Fabrication of polydimethylsiloxane nanofluidic chips under AFM tip-based nanomilling process
- 110 Downloads
In current research realm, polydimethylsiloxane (PDMS)-based nanofluidic devices are widely used in medical, chemical, and biological applications. In the present paper, a novel nanomilling technique (consisting of an AFM system and a piezoelectric actuator) was proposed to fabricate nanochannels (with controllable sizes) on PDMS chips, and nanochannel size was controlled by the driving voltage and frequency inputted to the piezoelectric actuator. Moreover, microchannel and nanochannel molds were respectively fabricated by UV lithography and AFM tip-based nanomilling, and finally, PDMS slabs with micro/nanochannels were obtained by transfer process. The influences of PDMS weight ratio on nanochannel size were also investigated. The bonding process of microchannel and nanochannel slabs was conducted on a homemade alignment system consisted of an optical monocular microscope and precision stages. Furthermore, the effects of nanochannel size on electrical characteristics of KCl solution (concentration of 1 mM) were analyzed. Therefore, it can be concluded that PDMS nanofluidic devices with multiple nanochannels of sub-100-nm depth can be efficiently and economically fabricated by the proposed method.
KeywordsAtomic force microscopy Nanomilling Nanochannels Nanofluidic chip
Atomic force microscope
Electric double layer
Due to their considerable potentials in chemical, medical, and biological fields, micro/nanofluidic systems are widely used in DNA analysis [1, 2, 3, 4], cell separation , protein research [6, 7, 8], food safety , and environmental monitoring . With the rapid development of nanofabrication technology, the demand for nanofluidic devices with the one-dimensional size smaller than 100 nm is continuously increasing . Nanofluidic chips can also be effectively used for virus detection , nanoparticle manipulation , and the study of ion diffusion . However, the detection efficiency and sensitivity of the nanofluidic chips depend on the feature dimensions and distribution of the nanochannels. It is indispensable to accurately control the feature dimensions of the nanochannels for nanofluidic-based label-free detection. How to fabricate nanochannels with controllable feature dimensions and distribution is still a challenge for the application in the nanofluidic field.
Till now, there are several methods that can be utilized for the fabrication of nanofluidic chips. Reactive ion etching , conventional photolithography , high-energy beam processing , interference lithography , nanoimprinting , and hot embossing technologies [20, 21] are most commonly used for the fabrication of nanofluidic devices; however, all of these methods manifest their own limitations. Reactive ion etching and conventional photolithography are the mainstream methods for micro/nanofluidic channel fabrication. However, the lateral dimensions of the fabricated channels depend on the wavelength of the incident light, thus the widths of the produced channels are often found in the micrometer scale, not in nanoscale . Besides, it is inconvenient to change the photomasks when fabricate micro/nanostructures have different features. Focused ion beam lithography (FIB) and electron beam lithography (EBL) both are high-energy beam processing methods, which can easily fabricate high-precision nanofluidic chip with sub-100 nm nanochannels. However, the investment for the fabrication facility is extremely high and the strict environmental requirement is necessary . Interference lithography (IL) is suitable for fabricating simple periodical structures over a large area; however, it is not suitable to machine a single nanochannel [24, 25]. The processing resolution of nanoimprinting depends on template properties, the crucial issue for this approach is how to fabricate the template with high-precision nanostructures . In addition, sacrificial molding and creak-based method are also adopted to fabricate micro/nanoscale devices [27, 28]; however, the accurate control of nanochannel size is very difficult in these approaches. Thus, a more feasible fabrication approach with the properties of high machining precision, ease of use, large processing range and low environment requirement is demanded for the fabrication of nanofluidic device.
