A low cost and quasi-commercial polymer film chip for high-throughput inertial cell isolation

Abstract

We propose a novel scheme for fast fabrication (<20 minutes) of ultra-low-cost (∼1.5 cents) polymer film chips using laser direct writing and roll-to-roll lamination. The microchannel architectures of the chip are defined by cutting through grooves within a polyvinyl chloride (PVC) hard film using a UV laser system, and then the patterned film is sealed by two sheets of laminating films through a laminator. To obtain the optimal process parameters for chip fabrication, we systematically analyse the influences of the critical factors of laser ablation and thermal lamination on channel dimensions. As an example application of our polymer chip, we develop an integrated device with high bond strength for the high-throughput inertial isolation of cancer cells from human blood. The experimental results show that our device can successfully isolate high purity cancer cells, proving the potential availability of our polymer film chip for biomedical applications. The advantages of low-cost and fast fabrication make our polymer film chip suitable for both laboratory scale development and industrial batch production.

---

Introduction

As an enabling technology for the miniaturisation of diagnostic instruments, microfluidics or lab-on-a-chip systems have made significant contributions to the field of life science during the last decades, and have offered a wide range of biomedical applications including DNA analysis,¹ polymerase chain reaction (PCR),² immunoassays,³ cell isolation⁴ and protein analysis.⁵ However, due to the constraint of materials, cost and fabrication process, it is still not easy for the microfluidic chips invented in laboratories to be successfully transferred into the mature commercial products in our daily life. For the industrial mass production of an ideal microfluidic product, characteristics of low-cost, disposable and amenable will be the most important considerations. Thus, researchers and manufacturers not only need to choose the appropriate materials which are compatible for the subsequent analytical and biomedical applications, but also have to develop reliable manufacturing processes for the low-cost and rapid fabrication of microfluidic products in an industry mode. For now, soft lithography⁶ remains the most popular scheme for the rapid prototyping of microfluidic devices in polydimethylsiloxane (PDMS). The unique advantages of this technique mainly reflect on the relatively mature fabrication procedures in laboratories and the large scale integration of various functional units within a chip.⁷ However, the elastomeric property of PDMS, the requirement of yellow clean-room environment, and the instability of fabrication process strongly limit the applicability of PDMS microfluidic chips for the industrial batch production.⁸

With the increasing propagation of thermoplastic polymers^9–11 for microfluidic applications in recent years, polymer chips had already exhibited their unique advantages in stable physical–chemical properties, disposable single use and inexpensive fabrication cost. In particular, the new arrival replication based technologies, such as hot embossing¹² and injection molding¹³ provide new pathways to the fabrication of microscale architectures within the polymer substrates at low-cost and in high-volume productions. The ability of repeatedly replicating the microstructures from the maser molds to the target workpiece enables the commercialization of polymer chips to be possible. Despite a high production throughput can be offered, both of these two techniques face the same problem of requiring precise and rigid molds which are usually very expensive and time-consuming especially when the optimization or redesign of architectures are needed. Thus, the above two methods are more suitable for the continuous and stable mass production of mature microfluidic products.

There are many other rapid prototyping methods for the fabrication of polymer chips, including computer numerical control (CNC) micromachining,¹⁴ laser ablation,¹⁵ xurography¹⁶ and 3D printing.¹⁷ These methods can directly machine architectures without using master molds and high-resolution photomasks, which are proved to be very cost-effective. The primary disadvantages of these methods are the poorer dimension tolerances and lower production throughout as compared with the above-mentioned replication based techniques. Due to the reasonable turnaround time, these methods are mostly acted as an effective fabrication scheme for the concept design validations or device optimizations before a final batch production.

In this work, we propose a novel method for the low-cost and fast fabrication of polymer film chips via UV laser direct writing and roll-to-roll thermal lamination methods. For concept demonstration, polyvinyl chloride (PVC) hard film is chosen to be the material of channel layers, and polyethylene terephthalate film with thermal sensitive ethylene vinyl acetate co-polymer (PET/EVA) is selected as the lamination materials. To fabricate the polymer film chip, a UV laser system is used to cut the through microchannel architectures within the PVC film, and then the patterned film is sealed by two sheets of laminating films through a laminator.

To explore the optimal process parameters for microfluidic chip fabrication, we systematically study the effects of laser parameters on dimensions of the fabricated ablation grooves, and also investigate the evolution of the channel dimensions as functions of critical factors in the lamination process. To probe into the practical application of the fabricated polymer film chip, an integrated device consisted of a spiral microchannel and a set of housing apparatus is built to achieve the high-throughput inertial isolation of cancer cells from human blood.

