Creating 3D Objects with Integrated Electronics via Multiphoton Fabrication In Vitro and In Vivo

3D objects with integrated electronics are produced using an additive manufacturing approach relying on multiphoton fabrication (direct laser writing, (DLW)). Conducting polymer‐based structures (with micrometer‐millimeter scale features) are printed within exemplar matrices, including an elastomer (polydimethylsiloxane, (PDMS)) have been widely investigated for biomedical applications. The fidelity of the printing process in PDMS is assessed by optical coherence tomography, and the conducting polymer structures are demonstrated to be capable of stimulating mouse brain tissue in vitro. Furthermore, the applicability of the approach to printing structures in vivo is demonstrated in live nematodes (Caenorhabditis elegans). These results highlight the potential for such additive manufacturing approaches to produce next‐generation advanced material technologies, notably integrated electronics for technical and medical applications (e.g., human‐computer interfaces).

demonstrated to be capable of stimulating mouse brain tissue in vitro. Furthermore, the applicability of the approach to printing structures in vivo was demonstrated in live nematodes (Caenorhabditis elegans). These results highlight the potential for such additive manufacturing approaches to produce next-generation advanced material technologies, notably integrated electronics for technical and medical applications (e.g., human-computer interfaces).

Introduction.
Advances in the manufacturing and miniaturization of electronics and components thereof (computers, microprocessors, transistors, etc.) has revolutionised our lives with the ubiquity of electronic devices in our daily lives, and underpins the economic success of countries across the world. 1 Electronic technologies employ conductors/semiconductors to fulfil specific roles within manufactured devices, and for a variety of reasons organic conductors and semiconductors (e.g., derivatives of carbon nanotubes, graphene, conjugated polymers, etc.) are playing an increasingly important role in these devices (e.g., in flexible/printable electronics, electronic interfaces for the body, etc.). [2][3][4][5][6][7][8][9][10][11][12][13][14][15][16][17][18][19] Integrated circuits used in electronics worldwide (e.g., for applications including, but not limited to, amplifiers, logic units, sensors, etc.) are typically mass produced in a layer-bylayer approach. 1 The manufacture of 3D objects with integrated electronics has become an area of intense research interest with a view to the development of flexible electronics. 3,7,11,20,21 22 There are a number of FDA-approved medical devices capable of electrical stimulation within the body, including cardiac pacemakers, bionic eyes, bionic ears and electrodes for deep brain stimulation; all of which are designed for long-term implantation (via a technically challenging surgical procedure). 3 Conducting polymers (e.g., polyaniline, polypyrrole (PPY), poly(3,4-ethylenedioxythiophene) PEDOT) can electrically stimulating cells in vitro, and have proven well-tolerated when implanted into small mammals (e.g., mice, rats and rabbits).
Their immunogenicity profile is comparable to FDA-approved non-conductive polymers such as poly(lactic-co-glycolic acid) (PLGA), supporting their safety in vivo; these preclinical studies suggest that conducting polymer-based biomaterials are promising for eventual clinical translation. 23 Furthermore, the tunable properties of conducting polymers (CPs, e.g., derivatives of polyaniline, polypyrrole, polythiophene) make them versatile components of electronic devices. 24 Various methods can be used for CP preparation (including solution phase synthesis, solid phase synthesis, electropolymerization, vapor deposition or photopolymerization), offering opportunities for inclusion in most standard electronic device manufacturing processes. 25,26 There are a number of approaches to prepare flexible bioelectronics, 3,11,[27][28][29][30] often involving layer-by-layer processing, 31 however, novel photochemical techniques are under development (e.g., for ion conductive hydrogels). 