In recent years, due to their high machining accuracies, ultra-precision machining methods, such as nanomilling, precision grinding, and ultra-precision turning, are widely used in micro/nanostructure fabrication [29, 30, 31, 32]. Moreover, since the invention of atomic force microscope (AFM) in 1986, AFM tip-based nanofabrication is a powerful method to prepare nanostructures . The traditional tip-based nanoscratching possesses some limitations, such as limited machining width and low fabrication efficiency. The width of the nanochannel fabricated by this approach is dependent on the geometry of the AFM tip, which signifies the nanochannels with controllable width which are inaccessible. In addition, the fabrication efficiency of the traditional tip-based nanoscratching process is relatively low especially for the case of employing a feed in the machining process to enlarge the depth and width of the obtained nanostructure. Due to its significant advantages, such as controllable machining size and high fabrication efficiency, tip-based nanomilling is widely adopted to fabricate nanochannels. Gozen et al. [34, 35] fabricated nanostructures on polymethyl methacrylate (PMMA) through a nanomilling process. Zhang et al. [36, 37, 38] prepared three-dimensional nanostructures using AFM and studied the effects of different machining parameters. Park et al.  investigated the mechanism of nanomachining process and found that the intensities of cutting force were significantly reduced; however, in the proposed system, the machining facilities were found to be relatively complicated and the material removal process was not investigated in details. The relationship between the machining parameters including the driving frequency and voltage and the feature dimensions of obtained nanochannel was not studied. In addition, their work did not focus on the application of the fabricated nanochannels. Therefore, more work is needed to explore the application scope of this AFM tip-based nanomilling approach. Polycarbonate (PC), due to its excellent machinability, is commonly used for nanofabrication ; nevertheless, it is rarely selected to fabricate nanofluidic chips. In contrast, polydimethylsiloxane (PDMS) is widely used to process microfluidic and nanofluidic chips. Mata et al.  studied the influences of PDMS weight ratio on tensile stress. Park et al.  developed a new method to improve the stiffness of PDMS. The applications of nanofluidic chips in label-free test field mainly depend on the electrical conductivity of nanochannels , thus the measurement results are often affected by the dimensional sizes of nanochannels .
Therefore, in order to overcome the disadvantages of traditional tip-based scratching process, the nanomilling approach is employed to conduct the fabrication process of nanochannel in this study. Moreover, PC sheet was selected as the experimental sample to mitigate tip wear as well as to reduce fabrication cost. Further, nanochannel size on PC sheet was controlled by the driving voltage and frequency inputted to the piezoelectric actuator. The influences of PDMS weight ratio on nanochannel size were also investigated. Furthermore, in order to verify the effects of different dimensional sizes on electric conductivity of nanochannels, the current measurement test was performed using KCl solution.
Nanomilling system setup
Fabrication of nanochannel and microchannel molds
Hence, the normal loads for nanomilling process were set as 17 μN and 25 μN. Further, for comparison, nanochannel molds on PC sheet were also fabricated without vibration, this method is called single scratching. The normal loads for the single scratching process were set as 25 μN, 33 μN, 42 μN, 50 μN, and 58 μN. The schematic diagram of nanochannel mold cross-section is displayed in Fig. 2(b2).
Microchannel molds were prepared by UV lithography process. The flowchart in Fig. 2(a1–a4) depicts the operation details of lithography process. The photoresist (SU-82015; MicroChem, USA) was spin-coated on Si substrate at 500 rps for 30 s and at 4000 rps for 120 s. A pair of “U”-shaped microchannels formed the microchannel chip (Fig. 2(a6)), which was bridged by nanochannels to form the final nanofluidic chip. The width of the microchannel was 30 μm and the diameter of the reservoir was 1 mm. Moreover, the distance between two “U”-shaped microchannels was 50 μm (Additional file 1: Figures S1 and S2).
Transfer printing of microchannels and nanochannels
The convex microchannel mold (Fig. 2(a4)) and the concave nanochannel mold (Fig. 2(b2)) were transferred by PDMS (Sylgard 184, Dow Corining, USA) to prepare the final nanofluidic chip. Figure 2 (b3)–(b6) present the technological process of nanochannel mold transfer, which consisted of two steps: first transfer and second transfer. To investigate the effects of the weight ratio of monomer to curing agent on nanochannel size, three different PDMS weight ratios (A-PDMS) were employed during both first and second transfer processes. The PDMS weight ratios for the first transfer printing process were set as 9:1, 7:1, and 5:1, whereas the values for the second transfer were set as 10:1, 9:1, and 8:1. Figure 2(a5) and (a6) displays the transfer process of microchannel mold using one-step transfer approach. The PDMS weight ratio of 10:1 was used for the transfer of convex microchannel. During all of the transfer printing processes, two-component PDMS elastomer was first uniformly stirred and then poured into a case to prepare the mold. The case was then kept in a vacuum desiccator for 30 min and degassed for 2–3 times to remove all of the trapped air bubbles. The prepared mold was kept in a heating oven at 80 °C for 4 h, and finally, the PDMS replica was gently peeled off from the mold.