Fabrication of polymer film chip

The laser direct writing and roll-to-roll lamination methods

For laser ablation in microfluidic applications, CO₂,^15,18 UV^19,20 and femtosecond lasers^21,22 are the most widely used approaches. Generally, due to the high heat concentration on the workpiece, CO₂ laser yields a minimum resolution of ∼100 μm, limiting the application areas to the analytical devices²³ whose structure resolutions are not highly required. In contrast, the short pulsed UV and femtosecond lasers are found to be more suitable for the precise machining of polymer materials. Since the extremely low heat effects generated in the laser processing, UV laser can machine structures with resolutions less than 50 μm, whereas the ultra-high resolution of several microns and even complex 3D architectures can be achieved via a femtosecond laser. However, due to the high investment cost and the tough maintenance constraints, the femtosecond laser systems are more likely to be employed as a laboratory instruments rather than equipments for the industrial large-volume production. To consider the balance between the machining resolution and the investment cost, we determined to choose the UV laser system (AWAVE 355-15W-30K, Advanced Optowave Corporation, USA) as the machining tool for patterning the polymer film. The maximum laser power of this system is 15 W, and the laser beam wavelength is 355 nm.

Following the laser machining of polymer film, a bonding process needs to be carried out to seal the patterned microchannels, and thus is another important step to achieve the final polymer film chip. This step would not only affect the total fabrication time and cost, but also determine the chemical or solvent compatibility of the fabricated chip. In addition, the optical properties, bonding strength, and dimensional homogeneity of the finally sealed chips are also needed to be considered carefully. Generally, there are many rapid methods (e.g., adhesive bonding,²⁴ solvent bonding,²⁵ localized welding²⁶ and surface modification²⁷) reported in previous works for bonding the polymer chips. Given the consideration of cost and bonding reliability, we proposed here a simple bonding method based on roll-to-roll thermal lamination technique for sealing the laser cut film chips. In this work, the bonding process was accomplished using a commercial grade laminator (LM8-330, RAYSON), which has 8 rollers and nine-level rolling speeds from 1.05 m min⁻¹ to 2.1 m min⁻¹, and the roller temperature can be regulated continuously from the room temperature to a maximum of 170 °C.

Fabrication process

The fabrication process of polymer film chip is illustrated in Fig. 1, and can be summarized as the following four steps.

Fig. 1 Schematic illustration of the fabrication process of the polymer film chip. (a) UV laser direct writing on a sheet of PVC film to cut through the channel circuit pattern; (b) chip lamination by using PET/EVA laminating films to seal and bond the ablation PVC film; (c) two kinds of chip-to-world connexions.

a. The design of microchannels. The desired structures of microchannels in the chip can be created via any commercial computer-aided design (CAD) softwares. In this work, we used AutoCAD software (AutoCAD 2013, AutoDesk, Inc.) to create the electronic drawing of the microchannels. The electronic file was saved to a USB mass storage device and exported to the UV laser system which automatically transfers the CAD file to be a readable format of the system software.

b. The fabrication of channels. This process step presents the laser direct writing method to cut a through groove within a PVC film. First, a sheet of PVC film was rinsed in isopropanol, cleaned in deionized water and then dried via the nitrogen gas stream. After that, the film was fixed onto the platform of the UV laser system, and the distance between the laser lens and the film was adjusted to obtain the optimum processing effect. Then, the laser beam was emitted to cut the contour of the channel circuit patterns. Finally, the film was cut through and the centre portion of the microchannel circuits was carefully removed with a tweezer.

c. The bond of PVC film. This process step presents the bonding process to seal the PVC film. First, the ablation film was cleaned with isopropanol and deionized water to remove the dirt and residues inside the channels and then dried in nitrogen gas. After that, the PVC film was sandwiched between two sheets of PET/EVA laminating films, and sealed using the laminator under the determined roller temperature and rolling speed. In the end, the laminated polymer chip was cut into the desired size.

d. The connexion of chip-to-world. The polymer chip has two types of chip-to-world connexions. The first approach is to use the commercial microfluidic adapters (NanoPort, Upchurch Scientific) to connect and seal the inlet and outlet of the chip. This connexion method is very simple and fast, which can be used in the laboratory. The second approach is to fabricate a set of rigid injection-molded housings to clamp the ship. As compared with the first method, the second one makes the integrated device more durable and portable, and thus is more suitable for using in the industrial productions.

Examples of the prototype polymer film chip and the images illustrating the channel cross-sections of different aspect ratios are shown in Fig. 2a–g. As can be observed from the figures, the edges of the microchannels cut by the UV laser system are relatively smooth and are approximately vertical to the upper and lower laminating films.