32 Such approaches are effective routes to functional electronic devices, the ability to prepare electronics with de-novo designed architectures via printing is appealing for technical and medical applications. It is possible to employ additive manufacturing (AM) techniques to produce components for electronic applications, 33,34 for example printing CP-based materials using various methods, including: extrusion, inkjet printing, photopolymerization, rotary printing, screen printing, etc.. 35 Multiphoton fabrication is an AM approach that potentially allows the manufacture of bespoke architectures with features on various length scales (i.e., nm/μm to mm scale) either free standing (e.g., on glass) or embedded within a matrix of another substance (e.g., in Nafion® sheets), useful for production of integrated circuits 36,37 within the complex geometry of 3D printed parts and addressing limitations in applications where a high level of customization is required. [38][39][40] Herein, the concept was applied to printing conducting polymer (PPY)-based structures 41 within insulators (e.g., PDMS and shape memory polymers [SMPs] [42][43][44][45] in vitro and in vivo in transparent nematode worms (C. elegans). The functionality of the structures for biomedical applications was exemplified by using the conducting polymer-based structures embedded in PDMS to stimulate electrical activity in nerve tissue (an in vitro brain tissue paradigm). Such 3D printed electronics may facilitate fundamental studies (in vitro and in vivo) of the nervous system and its connectivity (e.g., enabling precise, long-term and continuous monitoring of patients over their lifetimes); or indeed the production of bioelectronic devices capable of continuous monitoring and modulation of neural activity. A particularly exciting aspect of the 3D printed electrodes is the potential to tailor electrode array designs specific to patients and their needs. Integration with artificial intelligence and machine learning approaches [46][47][48][49] for the development and operation of smart neuromodulation systems and/or human computer interfaces (potentially also useful for the gaming and virtual reality industries), would further support the transition from Industry 4.0 (technology-driven manufacturing) to Industry 5.0 (human-centric design and resilient/sustainable bespoke manufacturing). [50][51][52] 2.0. Results and Discussion.

Additive manufacturing of conducting polymer-based electronics integrated in 3D objects in vitro.
A variety of computational approaches (with different length and time scales) can be applied to study materials and facilitate the development/production of advanced functional materials for a broad spectrum of technical and medical applications. 53, 54 The integration of computational materials engineering approaches in workflows applies the Materials Genome Initiative concept for accelerating the discovery, manufacture and deployment of advanced materials which underpin millions of jobs worldwide in an area of high economic growth. 55,56 We envision in silico approaches supporting the additive manufacturing of advanced functional materials for bioelectronic applications (e.g., in ink formulation, additive manufacturing process optimization, etc.). Understanding the cytocompatibility/biocompatibility of materials is important when contemplating their potential for various applications and their end-of-life. [57][58][59] In silico toxicity screening has been developed to predict negative outcomes in various organisms (mammals, humans, etc.) and in the environment if exposed to molecules (e.g., those being developed for agriculture/healthcare markets); 60-66 the large datasets offer a more reliable/robust method of assessing toxicity than individual measures such as the median lethal dose (LD50) 67 which are prone to variations between testing factors (administration method, environmental factors, genetics, species, etc.), 68 and moreover, conform to the most important principles of processes involving animals in ethically sound research and development (i.e., replacement, reduction and refinement, the 3Rs). 61,[69][70][71] We have previously employed Derek Nexus and Sarah Nexus (Derek Nexus is an expert rule-based system to identify structural alerts for several endpoints and Sarah Nexus is a statistical-based model focused on mutagenicity only) to assess the biocompatibility of biomaterials, [72][73][74] including PDMS, 75 which is popular in biomedical applications (e.g., coatings of cochlear implants) due to its flexibility and transparency. 7,[76][77][78] However, PDMS contains ether/organosilicon bonds which may be hepatotoxic/nephrotoxic, [79][80][81] it may degrade 82,83 and its surface chemistry may need to be tuned to minimize biofouling; 84-86 pyrrole (acknowledged in supplier's safety data sheets [SDSs], which are of variable quality, to display a degree of toxicity, with significant variation in LD50 between species and mode of administration), the photoinitiator (Irgacure D2959; SDSs indicating it to be somewhat toxic), and PPY (non-hazardous in supplier's SDSs), which were predicted to be non-sensitizers of skin, and non-mutagenic. 73 it to be non-hazardous) and polyethyleneglycoldimethyacrylate (PEGDMA, 2 kDa; SDSs variable, often indicated to be an eye irritant, skin sensitizer and toxic), depicted in Figure 1).