Results and discussion
Rotary trajectory of piezoelectric actuator
The two-dimensional piezoelectric actuator is a critical component to conduct rotary motion in an AFM tip-based nanomilling system. Therefore, to characterize its motions under a range of driving voltages and frequencies, preliminary scratching tests were carried out. Under the contact model with a scan range of 0 nm, the AFM tip first approached the surface of PC sheet under a given normal load and was kept static. The rotation of the two-dimensional piezoelectric actuator was controlled by a pre-set frequency and voltage. After the completion of scratching process, the AFM tip was lifted up from the surface of PC sheet. Thus the motion amplitude of the piezoelectric actuator was obtained as a function of driving voltage and frequency. The driving voltages were set in the range of 30–150 V with a spacing of 30 V, whereas the driving frequencies were set as 100 Hz and 1500 Hz. The relationship between measured amplitudes and driving voltages at two driving frequencies is displayed in Additional file 1: Figure S3. It is evident that the values of machining amplitude increased with the increasing driving voltages, and the value of machining amplitude at 1500 Hz was greater than that of 100 Hz. It was found that the widths of the nanochannel fabricated by our proposed method ranged from 350 nm to 690 nm.
Fabrication of nanochannel molds on PC sheet
The values of mean strain rate at 100 Hz were found in the range of 1.42 × 104 s-1~ 2.27 × 104 s-1. The values of local normal pressure (p) started to go up with the increasing of strain rates when the strain rates ranged from 1.42 × 104 s-1 to 2.27 × 104 s-1 . The value of τ was much smaller than that of p, it signifies that normal load mainly depended on p. Therefore, in order to keep the values of normal load (FN) constant during the entire machining process, the values of machining depth should be smaller at higher driving voltages. However, the final dimensional size of the fabricated nanochannel was affected by the recovery of the sample material. The recovery of the sample decreased with the increasing machining speeds in the range of 142~227 μm/s : thus, it indicates that a higher elastic recovery occurred at 30 V. Consequently, the depth of the fabricated nanochannel at 30 V (~142 μm/s) was shallower than that of 60 V (~161 μm/s). Additional file 1: Figure S4(a) and Fig. 5b are the typical AFM images of the nanochannel machined at 100 Hz under normal loads of 17 μN and 25 μN, respectively. It is obvious that the pile-up at the right side of the nanochannel is larger than the left. The rotary motion of the sample during nanomilling process is anticlockwise, and the cutting angle of the main cutting edge is changing with the rotation. The uncut chip thickness is too small to form chips at the beginning and end of a cycle of nanomilling process. The uncut chip thickness at the middle of a cycle of nanomilling process is relatively large; however, the small attack angle contributes to the formation of the pile-ups. Thus, more materials are pushed to the right side of the channel, and the pile-ups are thus asymmetric. The details for the formation of asymmetric pile-ups can be found in our previous study .
The values of α and β were kept constant during the entire machining process. The values of strain rate at 1500 Hz were found in the range of 2.03 × 105~3.66 × 105 s-1; hence, it can be assumed that local normal pressure (p) reached its limit value at 1500 Hz. Furthermore, machining speed manifested no effect on the recovery of the sample during machining at 30– 150 V (~ 2.03–3.66 mm/s) ; thus, the final dimensional sizes of the nanochannel were only determined by machining dimensions. The values of AS2 (Fig. 6(b)) were found to be smaller than those of AS1 (Fig. 6(a)) for larger machined widths, and according to Eq. (9), the value of D was larger for a smaller value of AS. Therefore, the values of machining depth increased with the increasing driving voltages. A typical AFM image of the nanochannel fabricated under a normal load of 25 μN, a driving voltage of 120 V, and a frequency of 1500 Hz is presented in Additional file 1: Figure S4(b). It is noticeable that materials were removed in both chip and pile-up formation, and the expelled materials accumulated only on one side of the nanochannel. Moreover, the expelled materials accumulated in chip formation at the bottom of the nanochannel during machining at 150 V under a normal load of 25 μN. Therefore, the size data of the fabricated nanochannel during machining at a voltage of 150 V and a frequency of 1500 Hz (under a normal load of 25 μN) was empty in Fig. 4d.
It is evident from Fig. 4c that nanochannel width started to increase with the increasing driving voltages. Moreover, when the values of normal load and driving voltage were kept constant, the width of the nanochannel fabricated at a frequency of 1500 Hz was wider than that of 100 Hz. Moreover, machining depth of the nanochannel fabricated at 1500 Hz was deeper than that of 100 Hz, and the cross-sectional size of the tip was found to be larger during the machining of a deeper nanochannel. Therefore, the nanochannels were fabricated wider when machining deeper.