Fig. 2 (a) Polymer film chip with a spiral microchannel; (b) microscope images representing the cross-sections of the sealed spiral microchannel, and channels with aspect ratios of (c) 1 : 2 and (d) 11 : 1; microscope images representing the inlet (e) and the outlet (f) of the spiral channel; (g) SEM image of the channel side wall, scale bars 100 μm.

Bond strength test

In order to characterize the bond strength of the polymer film chip, we constructed an experimental setup for the strength test, as illustrated in Fig. S1.† In this experimental setup, deionized water is driven by the condensed nitrogen gas to flow through a microfluidic device, and then pushed out to flow into a waste tank. During the test process, the input gas pressure is monitored by a pressure gage, and is regulated by a manual pressure regulating valve. The measured maximum non-leak gas pressure was recorded as the bond strength of the tested microfluidic device.

The measured bond strength of the polymer chip with a spiral channel is 45 kPa, which is equal to be about 3.7 N according to the channel area. This strength is not strong enough for the high-pressure-required microfluidic applications, thus the bond strength of the chip should be improved. From the strength test experiment, we found that the increased fluidic pressure breaks the bond interfaces between the laminating films and the PVC film, which directly results the deformations of laminating films. In order to prevent deformations of the laminating films, we used a set of rigid housings to clamp the chip, and then measured the bond strength again. The experimental results indicate that the non-leak gas pressure of the chip is increased to be over 8 bars, which is approximately 18 times higher than before. The improved bond strength enables our polymer film chip to be used in many microfluidic applications of high fluidic pressure requirement (high-throughput inertial isolation,^28,29 and viscoelastic focusing^30,31). For the requirement of higher bond strength in microfluidic chips, the previously reported tubing method may be applied.³²

Influences of critical process parameters on channel dimensions

Dependence of ablation groove dimensions on laser parameters

In the process of laser ablation, several variables (e.g., the material of workpiece, the laser power setting, the laser writing speed, the laser repetition number, and the distance between the lens of laser system and the workpiece) need to be considered. Among these variables, the parameters of laser power, speed and repetition number are the most important factors which directly affect the dimensions of the ablation grooves. Therefore, we will carry out a systematic investigation on the dependence of ablation groove dimensions on these factors in the following section.

To probe into the relationship between laser parameters and groove dimensions, we first used UV laser to machine several grooves within the PVC film under varied laser powers, writing speeds and passes. Then, we cut the film along a line vertical to the grooves using a razor blade and captured the cross-sectional images through an optical microscope. After that, we measured the groove depths and widths, and concluded the correlations between the laser parameters and the groove dimensions. Fig. 3a shows the dependence of the groove dimensions on the laser power at a given writing speed of 400 mm s⁻¹ and writing number of 20 passes. In order to eliminate the heat concentration caused by the continuous laser cutting, we set the laser system to emit laser beam every 500 ms. As can be observed from the figure, the groove depth increases slightly when the laser power is lower than 40%, and then the depth value increases dramatically. Instead, the groove width increases continuously until the power reaches 60%, and finally the width value saturates to be constant at 50 ± 5 μm. The captured microscope images in Fig. 3b show the cross-sectional profiles of the laser-cut films under the conditions of four different laser powers (26.7%, 46.7%, 60% and 66.7% in percent of 15 W). It can be seen that the film was totally cut through when the laser power was set to be 66.7%, and such a high energy finally results in the small bumps of molten material residues on the upper and lower edges of the through groove.

Fig. 3 Correlation of ablation groove dimensions to laser cutting parameters. (a) Dependence of the groove depth and width on the laser power, which is given in percent of 15 W; (b) microscope images representing the cross-sectional profiles of the laser-cut films under the conditions of different laser powers; (c) dependence of the groove depth and width on the laser writing speed; (d) dependence of the groove depth and width on the laser writing pass. Inset shows the relatively smooth walls of a through groove cut under a laser writing pass of 70. Scale bars 50 μm.

Fig. 3c shows the dependence of the groove dimensions on the laser writing speeds. The laser power here was fixed at 53.5%, and the number of laser writing passes was set to be 20, whereas the laser writing speed was changed from 100 mm s⁻¹ to 1000 mm s⁻¹. From the Fig. 3c, we observed that the PVC film is directly cut through when the laser writing speed is lower than 200 mm s⁻¹. With the increase of writing speed, the groove depth decreases linearly, whereas the groove width stays constant at first and then decreases gradually. We also found that the edges of the grooves are over burnt and deform seriously when the grooves are cut too slowly. In addition, at lower cutting speeds, the surface of the PVC film suffers from higher heat concentration than that at faster speeds, and more molten residues are ejected and accumulated around the grooves.

