Here we present a protocol for wide-area scanning probe nanolithography enabled by the iterative alignment of probe arrays, as well as the utilization of lithographic patterns for cell-surface interaction studies.
Scanning probe microscopy has enabled the creation of a variety of methods for the constructive ('additive') top-down fabrication of nanometer-scale features. Historically, a major drawback of scanning probe lithography has been the intrinsically low throughput of single probe systems. This has been tackled by the use of arrays of multiple probes to enable increased nanolithography throughput. In order to implement such parallelized nanolithography, the accurate alignment of probe arrays with the substrate surface is vital, so that all probes make contact with the surface simultaneously when lithographic patterning begins. This protocol describes the utilization of polymer pen lithography to produce nanometer-scale features over centimeter-sized areas, facilitated by the use of an algorithm for the rapid, accurate, and automated alignment of probe arrays. Here, nanolithography of thiols on gold substrates demonstrates the generation of features with high uniformity. These patterns are then functionalized with fibronectin for use in the context of surface-directed cell morphology studies.
Progress in nanotechnology is dependent on the development of techniques capable of efficiently and reliably fabricating nanoscale features on surfaces.1,2 However, generating such features over large areas (multiple cm2) reliably and at relatively low cost is a non-trivial endeavor. Most existing techniques, derived from the semiconductor industry, rely on ablative photolithography to fabricate 'hard' materials. More recently, lithographic techniques derived from scanning probe microscopy (SPM) have emerged as a convenient and versatile approach for the rapid prototyping of nanoscale designs.3 SPM-based techniques are able to conveniently and rapidly 'write' any user-defined pattern. The most well-known of these is dip-pen nanolithography (DPN), pioneered by Mirkin et al.,4 where a scanning probe is used as a 'pen' to transfer a molecular 'ink' to the surface producing features in a fashion analogous to writing. Under ambient conditions, as a probe is scanned across a surface the 'ink' molecules are transferred to the surface via a water meniscus that forms between the probe and the surface (Figure 1). DPN thus allows the nanolithographic deposition of a wide range of materials, including 'soft' materials such as polymers and biomolecules.5 Related techniques using probes engineered with channels for fluid delivery, variously referred to as 'nanopipettes' and 'nano-fountain pens', have also been reported.6,7,8
The main obstacle to the wider application of SPM-derived lithography is throughput, as it requires an excessively long time to pattern centimeter-scale areas with a single probe. Early efforts to address this issue focused on the parallelization of cantilever-based DPN, with both 'one- dimensional' and 'two-dimensional' (2D) probe arrays being reported for the lithography of centimeter-sized areas.5,9 However, these cantilever arrays are produced through relatively complex multistep fabrication methods and are relatively fragile. The invention of polymer pen lithography (PPL) addressed this issue by replacing the standard SPM cantilevers with a 2D array of soft siloxane elastomer probes bonded to a glass slide.10 This simple probe setup significantly decreases the cost and complexity of patterning large areas, opening up nanolithography to a wider range of applications. This cantilever-free architecture has also been expanded to hard-tip soft-spring lithography,11 which provides a hybrid of soft elastomeric backing with hard silicon tips giving improved resolution in comparison to patterns produced using soft elastomer tips.
A crucial factor in the execution of these 2D array technologies is that the probe array must be exactly parallel to the surface substrate so that when lithography is utilized, all the probes come into contact with the surface simultaneously. Even a small misalignment can cause a large difference in feature size from one side of the array to the other, since some probes will come into contact with the surface earlier during the descent of the array, while others will come into contact later or not at all.12 Exact alignment is especially important with PPL due to the deformability of the soft elastomer probes, where the probes contacting the surface earlier will be compressed, leaving a larger footprint on the surface.
The early work on PPL employed purely visual inspection to guide the alignment process, using a camera mounted above the array to observe the deformation of the pyramidal probes as they were brought into contact with the surface.10 Alignment was judged by observing which side of the probes came into contact with the surface first, then adjusting the angle and repeating the procedure in an iterative manner until the difference in contact on each side of the probe was indistinguishable to the eye. As this alignment procedure relies on subjective visual inspection by the operator, reproducibility is low.