In silico toxicity screening studies of the ink components (CSA, and PEG, Table S1) using Derek Nexus (Derek Nexus: 6.0.1, Nexus: 2.2.2) predicted them to be non-sensitizers of skin, and in silico mutagenicity screening studies using Sarah Nexus (Sarah Nexus: 3.0.0, Sarah Model: 2.0) predicted them to be non-mutagenic; by contrast, PEGDMA is predicted to plausibly cause chromosome damage, cause irritation of eyes/skin, be a sensitizer of skin (albeit non-mutagenic). In the case of the PPY electronics integrated in PDMS films, it is possible to contemplate their use as conformable bioelectrodes in vivo (however, the potential for slow degradation of PDMS 82,83 means they may need to be removed after some time in vivo, pending lifetime assessments), or indeed as bioelectrodes for in vitro studies.    Field stimulation experiments focusing on a single polymer electrode and its operational environment were simulated using a EM field solver in ANSYS Electronics Desktop 2020 R2 using Maxwell 3D with Electric Transient solution type (see Figure 4). Here, the tissue was simplified to be a homogeneous material with a conductivity of 0.33 S/m, which is similar to that of mouse brain tissue. 88 These simulations show how the electric field and current density is expected to evolve as the stimulation is applied if there were no cellular activity within the tissue and can be used to optimize experimentation. Simulations show that electric fields established initially between the polymer electrode and glass field electrode upon stimulation result in a flow of charge towards the glass field electrode, as expected. This acts to reduce the electric field gradient within the tissue in accordance with Gauss' law reaching a steady state (under constant voltage excitation) within ca. 5 µs. and current density (vector plot) as stimulation is initially applied.
To demonstrate that the conducting polymer electronics integrated in PDMS can be used as neural interfaces, we used the electrodes to stimulate a slice of mouse brain in vitro. The electrodes were positioned to stimulate the Schaffer collaterals in the stratum radiatum and a single CA1 pyramidal neuron was patch clamped using standard methodology, permitting CA3-CA1 synapses to be recorded ( Figure 5). A square potential step of 10 V was applied for 80 μs to the PDMS electrode. While the stimulus artifact was wider than that typically obtained with a conventional glass stimulating electrode, a corresponding physiological response was evoked by the PDMS electrode ( Figure 5, left trace), indicating that the electrodes interact with the nervous system. Importantly, typical CA3-CA1 synaptic properties were observed, whereby application of two stimuli at a 50 ms inter-pulse interval resulted in paired-pulse facilitation i.e., the second response was larger than the first, reflective of the low initial probability of presynaptic release of the neurotransmitter glutamate at these synapses. Moreover, as would be expected for excitatory currents in the central nervous system, a competitive antagonist of the postsynaptic α-amino-3-hydroxy-5methyl-4-isoxazolepropionic acid (AMPA)/kainate class of glutamate receptors, 6-cyano-7nitroquinoxaline-2,3-dione (CNQX), completely abolished the synaptic response, leaving only the stimulus artifact ( Figure 5, right trace), indicating that the evoked responses are physiological and indeed synaptic in nature. Printed electronics integrated in flexible substrates such as those described herein have significant potential for technical applications (e.g., display technologies) and medical applications (e.g., patient specific implantable electrodes for interaction with the central/peripheral nervous system). Printing conducting polymer structures integrated in shape-memory polymer-based materials may facilitate the development of switches, nerve cuff electrodes, etc., 42,[89][90][91][92][93][94] and proof-of-concept it would be possible to realize such applications was demonstrated by printing PPY structures in/on thin films of an optically transparent SMP (shape-memory polyimide, Figure S6). 95

Additive manufacturing of conducting polymer-based electronics integrated in 3D objects in vivo.
