plasma from a 5 μl droplet of real fresh blood (the required sample for blood testing in a POC device in terms of quantity and quality). Moreover, it does not require filters, tubes, or plasma manipulation, and due to its simple design, it can be easily mass produced. This paper is organized as follows. The principle of operation of the microfluidic device and the fabrication procedure are described in section 2. The proposed blood plasma separation microfluidic device is experimentally tested and validated in section 3, which also includes the experimental validation of the plasma quality with analyte detection.
2. Materials and methods 2.1. Principle of operation of the self-driven microfluidic device The proposed microfluidic design combines several aspects—including separation science, fluid dynamics, and blood rheology—to maximize the volume of extracted plasma. The phenomenon of RBCs clogging the entrance of the filtration channel is an inevitable and major issue for passive blood-plasma filtration. The proposed design addresses this issue by increasing the shear rate on the RBCs and using a symmetric arrangement of the capillary forces to extract the plasma from the whole blood. This technique moves RBCs away from the entrance of the filtration channel and consequently delays the RBC clogging, thereby increasing the extracted volume of blood plasma.
Biofabrication 7 (2015) 025007
H Madadi et al
Figure 1. Principle of operation of the self-driven microfluidic device. The yellow arrows show capillary forces on the extracted plasma in the MIMP filtration channel, while the red arrow shows the shear force on the RBCs due to the constriction design. The main channel includes a transport channel (300 μm × 10 μm) and two constrictions (100 μm × 10 μm), while two plasma-collected channels (100 μm × 10 μm) are placed on both sides of the main channel. (See video 1 in the supplementary information avaliable at stacks.iop.org/BF/7/025007/mmedia).
Figure 1 shows a schematic view of the microfluidic chip design, which is manufactured in two parts. The top part comprises the 10 μm-deep main channel (called the transport channel and used to deliver the blood drop sample) and two plasma-collected channels of the same depth. The bottom part of the microdevice is called the MIMP filtration channel and includes an arrangement of diamond microposts to decrease the flow resistance and increase the capillary force. The proposed MIMP filtration channel not only filters RBCs, but also paves the way for collecting the separated plasma in the top part of the microdevice using multiple collection channels. The MIMP design is the result of our previous work on developing a high-throughput microcapillary pump with efficient, integrated, low-aspect-ratio micropillars [40]. The obtained results indicate that the diamond-pillar array presents the minimum resistance flow compared to other pillar shapes. Furthermore, because of the deformability of human RBCs, the depth of the bottom part of the device, HFC , is designed to be less than 2 μm to successfully filter the RBCs (1.5 μm thickness). Two different transport-channel designs— straight and curved—have been investigated and compared, as seen in figure 2(a). In the straight transportchannel design, the blood flow in the transport channel is perpendicular to the extracted blood plasma flow, while in the curved design, both flows are parallel. The fluidic behavior of the whole device can be analyzed using an equivalent electrical network 3
(figure 2(b)). RT1 and ΔPT1 are the flow resistance and the pressure drop of the first part of the transport channel (before the constriction channel). RC and ΔPC are the flow resistance and the pressure drop of the constriction channel. RT 2 and ΔPT 2 are the flow resistance and the pressure drop of the second part of the transport channel (after the constriction channel). R F and ΔPF are the flow resistance and the pressure drop of the MIMP filtration channel, and R PC and ΔPPC are the flow resistance and the pressure drop of the plasma-collected channel, respectively. Once the hydrophilicity of the material is fixed (surface tension), the capillary pressure, ΔP, for each of the rectangular microchannels can be calculated using the Young–Laplace equation: ΔPi = σ cos θ
⎛1 1 ⎞⎟ ⎜ + ⎝h w ⎠
(1)
where σ is the surface tension, θ is the contact angle, h is the half depth of the microchannel, and w is the half width of the microchannel [41]. On the other hand, the flow rate, Q, of a liquid in a capillary microchannel depends on the viscosity of the liquid, the capillary pressure, and the microchannel flow resistance, as follows: Q=
1 ΔP μ R
(2)
where ΔP is the pressure difference between the area inside and in front of the liquid (equation (1)), μ is the viscosity of the liquid, and R is the microchannel flow resistance [41].