Subsequently, a more objective approach has been developed, consisting of a force sensor mounted beneath the substrate to measure the force applied upon contact of the probes on the surface.12 Alignment was thus achieved by adjusting the tilt angles to maximize the force exerted, which indicated that all the probes were simultaneously in contact. This method showed that alignment to within 0.004° of the surface parallel was possible. This 'force feedback levelling' has now been implemented into fully automated systems in two independent reports.13,14 Both use a triad of force sensors mounted either beneath the substrate or above the array and measure the amount of force exerted upon contact between the probe arrays and surface. These systems give high precision, reporting misalignments of ≤0.001° over a 1 cm length scale,14 or ≤ 0.0003°over 1.4 cm.13 These automated alignment systems also provide major savings in operator time and overall time taken to complete the lithography process.
One major application of high-throughput surface fabrication enabled by this technology is the generation of cell culture substrates. It is now well established that cell phenotype can be manipulated by controlling the initial interaction between cells and surface features, and that this can be enhanced at the nanoscale.15 Specifically, scanning probe lithography methods have been shown to be a facile method to produce a variety of nanofabricated surfaces for such cell culture experiments.16 For example, surfaces presenting nanoscale patterns of self-assembled monolayers and extracellular matrix proteins templated by PPL and DPN have been used to study the potential of nano-modified materials in material induced differentiation of stem cells.17
This protocol describes the utilization of a modified atomic force microscope (AFM) system that enables large-area PPL. We detail the detection of force using multiple force sensors as the means of determining probe-surface contact, together with an algorithm that automates the iterative alignment process. Subsequent functionalization of these patterns with the extracellular matrix protein fibronectin and the culture of human mesenchymal stem cells (hMSC) are described, as a demonstration of PPL-fabricated surfaces applied for cell culture.
1. Fabrication of the PPL pen array
2. Array preparation and substrate mounting
3. Preparation of gold substrates for PPL.
4. Automatic alignment of pen array
5. Polymer pen lithography (PPL)
6. Pattern visualization
7. Pattern functionalization with fibronectin
8. Cell culture on nanofabricated surfaces
To check whether the automated alignment had been successful, the graphs plotted from the alignment data (in the spreadsheet from step 4.8) were examined. Where the alignment process had been successful the two plots, corresponding to the angle by which the sample stage has been tilted along the θ and φ axes, showed a series of rising and descending data points. In each of the plots, two linear fits of the data points showed a well-defined intersect "peak" indicating the maximum z-extension and the corresponding angle at which alignment was achieved (Figure 4A and 4B). This process is repeated four times (i.e., twice for each axis) and plotted as a set of four coordinates. The intersection of each pair of coordinates thus shows the overall optimum angles (Figure 4C).13 In cases where the alignment was not successful, it can be observed that their corresponding θ and φ angle plots do not give good quality linear fits, or do not intersect (Figure 5). Such failed alignments are typically as a result of the arrays being improperly trimmed or mounted to the probe holder (steps 1.7, 1.8, and 2.2). In these cases, the arrays were discarded and a new one prepared and mounted (steps 1 and 2), and the alignment process repeated (step 4).
Upon successful alignment and lithography with MHA by PPL, patterned gold substrates were then imaged using lateral force microscopy to examine whether deposition had taken place. A larger area examination of the printed surfaces was also conducted by optical microscopy of the substrates after etching of the gold not protected by the deposited thiol (Figure 6 and Figure 7). However, the etched patterns cannot be used for further functionalization and should only be used to confirm patterning on representative samples of a batch of printed surface substrates. If the etched patterns show blank areas corresponding to individual pens (Figure 8), this result indicates that the production of probe arrays has not been done successfully, and that some probes are damaged or missing. This inhomogeneity of the probes may be due to the use of an old master where the perfluorinated coating has worn away (step 1.3), resulting in some probes being torn away when the array is separated from the master. In these cases, a new master should be used. The result may also be due to the presence of air bubbles trapped between the glass backing and the master (step 1.5), or if the probe array was not cleanly separated from the master after curing (step 1.8).