After setting the laser power and writing speed to the reasonable values (53.5% in percent of 15 W, 400 mm s⁻¹), a continuously increment was found on the ablation groove depth versus the number of the laser writing passes, as illustrated in Fig. 3d. The experimental results indicate that the groove width increases continuously during the first 40 passes and then keeps constant after that. In addition, we found that the PVC film is totally cut through when the laser cuts the groove for about 70 passes, and the groove width was measured to be about 50 μm. We also checked the top and cross-section of the groove, and found that the cutting section of the groove is relatively smooth (see the inset of Fig. 3d). Although a laser writing pass of 70 was employed, the machining of a through channel in the PVC film of 160 μm thick could be finished in 80 seconds.

Influence of lamination process on channel dimensions

The sealing performance of the polymer film chip is directly affected by the lamination process parameters (e.g., thickness of laminating film, roller temperature and rolling speed). Therefore, it is necessary to understand the influence of these parameters on channel dimensions. To probe into the relationship between the laminating film thickness and the channel dimension, several through channels of different aspect ratios (width : depth = 1 : 1, 6 : 1, and 11 : 1) were cut and laminated by laminating films of different thicknesses (i.e., 75 μm, 85 μm, 100 μm, 125 μm, and 150 μm). The roller temperature and the rolling speed were set to be 120 °C and 1.6 m min⁻¹, respectively according to the laminator instruction provided by its manufacturer. Before lamination, the laminator was preheated for at least 5 minutes to achieve a stable operational temperature. Fig. 4a shows the deformation values of the laminating films as a function of the film thicknesses. As can be found from this figure, the channels laminated with 100 μm thick films obtain the smallest deformation among the five film thicknesses. It is also found that the channels with an aspect ratio of 1 : 1 achieve a smaller deformation as compared with that of the high aspect ratio channels. In order to understand the deformation characteristics of the films, we plotted the cross-sectional profiles of the channels with an aspect ratio of 11 : 1, as shown in the right of Fig. 4a (the cross-sectional profiles of the channels with aspect ratios of 1 : 1 and 6 : 1 are provided in Fig. S2†). The results indicated that both the upper and lower EVA adhesions of the 75 μm and 150 μm thick films adhere to each other (see the inset images of Fig. 4a), resulting in the clogging of channels. As a comparison, the channel laminated with 100 μm thick film shows relatively smooth and flat surface (see Fig. 2d). The reason for the collapse is that the stiffness of the 75 μm thick film is not strong enough for the effective sealing of high aspect ratio channels due to the relatively thin thickness. Instead, the 150 μm thick film has the strongest stiffness among the three tested films, but it would suffer from the maximum pressure load from the rollers due to the over thick thickness, which results in its collapse. Therefore, the 100 μm thick film is the best choice for the high quality lamination in this work.

Fig. 4 Influence of the lamination process on the channel dimension. (a) Laminating film deformation values versus film thickness, and comparison of channel cross-sectional features with different laminating films; (b) laminating film deformation value versus roller temperature, and comparison of channel cross-sectional features under different roller temperatures; (c) laminating film deformation value versus rolling speed, and comparison of channel cross-sectional features under different rolling speeds. The bold lines inside the microchannel cross-sections in the right figures denote the profiles of EVA adhesive layers of the laminating films, and the thin lines in the outside represent the profiles of PET layers.

To investigate the influence of roller temperature on channel dimension, we laminated the PVC films under varied temperatures from 90 °C to 170 °C. The rolling speed was set to be 1.6 m min⁻¹, and the 100 μm thick laminating film was chosen due to its good lamination performance. Fig. 4b shows the deformation values of the laminating films versus the roller temperatures. From this figure, we observed that there is no deformation occurred in the channels with aspect ratios of 1 : 1 and 6 : 1 as the temperature increases from 110 °C to 130 °C while a stable deformation was observed in the channel with an aspect ratio of 11 : 1. However, with the increase of temperature (higher than 130 °C), all of the three channels start to deform, and the channel of high aspect ratio shows a bigger deformation than the low aspect ratio channel. In addition, the comparison of cross-sectional features in the right figure indicates that when the temperature is lower than 100 °C, the EVA adhesive does not melt completely and the laminating film exhibits much poorer optical property than the film at temperature of 110 °C. When the temperature is higher than 150 °C, although the film still shows a relatively flat characteristic, the EVA adhesive adheres to each other in the middle channel region due to the mutual attraction of the molten EVA liquids from the upper and lower surfaces of the laminating films. Therefore, it can be concluded that the lamination temperature should not be higher than 130 °C.