Researchers have previously generated abiotic conducting polymers using electropolymerization in the vicinity of live cells, 96 or indeed oxidative enzymes present in plants [97][98][99] and invertebrate Hydra, 100 in analogy to the production of natural melanins in a variety of organisms. 101 Multiphoton fabrication has been used to print free-standing conducting polymer structures used as an interface for mouse brain slices in vitro, 41 and to print non-conducting polymer-based hydrogels in the vicinity of live cells in vitro, 102 and C. elegans, 103 observing relatively low levels of cytotoxicity over the short timeframe of the experiments. The direct printing of conducting polymer structures directly on/in living organisms would enable real-time repairs of implanted bioelectronic devices and other applications (e.g., miniaturization/customization, precisely controlled reconfiguration of the electronics), 104 however, it has not yet been reported in the literature. 105 To facilitate proof-ofconcept that such a technological leap is within reach and it would be possible to realize such applications, we applied multiphoton fabrication to print PPY-based structures on/in live C. elegans (Figure 6), which complements reports on printing of non-conductive polymer structures employing near infrared (NIR) light sources. 106,107 We chose C. elegans for ethical reasons, but also for its high sensitivity to heat, desiccation, and physical injury, making it an ideal testing ground for biosafe laser-based in vivo printing approaches in biomedicine.
Achieving laser-printing on/in live C. elegans would require the lowest possible laser power that enables ink polymerization, and biocompatible ink components. The PPY-based formulations described earlier were thus first evaluated for C. elegans toxicity in order to design a biocompatible ink. While the photo-polymerized ink is inert and non-toxic, future biomedical applications would involve tissue exposure to the unpolymerized mixture prior to in vivo printing. It was thus necessary to determine the toxicity of each individual ink component in solution. Two toxicity assays were performed to measure (1) acute adult toxicity and (2) chronic developmental toxicity, for various concentrations of each compound.
The former involves exposing adult worms to compounds for 24-48h and relies on the labelfree automated survival scoring (LFASS) technique, 108 which exploits the fact that worms fluoresce in blue when they die to pinpoint median time of death. 109 The latter assesses the timing and duration of C. elegans successive larval stages and ability to reach reproductive age, using a transgenic strain that produces bioluminescence when the worm is metabolically active. 110 As worms progress through the four larval stages, they feed at increasing rates (commensurate with their size) and produce more bioluminescence. Between larval stages, worms undergo moults during which they cease feeding and appear metabolically quiescent, giving out little bioluminescence. Time-lapse recording of bioluminescence thus enables timing and measurement of developmental stages as bioluminescence rises and falls.
Acute adult toxicity assays revealed that all ink components are acutely toxic at concentrations of 6, 8 and 10 mg/mL but not below 3 mg/mL (Figure S7), while only CSA remained acutely toxic at 4 mg/mL. Conversely, neither of the compounds showed any strong developmental toxicity across the range of concentrations tested (10 μg/mL down to 156 ng/mL), as they all allowed worms to reach adulthood in a timely manner ( Figure S8); although higher doses (2.5-10 μg/mL) of Irgacure, Pyrrole and PEG led to modest developmental shifts ( Figure S9). These results indicated that the PPY ink components in solution were compatible with in vivo printing when employed at concentrations below 4 mg/mL. In particular, HA revealed more biocompatible than CSA and was thus chosen as photoinitiator in the subsequent phases of ink formulation refinement.
Next, to dilute ink component down to biocompatible concentrations, and because C. elegans would not naturally consume the ink alone, ink formulations were mixed with dietary E. coli OP50 bacterial paste. As ink dilution into bacterial paste is expected to impact printing performance, several ink to bacterial paste ratios were tested to determine the lowest ratio compatible with live laser-printing. Ink to bacterial paste ratios at 1:1 to 1:10 were tested first on polydimethylsiloxane (PDMS)-coated coverslips and the fidelity of intended printed structures was assessed ( Figure S10). While sub-millimetric structures could be printed at 1:10 ink to bacterial paste ratios (Figure S10A-C), resolving 10-30-micron scale structures was only fully achieved with pure ink formulations (Figure S10D-F).
C. elegans were then exposed to HA-based ink formulations mixed with dietary E. coli OP50 bacterial paste at 1:5 ratios as a compromise between biocompatibility and printing resolution.
