Disposable ultrasound-sensing chronic cranial window by soft nanoimprinting lithography

Abstract

Chronic cranial window (CCW) is an essential tool in enabling longitudinal imaging and manipulation of various brain activities in live animals. However, an active CCW capable of sensing the concealed in vivo environment while simultaneously providing longitudinal optical access to the brain is not currently available. Here we report a disposable ultrasound-sensing CCW (usCCW) featuring an integrated transparent nanophotonic ultrasonic detector fabricated using soft nanoimprint lithography process. We optimize the sensor design and the associated fabrication process to significantly improve detection sensitivity and reliability, which are critical for the intend longitudinal in vivo investigations. Surgically implanting the usCCW on the skull creates a self-contained environment, maintaining optical access while eliminating the need for external ultrasound coupling medium for photoacoustic imaging. Using this usCCW, we demonstrate photoacoustic microscopy of cortical vascular network in live mice over 28 days. This work establishes the foundation for integrating photoacoustic imaging with modern brain research.

Fig. 1

Process to fabricate usCCW: (a) Nanofabrication of MRR on transparent substrate using sNIL process; (b) assembling an MRR-based ultrasonic detector with matching optical fibers, onto a circular cover slide (8 mm in diameter); and (c) finally assembled usCCW device

Fig. 2

Fabrication process of the MRR ultrasonic detectors. The fabrication procedure consists of: (a) spin-coating of 200 nm photoresist; (b) patterning electron beam lithography process; (c) RIE pattern transfer from photo-resist into SiO₂ layer; (d) DRIE pattern transfer from SiO₂ layer into silicon substrate with high-aspect-ratio features; (e) removal of the SiO₂ layer; (f) casting the PDMS soft mold via thermal curing of the PDMS precursor; (g) peeling off PDMS soft mold from silicon master mold; (h) placing PDMS soft mold over a 400-nm-thick PS thin film spin-coated on a quartz coverslip; (i) molding process by heating the polystyrene film over its glass transition temperature; (j) after cooling down to room temperature, the PDMS stamp is peeled off; finally, (k) a 5-μm UV-curable PDMS thin film is spin-coated on the MRR waveguide and then cured by excessive UV exposure. The illustration of (l) silicon mold, (o) PDMS soft mold, and (r) fabricated MRR on quartz substrate and the magnified views shown as inset. The scanning electron microscope images of (m) silicon mold, (p) PDMS soft mold, and (s) MRR device. Their corresponding magnified image are shown as (n), (q), and (t), respectively. Scale bars: (m, p, s) 20 μm, (n, q, t) 2 μm

Fig. 3

Protecting MRR ultrasonic detector for long-term in vivo imaging. a Scanning electron microscope (SEM) image of a MRR ultrasonic detector fabricated by sNIL. Scale bar: 10 μm; (b) Magnified view of the waveguide coupling region highlighted by the orange dashed box in a. Scale bar: 500 nm. c Schematic and (d) SEM image of a cross-section of the fabricated MRR. The schematic illustrates potential contaminants attached to the waveguide. Scale bar: 500 nm. e Optical resonance diminished in 2 h due to contamination when unprotected MRR is exposed to whole blood. f Schematic and (g) SEM image of a cross-section of the MRR protected by an additional PDMS layer. Scale bar: 500 nm. The schematic illustrates that contamination to the waveguide is prevented. h Optical resonance remained unaffected after the protected MRR is exposed to whole blood for 2 h. i Comparison of Q-factors of MRRs with and without the protection layer in whole blood. Error bars show the standard deviation of Q-factor when fitting with the experimentally measured resonance spectrum

Fig. 4

In vivo PAM cortical imaging using a usCCW. a The usCCW is surgically implanted on the mouse skull after craniotomy. The inset shows the physical dimension of the usCCW with MRR and fibers attached, which is optically transparent, with a total thickness of 250 µm and a total weight of less than 1 g. The MRR ultrasonic detector is attached on an 8-mm diameter circular substrate and the sensing light is coupled through a pair of 30-cm flexible optical fiber. b Illustration of optical scanning through the usCCW. To excite the MRR resonance, a narrow-band continuous-wave tunable laser (New Focus, TLB-6712, wavelength from 765 nm to 781 nm) is coupled into the bus waveguide after passing through a fiber polarization controller, and collected by a multimode fiber on the other end of the bus waveguide. c Optical excitation and ultrasonic detection geometry along the cross section highlighted in b. The space between the MRR and the dura is 1 mm and is filled with 0.5% agarose gel. We seal the usCCW with dental cement to prevent infection and leakage. d Brightfield optical microscopy image of the cortical region through the MRR. e Depth-encoded maximum-intensity-projection (MIP) PAM image of the same area. The whole image is stitched from 9 acquisitions due to the limited laser-scanning field of view. f Three dimensional visualization of the vessel orientations and cortical curvacure. g PAM image of the hemorrhage area highlighted by the dashed box in d and e. h PAM B-scan image from the position highlight by the green dashed line in g, showing the hidden vessels beneath the hemorrhage area. i Visualizing vessels beneth the hemorrhage layer. Scale bars, (a–b) 0.5 mm and (g–i) 200 µm

Fig. 5

Longitudinal cortical PAM through the usCCW. a Photograph of a free-moving mouse in the breeding cage after the implantation. b Photograph of the mouse mounted under the microscope during the imaging. c Maginfied optical image shows the implanted transparent usCCW and the clear optical view of the cerebral cortax. d Q-factor shows marginal reduction from 4.2 × 10⁴ to 3.6 × 10⁴ over the 28-day-period. Error bars show the standard deviation of measured results. e Maximum amplitute projections of PAM images show cortical vasculature in the same area over a 28-day-period. Medium post-surgical bleeding occurs at Day 3 and the hemorrhage is gradually cleared in 4 days. Aggressive neovasularization is also clearly observed. The pairwised comparison of the measured cortical vasculature between (f) Day 0 and Day 2, (g) Day 0 and Day 14, and (h) Day 14 and Day 28. Superposition of two images rendered in distinct megenta and green colors is used to provide direct visualization of the neovascularization process. The overlapping regions are therefore rendered in white color