Biofabrication 7 (2015) 025007
H Madadi et al
Figure 2. (a) Transport microchannel design layouts. (b) The equivalent electrical network of the microdevice. (c) Diamond-shaped post-array in the filtration area.
On this basis, the minimization of the MIMP channel flow resistance, R F, will desirably increase the plasma flow rate in the MIMP filtration channel. To reach this goal, according to Madadi et al [40], a diamond-post shape with increased side distance between the pillars is the most efficient design, and it can be calculated through equation (3). RF =
12μL ⎛⎜ 1 ⎞⎟ wH 3 ⎝ 2ϵ − 1 ⎠
(3)
where μ is the fluid viscosity, L is the length of the microchannel, w is the half width of the microchannel, H is the height of microchannel, and ε is the porosity [40]. Since the fluid and the depth of the channel are fixed, we have maximized the porosity to achieve a manufacturable device. Figure 2(c) shows the dimensions of the selected MIMP filtration channel. 2.2. Fabrication procedure The microfluidic device is manufactured in two main parts—the top part and the bottom part, which are fabricated using different procedures. 2.2.1. Manufacturing process, top part SU-8, a negative-tone photoepoxy, is used to create a master mold via photolithography. PDMS, a common material in microfluidic chips, is used to manufacture the microdevice’s top part due to its outstanding characteristics—including its elastomeric properties, 4
biocompatibility, gas permeability, optical transparency, ease of bonding to glass, and low manufacturing cost [42, 43]—by means of single soft-lithography process. In spite of all the favorable physical and chemical properties, the hydrophobicity of PDMS is a handicap when pumping aqueous solutions through microchannels by capillary forces alone. To overcome PDMS’s hydrophobicity and to control PDMS’s hydrophobic recovery over a long period of time, a nonionic surfactant (Silwet L-77) is added to the PDMS before the curing process [44, 45]. Using the surfactant to modify PDMS’s hydrophobicity not only guarantees the shelf life of the final POC device for mass production (because of the negligible hydrophobic recovery [44]), it also eliminates the nonspecific absorption of proteins or biological material. Further details of the fabrication process are explained in the supplementary information avaliable at stacks.iop. org/BF/7/025007/mmedia. 2.2.2. Manufacturing process, bottom part The MIMP filtration channel is fabricated on a glass substrate using both photolithography and wet chemical methods. Figures 3(a)–(c) illustrate the fabrication steps of the MIMP filtration channel. Further details on the fabrication process are available in the supplementary information avaliable at stacks.iop.org/BF/7/ 025007/mmedia.
Biofabrication 7 (2015) 025007
H Madadi et al
Figure 3. (a)–(c) MIMP filtration channel manufacturing steps. The microdevice’s top and bottom parts (d) before and (e) after plasma bonding.