First transfer of nanochannel molds
Nanochannels machined by single scratching method under normal loads of 25 μN, 33 μN, 41 μN, 50 μN, and 58 μN were applied to the first transfer process. Moreover, nanochannel molds fabricated by nanomilling at a frequency of 100 Hz in the driving voltage range of 30–150 V (with a spacing of 30 V) were also used in the transfer process. Nanochannels (80 nm depth and 510 nm width) machined by single scratching method were termed as “nanochannel I”, whereas nanochannels (50 nm depth and 610 nm width, 90 nm depth and 630 nm width) fabricated by nanomilling were called as “nanochannel II” and “nanochannel III,” respectively. Three different PDMS weight ratios (5:1, 7:1, and 9:1) were used in the first transfer process.
Second transfer of nanochannel molds
The depths of nanochannels obtained from walls II and III were larger than the original machining size, whereas the depth obtained from wall I was smaller than the initial machining size. Furthermore, the changes in width were identical to the changes in depth. The aspect ratio of wall I was larger than those of walls II and III, thus each wall manifested different thermal expansion values. Hence, the changing trends of width and depth during second transfer were different though at the same PDMS weight ratio. The values of the depth and width of walls II and III at 9:1 and 8:1 were found to be closer to the original machining size compared with 10:1. Because the elastic recoveries of PDMS at 9:1and 8:1 are closer to 5:1 than 10:1, which indicates an almost similar recovery trend for PDMS at 9:1, 8:1, and 5:1.
Application of nanochannel devices in electric current measurement
In the present research, nanochannels with controllable sizes (sub-100-nm depth) were fabricated by AFM tip-based nanomilling, and for the first time, the machined nanochannels were applied to prepare nanofluidic devices. The multichannel nanofluidic devices were prepared in four steps: (1) fabrication of nanochannels by AFM tip and piezoelectric actuator, (2) fabrication of microchannels by lithography, (3) transfer of micro- and nanochannels, and (iv) bonding. Further, nanochannel sizes were controlled by changing the driving voltages and frequencies inputted to the actuator. The heights of the wall obtained during first transfer were smaller than the original machining size, whereas the widths were larger than the original machining size. The experiment results revealed that during second transfer process, nanochannel sizes affected PDMS weight ratios. Finally, micro-nanofluidic chips with three different nanochannel sizes were obtained by bonding a PDMS nanochannel chip on a PDMS microchannel chip. Moreover, the electrical current measurement experiment was conducted on the fabricated nanofluidic chips, and it was found that the values of current were affected by nanochannel sizes. Therefore, PDMS nanofluidic devices with multiple nanochannels of sub-100-nm depth can be efficiently and economically fabricated by the proposed method.
Compared with other fabrication approach, the proposed method for fabrication of the nanofluidic devices in the study is easy to use and low cost; besides, the nanochannels with controllable dimension size can be obtained easily. However, the commercial AFM system cannot equip with a large-scale high-precision stage due to the spatial limitation; thus, the maximum fabrication length of the nanochannel is confined as 80 μm. In addition, the tip wear cannot be neglected after long-term fabrication due to the high machining speed, which should be investigated in future work.
The authors gratefully acknowledge the financial supports of the National Natural Science Foundation of China (51705104, 51675134), the Foundation for Innovative Research Groups of the National Natural Science Foundation of China (51521003), Key Laboratory of Micro-systems and Micro-structures Manufacturing of Ministry of Education (Harbin Institute of Technology No. 2017KM005) and the China Postdoctoral Science Foundation (No. 2017 M610206 and 2018 T110289).
Availability of Data and Materials
All data generated or analyzed during this study are included in this published article.
YDY and YQG carried out the design and drafted the manuscript. JQW participated in the experiments. YG and ZF assisted with the optimization and proofed the manuscript. All authors read and approved the final manuscript.
The authors declare that they have no competing interests.
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
- 54.Geng Y, Li H, Yan Y, He Y, Zhao X (2017) Study on material removal for nanochannels fabrication using atomic force microscopy tip-based nanomilling approach. P. I. Mech. Eng. B-J Eng. 233:461–469Google Scholar
Open AccessThis article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made.