Fig. 4c shows the influence of rolling speed on channel dimension. In this experiment, we chose the 100 μm thick laminating film and set the roller temperature at 120 °C. The rolling speed was changed from 1.05 m min⁻¹ to 2.1 m min⁻¹. The experimental results in the left figure show that the channels with aspect ratios of 1 : 1 and 6 : 1 do not deform in the whole speed range, and a stable deformation is observed in the channel with an aspect ratio of 11 : 1. The measured cross-sectional profiles and the inset microscope image in the right of Fig. 4c further validate the above conclusions. Therefore, we can conclude that the rolling speed is not the decisive factor for the effective lamination of the polymer film chip in our experiment.

High-throughput inertial isolation application

The circulating tumour cells (CTCs) in the blood of cancer patients are regarded as the precursors for the generation of secondary tumors during metastasis^33,34 and can be used to assess therapeutic treatment effect³⁵ and to help personalization therapy.³⁶ Therefore, the isolation and enrichment of CTCs from blood is of great importance for monitoring and understanding the cancer metastasis. To the purpose of validating the practical application capability of our polymer chip, we developed an integrated microfluidic device, and then applied this device for achieving the high-throughput inertial isolation of cancer cells from human blood.

The integrated device is composed of a polymer chip, two transparent PMMA housings, two home-made stainless steel fixtures, and other auxiliary parts, as shown in Fig. 5a. The polymer chip includes a spiral microchannel of 500 μm wide and 120 mm long. To hold and clamp the chip, the PMMA housings with fluid access tubes were used to sandwich the chip, and the steel fixtures were fastened via four bolts. Leak-free connexion was achieved by inserting O-rings to the upper and lower interfaces between the chip and the housings. Fig. 5b shows the image of the finished device for cell isolation, and a schematic diagram of the spiral channel topology is illustrated in the inset of this figure. The underlying physics for cell inertial isolation in spiral microchannels can be interpreted by the coupling of inertial migration and Dean flow effects. Due to the influence of strong inertial lift force, large sized particles which satisfy the focusing criterion a_p/h > 0.07 (ref. 37) (here a_p is the particle diameter, and h is the channel depth) can be focused at a specific position near the inner wall region, while the small sized particles migrate to form a tight band or disperse to the channel width under the dominate Dean drag force. Therefore, the large and small particles can be isolated when the small particles migrate to form a particle band located near the outer wall at outlets. More details on the fundamentals of particle focusing and separation in spiral microchannels can be found in our and others' previous works.^28,38,39

Fig. 5 (a) Explosive drawing of the integrated microfluidic device for cell isolation; (b) photograph of the finished device. The inset shows the schematic diagram of the spiral channel in the polymer film chip.

To determine the optimal flow rate for efficient cell isolation, four differently sized polystyrene microbeads (5 μm, 7.32 μm, 15 μm, and 20 μm) were first employed to mimic polydisperse-sized cancer cells and blood cells, respectfully. Since the isolation performances of microbeads are directly influenced by the total Dean flow rate (combination of sample and sheath flow rates), it is of great importance to understand the transportation dynamics of particles under varied flow rates. During the particle isolation experiment, we fixed the sample flow rate to be 0.2 ml min⁻¹, and then regulated the sheath flow rate to match the sample fluid for achieving an effective isolation of particle mixtures. Fig. 6a shows the microbead isolation performances under varied total Dean flow rates from 1.4 ml min⁻¹ to 1.8 ml min⁻¹. For clear demonstration, we used green and red lines to respectively represent the lower boundaries of small microbead bands and the averaged focusing positions of large microbeads. We also made a quantitative characterization of the lateral focusing position of large microbeads and the band width of small microbeads as a function of flow rate, as illustrated in Fig. S3a.† It can be seen from these figures that with the increase of flow rate, the bands of small microbeads gradually narrow and shift away from the array of large microbeads at first, and then broadens again, resulting in the decrease of the gap between large and small microbeads. Instead, the large microbeads always focus near the inner wall region, and form a narrow bead array under all the tested flow rates. To quantitatively analyse the isolation performance under varied flow rates, the gap between large and small microbeads under varied flow rate was also measured (see Fig. S3b†). We found that the large and small microbeads are well separated under all the tested flow rates, and the best isolation performance is obtained under a flow rate of 1.65 ml min⁻¹. We also captured the microbead dynamics at different positions along the spiral channel under the optimal isolation flow rate, as illustrated in Fig. S3c.† From this figure, it is validated that a precision isolation of large and small microbeads is successfully achieved at outlets (the microscope images of collected samples are provided in Fig. S3d†).

Fig. 6 High-throughput inertial isolation experiments using our integrated device. (a) Composite images illustrating the isolation performances of polystyrene microbeads near the outlets under varied total Dean flow rates. Yellow lines represent the lower boundaries of the small microbead bands, and red lines indicate the focusing positions of large microbeads. (b) Bright field composite image illustrating the high-throughput isolation of cancer cells from blood cells, and the fluorescence composite image highlighting the focusing position of cancer cells. (c) The photograph of isolated cancer cell and blood cell samples collected from the inner outlet and the outer outlet. (d) Sampled microscope image of the isolated blood cell samples. (e) Sampled microscope (bright field or fluorescence) images of the enriched cancer cells. Cells inside the red circles in the bright field image are the stained cancer cells, and the blue circles denote the suspicious blood cells. All of the scale bars in these figures are 30 μm.