2.2.3. Microdevice assembly and test strip integration The PDMS microchannels (top part) and the etchedglass MIMP filtration channel (bottom part) are bonded via an oxygen plasma (Gambetti, Italy) treatment (see figures 3(d) and (e)). Oxygen at 200 Torr is introduced into the oxygen plasma chamber, and the plasma runs for 30 s. The oxygen plasma parameters are chosen to achieve maximum bonding strength [46]. Figure 4 shows a photograph of the completed blood plasma separation microfluidic device. To demonstrate and validate the potential diagnostic value of the microdevice and the quality of the separated plasma, the microfluidic chip was combined with a lateral flow immuno-chromatographic strip for TSH detection (NADAL® TSH Test, Germany). This test measures the levels of TSH, a hormone that is produced and released by the pituitary gland. The pituitary can sense whether there is enough thyroid hormone in the bloodstream, and it releases TSH when it detects insufficient thyroid hormone. The official normal range for TSH in blood runs from approximately 0.5 to 4.5/5.0 μIU ml−1. Therefore, a TSH level under 0.5 μIU ml−1 indicates hyperthyroidism (an overactive thyroid), and a TSH over 5.0 μIU ml−1 indicates hypothyroidism (an underactive thyroid). The separated plasma is driven to the entrance of half of a TSH test strip, as shown in figure 5. The filtration pad (fiberglass) and the absorption pad (at the end of test strip) are removed from the test strip. The test strip gold-conjugate pad is overlapped with the blood-plasma reservoir and sealed with tape. In the gold-conjugate pad, the plasma antigens are recognized and bound to the gold-conjugated 5
antibodies, as shown in figure 5(a). When the gold nanoparticle complex reaches the test line, a strong bond is formed with the immobilized antibodies embedded in the nitrocellulose membrane, and a colored line appears. Since the required volume of the blood plasma for wetting the conjugate pad is relatively high, 10 μl of phosphate-buffered saline (purchased from Life Technologies S A, Madrid, Spain) are added to the lateral-flow assay (LFA) strip after the plasma has reached the conjugate pad.
2.2.4. Sample preparation Fresh human whole blood collected from a healthy volunteer using the finger-prick method is used for experiments. The sample is collected using the lancet of a POC device in every experiment. To ensure the accuracy of the input volume, the blood sample is pipetted. The micropipette is gently held in a horizontal or slightly downward position near the channel inlet to introduce the blood sample. In this way, the blood sample loading does not affect in the blood flow behavior inside the device. For the TSH blood test, two different types of blood samples are prepared. The first type is a 5 μl drop of healthy fresh human whole blood, and the second type is a 5 μl drop of blood with a high concentration of TSH, which is obtained from reconstituted blood by mixing high-TSH dried plasma (provided by GRIFOLS) with RBCs from regular blood. Then, the collected blood samples are mixed with 0.5 μl heparin (purchased from Pfizer Co., Madrid, Spain), a standard anticoagulant.
Biofabrication 7 (2015) 025007
H Madadi et al
Figure 4. Photograph of (I) PDMS part (top part), (II) glass part (bottom part), and (III) the final microdevice.
Figure 5. (a) Schematic view of the microfluidic chip, including LFA strip. (b) The steps for embedding the TSH test strip inside the microfluidic device.
6
Biofabrication 7 (2015) 025007
H Madadi et al
Table 1. Comparison between the straight and curved transport channel geometries. Microchip
Blood sample volume
Straight transport channel Curved transport channel
∼5 μl ∼5 μl
Average volume of extracted plasma
Average time (min)
0.09 μl ± 0.01 0.07 μl ± 0.01
4±1 6±1
2.2.5. Experimental setup To record images and movies of the blood plasma separation process and analyze the results, a digital camera (Tucsen ISH500, 5.0 M pixel,) is connected to a microinspection lens system (Optem zoom 125C with a broad 12.5:1 zoom range and a 20× objective). All quantitative data are analyzed and expressed as the mean ±the standard deviation.