Lower-energy infrared two-photon 3D printing was chosen to reduce phototoxicity while enabling deeper tissue penetration (Figure S11). Two formulations were tested with subtoxic (3.3 mg/mL, Figure S11C) or mildly toxic (6.6 mg/mL Figure S11A-B, S11D) Irgacure concentrations. 6-10 μm size square and star shapes were then printed directly onto the skin and within the gut of live C. elegans roundworms (Figure 6). Thanks to the autofluorescence properties of the ink mix, the printed shapes were imaged and localized by confocal fluorescence imaging, demonstrating accurate and well-tolerated printing of polymer on live worms (Figure 6 and S11). Printing within the moving gut of the worm did not yield the intended shape ( Figure 6C). Faster printing could resolve this limitation, which may be achieved by improving the ink formulation photo-curing efficacy, and/or increasing laser power. However, as light propagates within a complex environment (here the body of the worm), printing accuracy and precision also decreases. Corrective strategies employing adaptive optics will thus likely be necessary to circumvent the issue when translating the approach to thicker vertebrate/human tissues; nevertheless, this represents a technological leap from examples of printing non-conductive structures in vivo. 106,107  Dotted circles indicate locations of printed structures. Asterisks mark the position of the terminal oocyte. Animals were alive following 3D printing. Scale bars represent 10 µm.

Ethics.
The research described in this paper has carefully negotiated the ethically sensitive aspects of the experimentation it entailed. For example, C. elegans was chosen as a model organism for the in vivo experimentation because it does not need ethical approval to use. In addition, the toxicity of the compounds used in the 3D printing process occur on such a small scale that there is a negligible risk of harm to the researchers or the environment. One potential concern is that as future stages of this research progress, there may be a growing scientific case to experiment on more complex organisms (e.g., mice) in vivo, and eventually in humans. Given the novel nature of this technology, some may find the thought of first-in-mammal or first-inhuman research to be ethically unsettling. However, if this research did reach more advanced stages that required more complex organisms for experimentation, this process would necessarily follow all the typical required safety, ethical and legal protocols. This would be no different than the development of a novel medical device or an analogous scientific procedure.
While the research carried out in this study is not particularly ethically contentious, there is nevertheless a need to be aware of the potential for 'dual-use dilemmas' to emerge as research progresses and becomes more ethically complex. A dual-use dilemma is an ethical dilemma that occurs when research is undertaken with a beneficial use in mind, however, the researchers also foresee that other users may employ this research in ways that could do harm. 111 When it comes to the scientific research outlined in this paper, the authors do not yet feel that the research has reached a point where it could be deployed in ways that could cause harm. However, if this research maintains a successful trajectory, the potential applications it may have at later stages will grow, and some of these applications may carry the risk of harm if misused. For example, human-computer interfaces could be used to beneficially treat medical patients with neurological conditions; however, such technology could also be used by a bad actor in such a way (e.g., hacking into them to control or obtain information from the device) that negatively affects the privacy and autonomy of the individuals using them. [112][113][114] One way to responsibly negotiate future dual-use dilemmas is to take steps to try and identify them in advance and subsequently have researchers work with regulators to creatively design ethical safeguards that can be engineered into and alongside the development of the technologies. For example, one safeguarding procedure may involve keeping key aspects of research knowledge secure (e.g., by withholding it) that would otherwise enable the harmful use of such research. Insofar as possible, researchers will need to endeavour to continue to engage in horizon scanning in relation to future stages of this research in order to identify and address ethical dilemmas in advance.

Conclusion.
Here we report the application of a multiphoton fabrication process to create 3D objects with integrated electronics: in silico toxicity screening of ink components identifies/confirms likely cytocompatible formulations; 3D printing through light-transmitting materials yields wellresolved conductive micron-scale features; stimulation of 3D printed PPY structures interfaced with live brain tissue can induce specific synaptic responses; and it is possible to 3D print PPY structures directly in vivo.
We showcase a range of examples made by this technology-driven manufacturing (Industry 4.0) process, including PDMS films and a living organism (C. elegans); and highlight potential applications such as electrodes capable of stimulating nerve tissue. We foresee significant potential for this technique's integration in human-centric design and bespoke manufacturing (Industry 5.0) processes of producing bioelectronics for telemedicine, and that this represents a possible roadmap for the broader development of direct-printing of electronic devices for biomedical applications in situ.