Florescent microscopy images of the fibronectin functionalized surfaces incubated hMSCs were also collected (Figure 9). In general, all substrates were stable within the in vitro culture environment and the cells adhered and adapted their morphology to the printed patterns in case of smaller isolated 20 x 20 array of features.
Figure 1. Schematic representation of polymer pen lithography showing molecular ink transport via a water meniscus on probe tip. (A) Side view and (B) top view of the polymer pen array indicate that when the probe array and surface substrate are fully aligned, all the probes come into contact with the surface simultaneously, resulting in parallelized lithography. Please click here to view a larger version of this figure.
Figure 2. Schematic diagram of polymer pen lithography set up. (A) Expanded side view of experimental set up where the prepared probe array is attached to probe holder and mounted to AFM scanner. The substrate is placed on the stage, below which are located the three force sensors. (B) A representation of the assembled instrumentation, showing the AFM scan head relative to the sample stage. (C) Bottom view showing force sensor location. Please click here to view a larger version of this figure.
Figure 3. Schematic representation depicting the spectroscopy program for the alignment procedure. The AFM scanner is set to move the probes toward sample by a distance of 10 µm within 100 ms, held at position for 250 ms, followed by a retraction of 10 µm within 100 ms, and then held for 250 ms at the retracted position. The motion is then repeated throughout the alignment process. Please click here to view a larger version of this figure.
Figure 4. Graphs illustrating a successful alignment. Graphs of z position against the tilt angles (A) θ and (B) φ for a successful alignment, where ● indicates the actual values measured and + indicates the best fit with the least-squares method. (C) Graph of φ against θ fitted angles with the four points where the maximum z-position was reached. The intersection point marked is the final optimum tilt angle across both axes. Please click here to view a larger version of this figure.
Figure 5. Graphs illustrating an unsuccessful alignment. Graphs of z position against the tilt angles (A) θ and (B) φ for an unsuccessful alignment, where ● indicates the actual values measured and + indicates the best fit with the least-squares method. (C) Graph of φ against θ fitted angles with the four points where the maximum z-position was reached. No clear optima or intersection point are observed and therefore the optimum alignment angles are not resolved. Please click here to view a larger version of this figure.
Figure 6. Illustrative optical microscopy and atomic force microscopy images of gold substrates that were patterned with MHA by the aligned PPL arrays and then etched. (A) and (B) are sequentially magnified optical microscopy images of the etched patterns; (C) is an AFM topography image of a single grid of patterns. Please click here to view a larger version of this figure.
Figure 7. Illustrative optical microscopy images of gold substrates that were patterned with MHA by the aligned PPL arrays and then etched. (A) and (B) are sequentially magnified optical microscopy images of etched patterns and (C) is a lower magnification image that shows large area homogeneous patterns. Please click here to view a larger version of this figure.
Figure 8. Illustrative optical microscopy image of a gold substrate that was unevenly patterned with MHA and then etched. The intended patterns (shown in the inset) were repeating grids of 20 dot lines arranged in 20 lines, with every two lines produced by increasing the z-axis extension by 1 µm (ranging from 5 to -5 µm). It can be seen that in some areas no patterns are generated, due to missing probes in those locations. In the areas where only two lines of dots are produced, this result indicates that a probe is present but it is not of the same height as the fully functioning probes, so only deposit features when the array is extended to the full z-axis distance. In this image, the contrast has been deliberately altered to enable observation of the partially printed areas. Please click here to view a larger version of this figure.
Figure 9. Epifluorescence microscopy images of hMSCs cultured on the fibronectin arrays templated by PPL. (A) and (B) are high magnification images showing individual cells. (C) Shows an example pattern of the fibronectin array without an adherent cell and (D) is a wide field image of the cells cultured in a grid arrangement (a schematic of the printed pattern is also shown in the inlay). The cells are stained to show fibronectin (red), F-actin (green) and cell nuclei (blue). Please click here to view a larger version of this figure.