3. Results and discussion To study the effect of the transport channel geometry on the blood-plasma separation, two different channel geometries are considered: a straight transport channel and a curved transport channel. The aim of the curved design is to maximize the first-instance filling rate of the MIMP channel before the RBCs clog the MIMP entrance due to the fact that the blood and extracted plasma flow in the same direction. The experiments are performed three times for each channel geometry. The volume of the extracted plasma is calculated by adding the volume of the MIMP channel, the plasma-collected channels, and the collected plasma in the outlet reservoir. The results are summarized in table 1, which shows the average of the recorded data. (See the supplementary information avaliable at stacks.iop.org/BF/7/025007/mmedia for the whole data). 3.1. Curved transport channel A 5 μl drop of fresh human whole blood is introduced into the channel inlet of the transport channel. The blood sample is drawn into the microdevice by capillary forces alone, due to the hydrophilicity of the surfactant-added PDMS (Silwet L-77). The separation process starts as soon as the blood flows into the transport channel and arrives at the MIMP filtration channel. Since the MIMP filtration channel depth is less than 1.5 μm, RBCs cannot penetrate into the plasma channel. In the separation zone, the RBCs’ velocity in the transport channel is slowed down, which leads to an accumulation of RBCs in the entrance of the separation area. The purpose of the constriction in the middle of the transport channel is to generate a local flow acceleration and delay the RBCs clogging of the filtration area. In the curved transport-channel design, although the constriction postpones the RBC-related clogging, the centrifugal force due to the curvature and the nonsymmetrical arrangement of the separation zone leads to precocious RBC clogging in the entrance of the 7
Purity % >98 80–85
MIMP filtration channel. Figure 6 illustrates the filtration performance for the plasma separation from whole blood in the curved transport channel. Using a curved transport channel, an average of 0.07 μl of blood plasma are extracted from a 5 μl droplet of fresh whole blood. In the separation process, the MIMP filtration channel is filled within 1 min due to the fact that the plasma flow direction is parallel to the blood flow in the transport channel (see figure 6(b)), while the plasma-collected channel is filled in 5 to 7 min (see figures 6(g) and (h)). Figures 6(b) and (c) show that the plasma-collected channel is initially filled from the top surface of the pillars in the collected channel due to the higher capillary force compared to the other zones. It is noticeable that although some air bubbles form inside the separation area and the plasma-collected channel during the first few minutes, they disappear in the following steps of the process due to the permeability of the PDMS. Obviously, these trapped air bubbles reduce the separation efficiency because they create a resistance to the capillary flow, but the formation of air bubbles can be controlled by unifying the wetting properties of the modified PDMS and the glass [47]. The purity of the extracted blood plasma is defined as: Purity = 1 −
CP CM
(4)
where C M is the number of RBCs in the main channel and C P is the number of RBCs in the plasma-collected channel [8]. Image processing is used to identify the purity of the extracted blood plasma from whole blood, which is around 80 to 85% for the curved transport-channel topology, as seen in figure 7(a). The measurement is done based on 90 images taken from three different regions of the filled plasma-collected channel using the described experimental setup. ImageJ software is used to count the number of RBCs in the same specified area (the considered volume is 5 × 105 μm3) of every image region. The average of the obtained results (the number of RBCs in the plasma-collected channel) compared to the hematocrit concentration of the initial sample (fresh whole blood) is 45% in our experiments. The curved transport channel fills the separation MIMP channel in less than 1 min. However, the centrifugal force on the RBCs attributed to this geometry and the nonsymmetrical capillary forces reinforce the RBCs’ accumulation in the entrance of the MIMP filtration channel and drive them into the MIMP
Biofabrication 7 (2015) 025007
H Madadi et al
Figure 6. Blood plasma separation process using the curved transport channel geometry. (a) The hollow transport channel (blood channel) and the plasma-collected channel before the introduction of the whole fresh blood in the microchip inlet. (b) The blood channel filled with fresh blood and the filled MIMP filtration channel with the separated blood plasma. (c) The starting point of the blood plasma filling process in the plasma-collecting channel. (d)–(f) The blood plasma filling process in the plasma-collected channel. (g) and (h) The plasma-collected channel completely filled with the extracted blood plasma after 5 to 6 min (see video 2 in the supplementary information avaliable at stacks.iop.org/BF/7/025007/mmedia).
Figure 7. (a) A comparison between the concentration of RBCs in the main blood channel and their concentration in the plasmacollected channel. (b) The effect of the curved geometry and the resulting centrifugal force on the RBCs (see video 2, supplementary information avaliable at stacks.iop.org/BF/7/025007/mmedia).