Computational approaches (e.g., including in silico toxicity screening, multiscale modelling of electrical/mechanical/physicochemical properties, etc.) will facilitate ink formulation development for the production of 3D objects with integrated electronics, offering insights into the properties (e.g., electrical conductivity) of the conjugated polymers which underpins their function in devices. Simulation of the interactions between oligomers of components in the bioelectronic devices (in this case, PDMS, PEG and PPY in water, mimicking the hydrated state of the bioelectrodes if used in vitro or in vivo) would offer insight into aggregation/clustering due to intramolecular/intermolecular interactions, and potentially phase separation of polymer phases (e.g., associative phase separation). In the long term we believe that such an approach may offer insights that accelerate the discovery, manufacture and deployment of complex composites used in advanced materials technologies generated by additive manufacturing approaches, for example, component selection and composition tuning to achieve optimal properties and device performances. conditions that increase with ageing (e.g., epilepsy, strokes), as well as rises in traumatic brain injury. In this context, potential applications of the approach described herein may include: improved electrodes (smaller, lower morbidity, better tolerated etc.) for deep brain stimulation (e.g., treating Parkinson's disease, epilepsy, etc.), improved monitoring/diagnosis (e.g., advanced epilepsy work-up to identify epileptogenic focus), novel neuroprosthetics (e.g., for traumatic brain injury), new approaches to monitoring/treatment of (peripheral) neuromuscular disorders (e.g., degenerative muscular and peripheral nerve conditions), novel approaches to neuromodulation for pain. In short, we foresee the technique described herein as having significant potential for technical and medical applications, with potential economic, environmental, health and societal impacts.  Table S2).
In silico toxicity screening: In silico toxicity screening was carried out using Derek Nexus (v. The holder was then flipped so that the PDMS/resist face is upwards and oil face is downwards. A drop of one of the clear ink stock solutions was placed on the substrate and left for 5 minutes to infiltrate the PDMS. The structures to be printed were designed using the computer aided design (CAD) package (Fusion360 from AutoCAD:Autodesk), the structures were exported to the Nanoscribe ® software (DeScribe) to do the scripting. The Nanoscribe ® Photonic Professional GT 700 instrument was equipped with a light source (Topica FemtoFiber pro, Er-doped fiber laser of wavelength 780 nm, pulse duration <150 fs, repetition rate 100 MHz, an aperture of 7.3 mm, a diode voltage of 1.34-5, and 50 mW power at the focus point at 100% power). The Nanoscribe ® was controlled by the Nanowrite software (version 1.8.14) and the camera software within the Nanoscribe ® is AxioVision LE (version 4.8.2.0). The structures were printed on PDMS-coated glass slides using Galvo writing mode by moving beam fixed stage (MBFS) -a fast layer-by-layer writing approach, suitable for micrometre to millimetre scale structures with a computer-controlled piezoelectric scanning stage range 300 x 300 x 300 µm 3 . Positioning was achieved using a computer-controlled motor stage, range 100 x 100 mm 2 . Laser power and speed for printing on PDMS-coated glass substrates the laser power was ca. 50-60 % and the speed was ca. 5000 µm/s. After printing structures, the substrates were washed with water-ethanol and water to remove low molecular weight contaminants, and the structures were dried under N2.
Electrical measurements: Direct electrical characterization of the structures was performed using a Keithley 2602B source measure unit, connected to a Wentworth Laboratories SPM197 probe station, with tungsten probes with a 1μm tip diameter. Each wire was swept in a range from 0V to +1V DC. 900 x 40 x 8 μm CSA-doped conductive wires were fabricated within PDMS using 2PP, with circular pads (75 μm diameter) to aid electrical measurements. The electrical properties of the wires were measured with a source-measure unit connected to a probe station. The wires exhibited linear conductance. Their resistance was calculated to be RPPy = 570 ± 50 kΩ. Modelling of the structure, we can estimate the resistivity of the wires using: Where R is the measured resistance of the structure, l is the length of the resistor, and A is it's cross-sectional area. The doped PPy structures had σPPy = 3.9 ± 0.3 S/m.

