Controlled growth of a single carbon nanotube on an AFM probe

The study addresses the need for accurate, nondestructive 3D morphological characterization of complex, high–aspect-ratio nanostructures using AFM. Conventional silicon AFM probes, with pyramidal or conical tips, suffer from limited lateral resolution on steep features and poor wear resistance. Carbon nanotubes, with their small diameter, high aspect ratio, and durability, are promising as AFM tips. Existing CNT-on-AFM fabrication methods are either complex or poorly controlled in placing a single CNT of suitable length and orientation at the apex. The authors aim to develop a simple, controllable process to selectively grow a single CNT directly on an AFM probe tip, enabling improved imaging of complex nanostructures without post-growth cutting.

Two main approaches exist: (1) manual assembly, where CNTs are transferred to tips under an optical microscope and bonded with adhesives, often resulting in bundles, variable length/straightness, and low reliability; multistep transfers monitored by SEM are time-consuming and can still yield variability. (2) Direct CVD growth offers stronger bonding. Pore growth on flattened tips with anodized nanopores and electrodeposited catalysts can grow CNTs, but placement on the optimal apex is difficult, perpendicularity is hard to achieve, and lengths often require post-growth cutting. Catalyzed CVD on arrays can anchor CNT networks but lacks precise catalyst placement and often produces looped tips. Prior methods therefore struggle to reproducibly place a single, well-oriented CNT of suitable length at the exact tip apex.

The authors propose a two-step pick-up and growth method using standard AFM control to regulate the amount of catalyst-containing growth solution adhered to the tip via the trigger threshold (cantilever bending/deflection). Preparation: a growth solution (inorganic chlorides AlCl3·6H2O, SiCl4, FeCl3·6H2O; triblock copolymer; alcohol) is dripped onto an oxygen plasma-treated mica or silicon substrate. The AFM probe approaches the droplet; the trigger threshold sets the cantilever deflection and thus the immersion depth of the tip into the droplet. Force–distance (pull-off) curves quantify immersion depth: the onset of immersion (P) to maximum immersion (Q) distance defines depth. Example measurements show immersion depths of 56 ± 0.7 nm at 0.25 V and 98 ± 0.40 nm at 0.55 V trigger thresholds. Retract curves exhibit increased interaction forces at higher thresholds, indicating larger contact area and more solution pickup. After contact, the adhered solution on the tip is allowed to solidify for 3–5 minutes. The probe is then placed in a CVD furnace for CNT synthesis (metal-catalyzed CVD; fixed growth conditions for solution type, temperature, and time; details referred to Methods). By varying only the trigger threshold, the amount and footprint of catalyst at the apex is controlled, enabling selective growth of a single CNT (for specific threshold ranges) versus multiple CNTs (at higher thresholds).

• Selective single-CNT growth is achieved by controlling the AFM trigger threshold; single CNTs grow reliably for thresholds 0.25–0.50 V, while 0.55 V tends to yield multiple CNTs. • Immersion depths measured via force curves: 56 ± 0.7 nm at 0.25 V and 98 ± 0.40 nm at 0.55 V. • Diameter of solidified growth solution footprint on tip: d0.25 = 128 ± 1.20 nm at 0.25 V; d0.55 = 224 ± 1.36 nm at 0.55 V; ratio d0.25/d0.55 = 0.571. • CNT characteristics (single-CNT regime 0.25–0.50 V): average length 589.17 nm with overall SD 27.60 nm; length increases by 16 ± 2.0 nm per 0.05 V increase in trigger threshold; average diameter 50 ± 1.53 nm. Multiwalled CNTs are formed. • Yield: A perpendicularity criterion of ±5° from tip axis defines success; yield of single-CNT probes meeting this is 93.33%. • Imaging performance: On a standard grating (true height 568 ± 2.6 nm, period 3 ± 0.01 µm), a conventional AFM probe measured 344 ± 2.09 nm, whereas the CNT probe measured 568 ± 0.8 nm, accurately capturing depth and steep curvature. • High–aspect–ratio nanoholes in fused silica: CNT probes produced clearer edges and profiles closer to the true morphology. Representative section measurements include AFM probe: X_AB = 438 ± 7.04 nm, Y_AB = 299 ± 13.6 nm; CNT probe: X_CD = 471 ± 10.96 nm, Y_CD = 388 ± 9.76 nm; additional positions X_A1B1 = 439 ± 1.60 nm, Y_A1B1 = 637 ± 1.36 nm; X_C1D1 = 465 ± 0.56 nm, Y_C1D1 = 723 ± 1.04 nm.

Regulating the AFM trigger threshold controls the immersion depth and thus the quantity and footprint of catalyst-bearing growth solution adhered to the tip, enabling deterministic placement and growth of a single CNT at the apex without complex lithography or post-growth cutting. Within a defined threshold window (0.25–0.50 V), single multiwalled CNTs of consistent length and diameter are reproducibly grown with high yield and near-axial alignment, addressing key shortcomings of prior manual assembly and pore-growth CVD methods. The strong tip–CNT bonding from direct CVD and the controlled, sub-micrometer CNT length translate to improved AFM imaging of steep, high–aspect–ratio nanostructures, as evidenced by accurate depth recovery on standard gratings and clearer, more faithful profiles of nanoholes compared to conventional probes.

A simple, two-step AFM-controlled pick-up and CVD growth process enables the selective fabrication of a single, well-aligned CNT at the apex of standard AFM probes by tuning the trigger threshold to control catalyst pickup. The method yields multiwalled CNT tips with consistent sub-micrometer length, obviating post-growth cutting, and achieves a high success rate (>93%). The resulting CNT probes substantially improve imaging of complex, high–aspect–ratio nanostructures and accurately recover true depths that conventional probes underestimate.

The pickup process is stochastic at the catalyst particle level; while length variability is small at fixed thresholds, the precise catalyst particle distribution cannot be completely controlled. Droplet thickness can vary between depositions, necessitating force-curve control to standardize immersion depth. Higher trigger thresholds (e.g., 0.55 V) tend to produce multiple CNTs rather than a single CNT. Very long CNTs (>1 µm) may introduce vibration during measurement (though the proposed process targets shorter lengths to avoid post-growth cutting).