channel, and thus into the plasma-collected channel (see figure 7(b)). A video clip of the performance of the blood plasma separation microdevice using the curved channel is provided in video 2 in the supplemental information avaliable at stacks.iop.org/BF/7/ 025007/mmedia. 3.2. Straight transport channel To increase the separation efficiency and the purity of the extracted plasma and to avoid the effect of the 8
centrifugal force due to the curvature design, the straight transport channel is tested. An example of the microdevice’s performance can be seen in figure 8. The mechanism of separation is similar to the curved transport channel, but there are two main differences. First, the direction of the blood flow in the transport channel is perpendicular to the extracted plasma flow in the MIMP channel. Second, the capillary forces due to the plasma extraction are applied symmetrically on both sides of the RBCs that flow in the transport
Biofabrication 7 (2015) 025007
H Madadi et al
Figure 8. Blood plasma separation process using the straight main channel geometry. (a) The hollow transport channel (blood channel) and the plasma-collected channel before the introduction of the whole fresh blood droplet in the microchip inlet. (b) The blood channel filled with fresh blood and the MIMP filtration channel filled with the separated blood plasma. (c) The starting point of the blood plasma filling process in the plasma-collected channel. (d)–(f) The blood plasma separation and filling process in the plasma-collected channel. (g) and (h) The plasma-collected channel completely filled with the extracted blood plasma after 3 min. (see video 3 in the supplementary information avaliable at stacks.iop.org/BF/7/025007/mmedia).
channel. This symmetric plasma extraction does not disturb the RBCs’ trajectory in the transport channel, which leads to a delay in the RBC-related clogging of the filtration area. In the case of the straight transport channel, an average of 0.09 μl of blood plasma is successfully separated from a 5 μl droplet of whole fresh blood. According to the results, the plasmacollected channel is filled in 3 to 5 min with a purity of more than 98%. The obtained results confirm that the perpendicularity of the plasma flow decreases the amount of time required to fill the plasma-collected channel [48]. A video clip of the performance of the blood plasma separation microdevice using the straight channel is provided in video 3 in the supplementary information. Table 1 shows a comparison of the results between the straight and curved transport channel geometries. To investigate the effect of the passive capillary filling flow rate on the microdevice’s performance, three types of microdevices with different surfactant concentrations (Silwet L-77 concentration 1%, 2%, and 3% in PDMS) are fabricated. More details about the experiments and the obtained results are explained in the supplementary information. According to the obtained results, the whole 9
microdevice performance time is decreased by more than 25% to less than 3 min by increasing the surfactant (Silwet L-77) concentration from 1% to 3%. This reduction in filtration time addresses the bottleneck in the sample preparation procedure for some blood analyses in order to achieve the appropriate performance of the POC device. 3.3. Double-embedded plasma-collected channel To further investigate the ability of the proposed microdevice to perform different assays simultaneously on the same blood sample, another plasmacollected channel is added to the other side of the main blood-flow channel (1 mm apart). The filling of this second plasma-collected channel not only opens the possibility of performing multifunctional blood analysis, but also increases the volume of extracted plasma up to 0.1 μl from a 5 μl droplet of whole fresh blood with the same purity (more than 98%). Since the presented blood-plasma separation device is driven by capillary forces alone, no hemolysis is observed. More details about the hemolysis study can be found in the supplementary information. Moreover, a video clip of the performance of the double-embedded plasma-
Biofabrication 7 (2015) 025007
H Madadi et al
Figure 9. The results of the TSH test. (a) Negative TSH result using a hybrid microfluidic chip. (b) Positive TSH result using a hybrid microfluidic chip. (c) Negative TSH result using an LFA strip.