Scanning electron microscopy (SEM) and energy dispersive X-ray spectroscopy (EDX):
Samples were mounted on stubs and coated with a layer of gold (ca. 5 nm, 60 s, 20 mA, 8 × 10 -2 mBar) using a Quorum Q150RES sputter coater (Quorum Technologies Ltd., Lewes, UK). Images were obtained using a JEOL JSM-7800F (Field Emission SEM, FE-SEM) fitted with an EDX system (X-Max50, Oxford Instruments, Abingdon, UK) at 10 mm working distance and 10 kV voltage, three measurements were performed per sample and average results are presented.
Fourier transform infrared (FTIR) spectroscopy: Infrared spectroscopy was carried out on Agilent Cary 630 FTIR Spectrometer. Spectra were recorded in ATR mode, with a 1 cm -1 resolution and 64 scans (corrected for background and atmosphere using OMNIC software supplied with the spectrometer).

Optical coherence tomography (OCT): A line-field optical coherence tomography (LF-OCT)
system was employed to measure the electrode sample. The LF-OCT can acquire crosssectional images (B-scan) in a single shot, 115 which can significantly reduce the motioninduced image distortion and artefacts. The resolution of our system is 8µm axially and 17.5 µm laterally, which enables high-resolution tomography and topography imaging. Figure S1 shows the LF-OCT system setup, which uses a superluminescent diode (SLED) source to  (10 mol % concentration relative to pyrrole)) with laser writing powers of ca. 80-100 %, followed by ethanol washing, yielding structures depicted in Figure S6. spin-coated at 700 rpm on 12-inch vinyl record chunks (typically 2-3 inches wide) for 1 min, baked at 60°C for 3h, cut in 1 cm 2 slabs, and bound to 22 mm and 30 mm diameter coverslips using oxygen plasma activation of silicon surfaces. Coated coverslips were stored in a dust-free environment until use. Before use, they were briefly rinsed with isopropanol and dried with an air gun.
3D printing in vivo in C. elegans: Wildtype C. elegans Bristol N2 worms were maintained as previously described. 121  and M9 medium, between two 1 cm 2 slabs of PDMS grooves bonded onto a 22 mm-diameter and a 30 mm-diameter coverslips (PDMS film were prepared by spin-coating at 700 rpm on vinyl records for 1 min, baked at 60°C for 3h, cut in 1 cm 2 slabs, and bound to coverslips by plasma-cleaning). During worm mounting, worms were aligned inside the grooves using a platinum pick, and vacuum grease was used to seal the edges of the coverslips. Coverslips were then mounted into the Nanoscribe ® holder (Herma glue was dotted around the edges of the coverslip binding it to the holder and left to dry for 10 minutes) and a drop of immersion oil (immersol 518F) was applied. The sample was then processed for 3D nano-printing through a Zeiss 63X 1.4NA lens from a Nanoscribe ® Photonic Professional GT 700 instrument operating in galvo mode at speeds of 500-1,000 μm/s with a laser power output of 40 -60 mW (Topica FemtoFiber pro Er-doped fiber laser with a pulse duration <150 fs, a repetition rate of 100MHz, a wavelength of 780 nm, an aperture of 7.3 mm and a diode voltage of 1.34-5).
Confocal imaging of C. elegans: Following 3D printing, worms were collected by transferring the coverslips onto a microscopy slide with M9 added to prevent desiccation. The slides were then imaged through a Zeiss 40X 1.4NA oil immersion lens with a Zeiss LSM880 confocal microscope exciting the sample at 488nm and collecting fluorescence signal at 543nm and operated by the Zen Black software. Z-stacks were acquired at 1024x1024 pixel 2 every 0.45 µm across the first half of the worm thickness (about 30 µm). single plane sequences were acquired at maximum speed (1.5 s per frame after averaging). Images were then exported and processed with FiJi for figure preparation. 122

Supporting Information
Supporting Information is available from the Wiley Online Library or from the author.