Based on the determined optimal flow rate, we performed the high-throughput inertial isolation of cancer cells from the human blood. The overlaid bright-field (for blood cells) or fluorescent (for cancer cells) images of cell migration near the outlets are illustrated in Fig. 6b. As can be seen from these images, most of the small sized blood cells are collected via the outer outlet, while almost all of the large sized cancer cells focus near the inner wall and are exported through the inner outlet. Fig. 6c shows the photograph of isolated sample fluids of cancer cells and blood cells collected from the two outlets. It can be seen from this figure that the suspension collected from the inner outlet is colourless, whereas the suspension from the outer outlet appears light red due to the containing of overwhelming quantities of blood cells. To further characterize the cell isolation efficiency, optical microscope images of the two collected sample fluids were captured, as illustrated in Fig. 6d and e. Fig. 6d shows the microscope image of the sample collected from the outer outlet. As can be observed in the figure, no cancer cells are found in the visual field. As a comparison, from the bright filed and the fluorescent image of the suspension collected from inner outlet (see Fig. 6e), we found that the cancer cells are highly enriched and only a little blood cells are mixed into the isolated cancer cell suspension, which indicates a good isolation efficiency of our polymer chip. The successful cancer cell isolation from human blood well validates the applicability of our polymer film chip for biomedical applications where an extremely high processing throughput is required for timely analyse.

Discussion

In this work, UV laser direct writing technique was employed to cut channel patterns within a PVC film. The maximum dimension tolerance between the actually measured channel dimension and the designed dimension is ∼20 μm in this work due to the inherent diameter of the focused laser beam. This can be overcome by appropriately lowering down the designed dimension of channels (i.e., a desired channel dimension of 500 μm wide can be designed as 480 μm for laser machining). For most microfluidic applications with large channel structures (e.g., inertial isolation in this work), this machining tolerance can also be neglected. The laser spot size of our UV laser system is ∼20 μm, which should theoretically allow for channel width just slightly >20 μm. In order to cut through the material completely, the laser spot should be adjusted beneath the lower surface of the film. Since the cone shape of the laser beam, the minimum achievable channel width in this work is ∼50 ± 5 μm according to the used relatively thick film of ∼160 μm. In addition, as compared with other more advanced laser systems (e.g., femtosecond lasers), complicated 3D architectures cannot be machined via the UV lasers, which may be solved through tacking several films of different architectures.

Although the UV laser direct writing technique imposes structure limitations and dimension tolerance restrictions on our device, we still successfully fabricated polymer film chips for high-throughput cell inertial isolation application. The overall isolation performance of our polymer film chip is similar to that of reported PDMS devices. It is noted that the significant advantages of our polymer film chip are the ultra-low cost and the fast fabrication capability over other rapid prototyping methods (a comprehensive comparison of current rapid prototyping methods is provided in Table S1†). Since the raw materials of the chip are inexpensive office supplies, the chip cost can be reduced to be approximately 1.5 cents per chip. In addition, due to the fast cutting speed of the laser system and the simple channel sealing process via lamination, the design-to-device time of the chip can be finished within 20 minutes in a non-clean room. Another appealing advantage of our polymer film chip is the simple in-process dimension control characteristics. Since the channel depth is directly dependent on the thickness of the PVC film, manufacturers only need to check and control the initial film dimensions of the incoming raw materials, while the channel width and the cross-sectional profiles can be easily monitored by using the in-process electronic visual sensors. In addition, since the technique does not need any precise master mold, and thus is very flexible in updating or redesigning of the architectures of their developing microfluidic devices. Given these special features, this technique can be applied both in laboratory scale and industrial batch production mode. Moreover, due to the ultra-low-cost and fast fabrication characteristics of the polymer film chip, the integrated microfluidic systems will have a good prospect for the potential commercialization in the future.

Materials and methods

Materials of polymer film chip

Transparent PVC hard film (160 μm thick) was introduced as the material for the fabrication of microchannel layer. Transparent PET film with thermal sensitive EVA on one side was acted as the laminating film, which was used to seal and bond the patterned PVC film. Five types of the PET/EVA laminating films with different thicknesses (thicknesses of PET film/thicknesses of EVA adhesive = 50/25 μm, 57/28 μm, 70/30 μm, 75/50 μm, and 85/65 μm) were employed in this work. All the PVC film and PET/EVA laminating films are inexpensive and easily available office supplies, which were directly purchased from the local market.