collected channel is provided in video 4 in the supplementary information avaliable at stacks.iop. org/BF/7/025007/mmedia. 3.4. Integration of plasma separation with analyte detection for TSH blood test To demonstrate the utility of the quantity and quality of the extracted plasma, the plasma from the collected channel is driven to the entrance of an LFA to qualitatively detect the TSH level. Two different types of tests are performed based on the 5 μl prepared samples of healthy fresh blood and reconstituted blood. Figure 9 shows an example of the results of the TSH tests using both the test-strip device and the hybrid microfluidic chip (the implemented test strip in the presented microfluidic device). Note that in figure 9, the hybrid microfluidic chip is shown from the back view due to the rotation of the test strip in the implementation process (see figure 5(b)). Three test devices are tested for every sample type, and the colored line appeared in the control section of the test strip 18 ± 2 min after adding the blood droplet. Figure 9 shows a picture of the obtained results from the hybrid microfluidic chip compared to the standard NADAL® TSH test. The TSH test result is negative and the colored line that appears in the control ‘C’ section confirms the correct operation of the TSH test in the presented hybrid microfluidic chip. Figure 9(b) also shows the results when high-TSH blood is processed. The results are positive, since a result line appears in addition to the control line. According to the results, the proposed design shows its potential as a clinical tool compared to previous works. First, the initial volume of the required sample, 5 μl, is considerably reduced compared to other works [34, 38]. Second, the volume of the extracted blood plasma, 0.1 μl, is enough to implement the TSH test and also is more than the obtained volume in previous filter designs [33, 35]. 10
Importantly, the use of a simple design, which is made of only two parts without post-processing (such as in [37, 39]) or embedding additional parts for filtration [34], makes it suitable for batch manufacturing. Furthermore, the capability of the proposed microdevice for separating and gathering blood plasma within more than one plasma-collected channel paves the way for blood panels in a single POC test.
4. Conclusion In this work, a novel, self-driven blood plasma separation microfluidic chip is designed, built, and validated through integration with a TSH qualitative test. The test works with a single droplet of whole blood (∼5 μl), and 0.1 μl of blood plasma is successfully separated and collected in 10 μm-deep channels in 3 to 5 min, with a high purity level of more than 98%. The long-term hydrophilicity of the employed material, the straightforward manufacturing process, and the achieved plasma volume (0.1 μl) pave the way for doing multiplex blood analyses simultaneously in the same device, as is promised in portable medical POC testing. Supporting information Additional information as noted in text. Ethics statement The whole human blood sample for the experiments was collected from a healthy volunteer using the finger-prick method. The protocol was approved by the committee of the ethics of the Technical University of Catalonia, Barcelona-Tech.
References [1] [2] [3] [4] [5]
Whitesides G 2014 Lab Chip 14 3125–6 Psychogios N et al 2011 PLoS One 6 e16957 Wagner J 2012 Biochemia Medica 22 24–38 Polanski M and Anderson N L 2007 Biomarker Insights 1 1−48 Franklin B S, Vitorino B L F, Coelho H C, Menezesneto A, Santos M L S, Campos F M F, Brito C F, Fontes C J, Lacerda M V and Carvalho L H 2011 PLoS One 6 e19842