This protocol serves to provide users with a convenient methodology to rapidly carry out nanolithographic patterning with high uniformity and controllable feature size over large (cm2) areas. Substrates bearing these large area nanopatterns can then be further elaborated for a variety of applications. One major application of this technology is in the generation of nanofabricated surfaces for cell-surface interaction studies. This report shows some illustrative examples of cell culture on these materials, demonstrating control of hMSC morphology by nanofabricated substrates.
The key enabler of this protocol is the automation of the alignment procedure (step 4) that allows highly uniform and high-throughput production of features on surfaces, down to nanoscale resolution, which enables the rapid turnover of cell culture experiments. The polymer pen lithography carried out using this alignment algorithm is able to generate nanoscale features within approximately 30 min. The reproducibility and accuracy of automated alignment, and thus the uniformity of the patterned features, is however critically dependent on the quality of the probe arrays that are produced (step 1 and 2). Any flaws in their preparation that result in blunt, broken or missing probes; such as trapped air bubbles (step 1.5) or improper separation of the probes from the master (step 1.8) can result in inaccurate alignment and poor quality lithography.
This reported method shares a limitation in common with other alignment methods that rely on force feedback. The accurate determination of when the probes are in contact with the surface is constrained by the need to account for background vibrations caused by the ambient environment and the movement of the sample stage. In general, the sensors have a force sensitivity in the µN regime (2 µN in this case), but the alignment algorithm is designed to only register a force of at least 490 µN as definitive contact between the probes and the surface, in order to avoid any 'false positives' resulting from background noise.13 Thus, this method tends to produce large features (1-2 µm) since the probes must extended a large distance on the z-axis (with a consequent higher force) in order to register contact. In order to compensate, smaller features can be generated by reducing the z-axis distance travelled during the lithography step (e.g., entering the 'Black' setting in step 5.2.3.2 as 3 µm instead of 5 µm).
Nevertheless, even with this limitation, the automation algorithm is able to address a critical aspect in the application of parallelized scanning probe lithography methods, as alignment was previously the most time demanding and imprecise step in the implementation of these techniques. This automation now shifts the rate-limiting step of the fabrication process from the alignment to the lithographic writing itself. While this protocol demonstrates the application of this alignment procedure to PPL, the framework could be applied to a number of SPL techniques such as lipid-DPN26 and matrix-assisted lithography27 as well as potential future catalytic probe systems.28
The authors have nothing to disclose.
The authors acknowledge financial support from a variety of sources including the UK Engineering and Physical Sciences Research Council (grant refs. EP/K011685/1, EP/K024485/1) and a graduate studentship for JH; the Leverhulme Trust (RPG-2014-292); the Wellcome Trust Institutional Strategic Support Fund (105610/Z/14/Z); the British Council (216196834); and the University of Manchester for a University of Manchester Research Institute (UMRI pump priming fund) and a Presidential Doctoral Scholarship to SW. Technical assistance by Dr. Andreas Lieb (Nanosurf AG) is also gratefully acknowledged.