Sample preparation

Blood samples were obtained from healthy donors. This study was reviewed and approved by the institutional review board of the affiliated Zhongda hospital. In order to reduce the background noise of the overwhelming red blood cells (RBCs), we extracted the buffy coat from the healthy blood samples after a standard centrifugation, which leaves approximately 20% of total RBCs in the extracted sample. Human breast adenocarcinoma cell lines MCF-7 from cancer patients were used to mimic cancer cells, and the cells were stained with calcein AM (acetoxymethyl ester) (C3100MP, Invitrogen) for the fluorescence observation. To better observe the isolation performance of cancer cells, sample buffer fluid (1× PBS) was added with a relatively high concentration of cancer cells (∼10⁵ cells per ml) and extracted blood cell samples (∼10⁸ cells per ml). Mixed suspension, composed of four polystyrene microbeads (1% solids) with different diameters (5 μm, 15 μm, and 20 μm from Thermo Fisher Scientific, Inc., and 7.32 μm from Bangs Laboratories, Inc., each accounting for 0.005% of the volume) was introduced to investigate the optimal flow rate for cell isolation.

Device characterization

The image of channel side wall was captured through SEM observation (Helios NanoLab 600i). For inertial isolation experiments, sample and sheath fluids (1× PBS) were filled in the 10 ml and 20 ml syringes, and two syringe pumps (LEGATO 270, KD Scientific) were used for the accurate and stable introduction of the two fluids, respectfully. The integrated microfluidic device was fixed on the stage of an inverted fluorescent microscope (IX71, Olympus) equipped with a high-speed CCD camera (Retiga EXi, QImaging). The motion trajectories of microbeads/cells inside the microfluidic channel were observed through a 10× objective lens, and the experimental images were captured by the CCD camera, and subsequently analysed using IMAGE-PRO Express software (Media Cybernetics, Inc.).

Acknowledgements

This research work was supported by the National Natural Science Foundation of China (51505082, 51505083, 51375089), and the Natural Science Foundation of Jiangsu Province (BK20150606).