Biofabrication 7 (2015) 025007
H Madadi et al
[6] Rainer T H, Wong L K S, Lam W and Yuen E 2003 Clin. Chem. 49 562−9 [7] Schwarzenbach H, Hoon D S B and Pantel K 2011 Nat. Rev. Cancer 11 426−37 [8] Kersaudy-Kerhoas M and Sollier E 2013 Lab Chip 13 3323−46 [9] Vella S J, Beattie P, Cademartiri R, Laromaine A, Martinez A W, Phillips S T, Mirica K A and Whitesides G M 2012 Anal. Chem. 84 2883–91 [10] Yang X, Forouzan O, Brown T P and Shevkoplyas S S 2012 Lab Chip 12 274−80 [11] Songjaroen T, Dungchai W, Chailapakul O, Henryand C S and Laiwattanapaisal W 2012 Lab Chip 12 3392−8 [12] Gorkin R, Park J, Siegrist J, Amasia M, Lee B S, Park J M, Kim J, Kim H, Madou M and Cho Y K 2010 Lab Chip 10 1758−73 [13] Amasia M and Madou M 2010 Bioanalysis 2 1701−10 [14] Sollier E, Rostaing H, Pouteau P, Fouillet Y and Achard J 2009 Sensors Actuators B 141 617−24 [15] Pamme N 2007 Lab Chip 7 1644–59 [16] Mukherjee S, Kang T G, Chen Y and Kim S 2009 Crit. Rev. Biomed. Eng. 37 517−29 [17] Petersson F, Aberg L, Sward-Nilson A M and Laurell T 2007 Anal. Chem. 79 5117−23 [18] Doria A, Patel M and Lee A P, μ TAS 2011 15th Int. Conf. on Miniaturized Systems for Chemistry and Life Sciences (Seattle, WA, 2–6 October 2011) [19] Kralj J G, Lis M T, Schmidt M A and Jensen K F 2006 Anal. Chem. 78 5019−25 [20] Liao S H, Chang C Y and Chang H C 2013 Biomicrofluidics 7 024110 [21] Furlani E 2007 J. Phys. D: Appl. Phys. 40 1313−19 [22] Han K H and Frazier A B 2006 Lab Chip 6 265−73 [23] Tachi T, Kaji N, Tokeshi M and Baba Y 2009 Anal. Chem. 81 3194−8 [24] Zhang X B, Wu Z Q, Wang K, Zhu J, Xu J J, Xia X H and Chen H Y 2012 Anal. Chem. 84 3780−6 [25] Li C, Liu C, Xu Z and Li J 2012 J. Biomed. Microdevices 14 565−72 [26] Aota A, Takashi S, Mawatari K, Tanaka Y, Sugii Y and Kitamori T 2011 Anal. Sci. 27 1173−8 [27] Sollier E, Rostaing H, Pouteau P, Fouillet Y and Achard J L 2009 Sensors Actuators B 141 617−24
11
[28] Huang L R, Cox E C, Austin R H and Sturm J C 2004 Science 304 987−90 [29] Inglis D W, Morton K J, Davis J A, Zieziulewicz T J, Lawrence D A, Austin R H and Sturm J C 2008 Lab Chip 8 925−31 [30] Rodriguez-Villarreal A I, Arundell M, Carmona M and Samitier J 2010 Lab Chip 10 211−19 [31] Yang S, Undar A and Zahn J D 2006 Lab Chip 6 871–80 [32] Kersaudy-Kerhoas M, Kavanagh D M, Dhariwal R S, Campbell C J and Desmulliez M P Y 2010 Lab Chip 10 1587–95 [33] Crowley T A and Pizziconi V 2005 Lab Chip 5 922–9 [34] Gervais L and Delamarche E 2009 Lab Chip 9 3330−7 [35] Kim Y C, Kim S H, Kim D, Park S J and Park J K 2010 Sensors Actuators B 145 861−8 [36] Dimov I K, Basabe-Desmonts L, Garcia-Cordero J L, Ross B M, Ricco A J and Lee L P 2011 Lab Chip 11 845–50 [37] Sakamoto H, Hatsuda R, Miyamura K and Sugiyama S 2012 Micro Nano Lett. 7 64−7 [38] Liu C, Mauk M, Gross R, Bushman F D, Edelstein P H, Collman R G and Bau H H 2013 Anal. Chem. 85 10463−70 [39] Lee K K and Ahn C H 2013 Lab Chip 13 3261−7 [40] Madadi H, Casals-Terre J, Lopez R C and Anfrez M S 2014 Microfluid Nanofluidics 17 115−30 [41] Bruus H 2007 Theoretical Microfluidics (Oxford: Oxford University Press) [42] Mata A, Fleischman A J and Roy S 2005 Biomed. Microdevices 7 281–93 [43] Van Poll M L, Zhou F, Ramstedt M, Hu L and Huck W T S 2007 Angew. Chem. Int. Ed. Engl. 46 6634–7 [44] Madadi H and Casals-Terre J 2013 J. Microsyst. Technol. 19 143–50 [45] Kim H T, Kim J K and Jeong O C 2011 Japan. J. Appl. Phys. 50 06GL04 [46] Bhattacharya S, Datta A, Berg J M and Gangopadhyay S J 2005 Microelectromech. Syst. 14 590−7 [47] Le N B and Hsu Y C 2010 Adv. Nat. Sci.: Nanosci. Nanotechnol. 1 015006 [48] Kim D, Yun J Y, Park S J and Lee S S 2009 Microsyst. Technol. 15 227–33