Equipment | |||
FlexAFM mounted on a motorised 5-axis (XYZΘΦ) translation and goniometer stage | NanoSurf | P40008 | |
AFM control software | NanoSurf | C3000 | |
Engraving pen | Sigma-Aldrich | Z225568 | |
Plasma Cleaner | Harrick plasma | PDC-32G-2 | |
PlasmaFlo | Harrick plasma | PDC-FMG-2 | |
Economy Dry Oxygen Service Pump | Harrick plasma | PDC-OPE-2 | |
Tube Rotator | Stuart | SB3 | |
Vacuum Desiccator | Thermo Fisher Scientific | 5311-0250 | |
Milli-Q Water Purification System | Merck Millipore | ZRXQ015WW | |
Modular Humidity Generator | proUmid | MHG32 | |
Proline Plus Pipette 100-1000 µL | Sartorius | 728070 | |
Silicon masters | NIL Technology | custom-made | |
Upright snapshot fluorescence microscope | Olympus | BX51 | |
Microscope objectives | Olympus | 10x and 60x UPlan FLN ∞/-/FN 26.5 | |
Upright bright field microscope | Leica | DM 2500M | |
Ultrasonicator | Ultrawave Ltd. | U95 | |
Spreadsheet for recording and intepreting automated alignment results | Microsoft | Excel | |
Reagent | |||
2-propanol | Sigma-Aldrich | 34863 | FLAMMABLE |
Microscope Sildes, Clear, Ground | Thermo Fisher Scientific | 451000 | |
(7–8% vinylmethylsiloxane)-dimethylsiloxane copolymer, trimethylsiloxy-terminated | Gelest | VDT-731 | |
1,3,5,7-tetramethyl-1,3,5,7-tetravinylcyclotetrasiloxane | Gelest | SIT7900.0 | |
Platinum(0)-1,3-divinyl-1,1,3,3-tetramethyldisiloxane complex solution | Sigma-Aldrich | 479527 | HARMFUL, TOXIC |
(25–35% methylhydrosiloxane)-dimethylsiloxane copolymer, trimethylsiloxane-terminated | Gelest | HMS-301 | |
Weigh Boat 100 mL | Scientific Laboratory Supplies | BALI828 | |
Pasteur pipette | Appleton Woods | KS230 | |
Petri dish | SARSTEDT | 82.1473 | |
Razor blade | Thermo Fisher Scientific | ST10-031T | |
Adhesive Carbon Tape | Agar scientific | AGG3939 | |
16-Mercaptohexadecanoic acid | Sigma-Aldrich | 448303-1G | HARMFUL, TOXIC |
Ethanol | Sigma-Aldrich | 34852 | FLAMMABLE |
Gold coated microscope slide | Sigma-Aldrich | 643203 | Once opened gold will remain reactive to thiols for at least 1 month |
Thiourea | Sigma-Aldrich | T8656 | HARMFUL, TOXIC |
Iron(III) nitrate nonahydrate | Sigma-Aldrich | 529303 | HARMFUL, TOXIC |
Hydrochloric acid | Sigma-Aldrich | 84415 | HARMFUL, TOXIC |
(11-Mercaptoundecyl)hexa(ethylene glycol) | Sigma-Aldrich | 675105 | HARMFUL, TOXIC |
Fibronectin from human plasma | Sigma-Aldrich | F0895 | |
Cobalt(II) nitrate hexahydrate | Sigma-Aldrich | 203106 | HARMFUL, TOXIC |
Dulbecco’s Phosphate Buffered Saline | Sigma-Aldrich | D8537 | |
MSCGM Mesenchymal Stem Cell Growth Medium | Lonza UK | PT-3001 | |
Human Mesenchymal Stem Cells | Lonza UK | PT-2501 | |
Trypsin-EDTA | Sigma-Aldrich | T4174 | |
Heraeus Multifuge X1 Centrifuge | Thermo Fisher Scientific | 75004210 | |
CELLSTAR Centrifuge Tubes | Greiner Bio-One | 188261 | |
Paraformaldehyde | Fisher Scientific | P/0840/53 | HARMFUL, TOXIC |
Alexa Fluor 488 Phalloidin | Thermo Fisher Scientific | A12379 | |
Triton X-100 | Sigma-Aldrich | T8787 | "Detergent" in manuscript |
VECTASHIELD Antifade Mounting Medium with DAPI | Vector Laboratories | H-1200 | |
Rabbit anti-fibronectin antibody | Abcam | ab2413 | |
Goat anti-Rabbit IgG (H+L) Secondary Antibody, Alexa Fluor 594 | Thermo Fisher Scientific | R37117 | |
Bovine Serum Albumin | Sigma-Aldrich | A3912 | |
12-well plate | Thermo Fisher Scientific | 10253041 | |
T75 tissue culture flask | Thermo Fisher Scientific | 10790113 | |
cantilever | BudgetSensor | ContAl-G |