Notes and references

1. Y. Sun and Y. C. Kwok, Anal. Chim. Acta, 2006, 556, 80–96.
2. E. A. Oblath, W. H. Henley, J. P. Alarie and J. M. Ramsey, Lab Chip, 2013, 13, 1325–1332.
3. B. S. Lee, Y. U. Lee, H. S. Kim, T. H. Kim, J. Park, J. G. Lee, J. Kim, H. Kim, W. G. Lee and Y. K. Cho, Lab Chip, 2011, 11, 70–78.
4. N. M. Karabacak, P. S. Spuhler, F. Fachin, E. J. Lim, V. Pai, E. Ozkumur, J. M. Martel, N. Kojic, K. Smith, P. I. Chen, J. Yang, H. Hwang, B. Morgan, J. Trautwein, T. A. Barber, S. L. Stott, S. Maheswaran, R. Kapur, D. A. Haber and M. Toner, Nat. Protoc., 2014, 9, 694–710.
5. M. Wu and A. K. Singh, Curr. Opin. Biotechnol., 2012, 23, 83–88.
6. D. C. Duffy, J. C. McDonald, O. J. Schueller and G. M. Whitesides, Anal. Chem., 1998, 70, 4974–4984.
7. J. Melin and S. R. Quake, Annu. Rev. Biophys. Biomol. Struct., 2007, 36, 213–231.
8. R. Mukhopadhyay, Anal. Chem., 2007, 79, 3248–3253.
9. T. W. de Haas, H. Fadaei and D. Sinton, Lab Chip, 2012, 12, 4236–4239.
10. S. Miserere, G. Mottet, V. Taniga, S. Descroix, J. L. Viovy and L. Malaquin, Lab Chip, 2012, 12, 1849–1856.
11. A. Toossi, H. Moghadas, M. Daneshmand and D. Sameoto, J. Micromech. Microeng., 2015, 25, 085008.
12. L. Peng, Y. Deng, P. Yi and X. Lai, J. Micromech. Microeng., 2014, 24, 013001.
13. P. Utko, F. Persson, A. Kristensen and N. B. Larsen, Lab Chip, 2011, 11, 303–308.
14. D. J. Guckenberger, T. E. de Groot, A. M. Wan, D. J. Beebe and E. W. Young, Lab Chip, 2015, 15, 2364–2378.
15. M. I. Mohammed, E. Abraham and M. P. Y. Desmulliez, J. Micromech. Microeng., 2013, 23, 035034.
16. D. Bartholomeusz, R. W. Boutté and J. D. Andrade, J. Microelectromech. Syst., 2005, 14, 1364–1374.
17. A. K. Au, N. Bhattacharjee, L. F. Horowitz, T. C. Chang and A. Folch, Lab Chip, 2015, 15, 1934–1941.
18. P. Spicar-Mihalic, B. Toley, J. Houghtaling, T. Liang, P. Yager and E. Fu, J. Micromech. Microeng., 2013, 23, 067003.
19. L. E. Criales, P. F. Orozco, A. Medrano, C. A. Rodríguez and T. Özel, Mater. Manuf. Processes, 2015, 30, 890–901.
20. S. A. Mahoney, T. E. Rufford, V. Rudolph, K.-Y. Liu, S. Rodrigues and K. M. Steel, Int. J. Coal Geol., 2015, 144–145, 48–57.
21. B. B. Xu, Y. L. Zhang, H. Xia, W. F. Dong, H. Ding and H. B. Sun, Lab Chip, 2013, 13, 1677–1690.
22. K. Sugioka, J. Xu, D. Wu, Y. Hanada, Z. Wang, Y. Cheng and K. Midorikawa, Lab Chip, 2014, 14, 3447–3458.
23. D. Y. Stevens, C. R. Petri, J. L. Osborn, P. Spicar-Mihalic, K. G. McKenzie and P. Yager, Lab Chip, 2008, 8, 2038–2045.
24. F. C. Huang, Y. F. Chen and G. B. Lee, Electrophoresis, 2007, 28, 1130–1137.
25. M. T. Koesdjojo, Y. H. Tennico, J. T. Rundel and V. T. Remcho, Sens. Actuators, B, 2008, 131, 692–697.
26. R. Truckenmüller, R. Ahrens, Y. Cheng, G. Fischer and V. Saile, Sens. Actuators, A, 2006, 132, 385–392.
27. Y. Wang, H. Chen, Q. He and S. A. Soper, Electrophoresis, 2008, 29, 1881–1888.
28. N. Xiang, H. Yi, K. Chen, D. Sun, D. Jiang, Q. Dai and Z. Ni, Biomicrofluidics, 2013, 7, 44116.
29. M. E. Warkiani, A. K. Tay, G. Guan and J. Han, Sci. Rep., 2015, 5, 11018.
30. C. Liu, C. Xue, X. Chen, L. Shan, Y. Tian and G. Hu, Anal. Chem., 2015, 87, 6041–6048.
31. X. Lu, J. DuBose, S. W. Joo, S. Qian and X. Xuan, Biomicrofluidics, 2015, 9, 014108.
32. J. Wang, W. Chen, J. Sun, C. Liu, Q. Yin, L. Zhang, Y. Xianyu, X. Shi, G. Hu and X. Jiang, Lab Chip, 2014, 14, 1673–1677.
33. C. Jin, S. M. McFaul, S. P. Duffy, X. Deng, P. Tavassoli, P. C. Black and H. Ma, Lab Chip, 2014, 14, 32–44.
34. J. Wang, W. Lu, C. Tang, Y. Liu, J. Sun, X. Mu, L. Zhang, B. Dai, X. Li, H. Zhuo and X. Jiang, Anal. Chem., 2015, 87, 11893–11900.
35. D. F. Hayes, M. Cristofanilli, G. T. Budd, M. J. Ellis, A. Stopeck, M. C. Miller, J. Matera, W. J. Allard, G. V. Doyle and L. W. Terstappen, Clin. Cancer Res., 2006, 12, 4218–4224.
36. K. Ameri, R. Luong, H. Zhang, A. A. Powell, K. D. Montgomery, I. Espinosa, D. M. Bouley, A. L. Harris and S. S. Jeffrey, Br. J. Cancer, 2010, 102, 561–569.
37. D. Di Carlo, D. Irimia, R. G. Tompkins and M. Toner, Proc. Natl. Acad. Sci. U. S. A., 2007, 104, 18892–18897.
38. H. W. Hou, M. E. Warkiani, B. L. Khoo, Z. R. Li, R. A. Soo, D. S. Tan, W. T. Lim, J. Han, A. A. Bhagat and C. T. Lim, Sci. Rep., 2013, 3, 1259.
39. X. Zhang, N. Xiang, W. Tang, D. Huang, X. Wang, H. Yi and Z. Ni, Lab Chip, 2015, 15, 3473–3480.

---

Footnote

† Electronic supplementary information (ESI) available: Further information about the comparison of various rapid prototyping methods, the experimental setup for bond strength test, the measured results of cross-sectional features of channel with aspect ratios of 1 : 1 and 6 : 1, the quantitative analysis of microbead isolation performance, and a video illustrating the high-throughput cell inertial isolation can be found in the ESI. See DOI: 10.1039/c5ra27092h

---