Fiber-based optrode with microstructured fiber tips for controlled light delivery in optogenetics

Objective. Optogenetic modulation of neuronal activity requires precise and flexible light delivery to deep brain regions. Flat cleaved optical fibers combined with electrodes are widely used in implantable optogenetic devices for light delivery and electrical monitoring of neural activity. However, the flat fiber tip geometry induces serious tissue damage upon insertion, and makes it difficult to adjust and control the spatial extent of illumination within the brain. With their strongly increased tissue-compatibility and the possibility of spatial illumination control, tapered fibers outperform cleaved fibers in targeted neural photo-stimulation. Approach. In this work, we describe our device concept, and present a novel approach for reproducible fabrication of tapered fiber tips via grinding. Furthermore, we characterize recording electrodes by commenting data obtained from electrochemical impedance spectroscopy (EIS). We also investigate the impact of different cone angles (14°, 30°, 60°, and 90°) on the illumination profile and optical throughput. Main results. We fabricated a fiber-based optrode with cone tip and two deposited electrodes. Custom grinding setup for fabrication of tapered fiber tips with various cone angles is developed as a part of our research. Microscope images showed very good optical quality of cone tips. The results of transmitted optical power measurements performed with integrating sphere suggest that, compared to the flat cleaved optical fiber, transmitted power decreases exponentially with cone angle reduction. Obtained emission profiles (as induced fluorescence in Rhodamine 6G water solution) indicate very strong effect of cone angle on shape and size of illumination volume. Results obtained from EIS show the effect of electrode size on its recording capability. Significance. Compared to optrodes with flat cleaved optical fiber, the demonstrated fiber-based optrode with cone tip allows controlled light delivery with reduced invasiveness. The possibility to fabricate reproducible fiber tips with various cone angles enables control of light delivery in optogenetic experiment. The results presented here give neuroscientists the possibility to choose the appropriate tissue-compatible cone geometry depending on their stimulation requirements.


Introduction
Over the past decade, integration of optics and genetics opened a new door for neuroprosthesis application and provided novel possibilities for neuroscientists to study Parkinson's disease [1], epilepsy [2], brain mapping [3], retinal prosthesis [4] and depression [5], in a way that cannot be achieved using conventional electrical stimulation and sensing. To achieve light delivery and simultaneous electrophysiological recording from the nervous system, scientists and engineers designed dual optical and electrical probes. The combination of optical and electrical elements in one single probe, enabling simultaneous optical control and electrical monitoring of neural activity, represents the device called optrode (opto-electrode).
Some of the earliest optrodes were composed of standard commercial optical fibers, connected to a single light source, and a separately implanted single recording electrode [6,7]. It took years of development to obtain more sophisticated optrode designs incorporating multiple light delivery and electrical recording channels [8][9][10][11][12][13]. Although optogenetic techniques have lately improved, neuroscience labs widely rely on optrodes with flat cleaved optical fibers [8,10,[14][15][16][17] Beside the fact that, upon insertion, flat cleaved fibers substantially damage tissue and induces glial activation around the device even long after the implantation [18][19][20], this widely used approach also encounters important limitations when it comes to illumination within the brain. Since the spatial extent of illumination within the brain tissue cannot be controlled and adjusted using flat cleaved fibers, a precise spatiotemporal control of neural processes, which is a main goal of optogenetics, is very difficult to achieve. Compared to flat cleaved fibers, that yield a spatially heterogeneous illumination in a relatively small and restricted brain volume near the fiber facet, optical fibers with tapered tips can be used for illumination of spatially restricted or bigger brain volumes [19]. In addition to being more compatible upon insertion than flat cleaved fibers, tapered fibers offer a great advantage for in-vivo control of neural activity in different animal models. Recently, the application of a tapered fiber based optrode provided a wide-volume illumination in the motor cortex and in the striatum of both free-moving and head restrained mice [19], as well as in the frontal eye field of non-human primates [21]. Aside from their application in optogenetics, tapered fibers are also used for in-vivo fluorescence detection in a field of fiber photometry because of their unique light collection features.
Fiber tips are usually tapered employing the heat and pull method or chemical etching procedures. The heat and pull method is mostly used to obtain very small smooth cone angles (∼ 4 o ), while chemical etching yields reproducible fiber tips with a variety of cone angles. Compared to chemical etching, fiber tips developed using heat and pull method have a lower throughput power due to simultaneous reduction of the fiber cladding and core along the tapered region. Although chemical etching allows obtaining highly reproducible taper geometries, the sensitivity of the tip shape to environmental influences during the etching leads to large non-smooth taper angles whose rough surface influences the reflection of the guided light in the tapered region [22,23]. In many optrode designs incorporating tapered optical fibers [24][25][26][27], fiber tips were fabricated employing chemical etching procedure with very little information about the effect of cone angles on optical throughput and light emission profiles.
Here we present our optrode design and propose a novel approach for reproducible fabrication of smooth tapered fiber tips with various cone angles using a custom grinding setup. We demonstrate and comment on emission profiles of multiple fiber tip geometries, imaged from the induced fluorescence in a non-scattering medium, and experimentally quantify the total optical throughput using an integrating sphere. Based on recorded beam profiles, we comment on the effect of cone angle on intensity distribution at various distances from tip end. Furthermore, we comment on electrode recording capabilities according to results obtained from electrochemical impedance spectroscopy (EIS).

Materials and methods
Optrode design and assembly, as well as detailed explanation of fabrication and characterization methods will be presented below. Starting from commercial optical fiber (FG200UEP ThorLabs, core diameter: 200 µm), optrode fabrication can be divided in three stages: development of tapered fiber tip, electrode deposition, and optrode assembly. Optical characterization includes transmission measurements, and investigation of light emission and beam profiles. EIS is used as a tool for investigation of optrode's recording capabilities.

Device design
Single tip fiber-based optrode, fabricated in this work, is designed to enable simultaneous light delivery and electrophysiological recording. It consists of tapered multimode optical fiber (FG200UEP ThorLabs, core diameter: 200 µm) with two equidistant gold electrodes deposited on fiber surface (figure 1). Since electrodes were deposited on a surface of bare optical fiber, biocompatible polytetrafluoroethylene (PTFE) heat shrink tube (PTFE Sub-Lite Wall®, Zeus Industrial Products, Inc.) with wall thickness of approximately 40 µm after curing was applied to increase mechanical stability. In addition, precise positioning of heat shrink tube enables control of exposed electrode portion, which affects its recording capability. The size of applied PTFE heat shrink tube depends on optrode length. Tube is applied such to cover most of deposited electrode leaving small electrode parts (close to tapered fiber tip, and on opposite optrode end) uncovered. FR4 (flame resistant) substrate (1.5 mm thick), covered with copper on one side, is cut in square pieces (10 × 10 mm). After drilling a small hole (300 µm diameter) in the middle of substrate, 100 µm wide and 500 µm deep groove is formed such to pass through hole center and divide copper layer into two separate parts. Optrode with heat shrink tube is positioned perpendicular to the surface of FR4 substrate, and moved through the hole until the copper covered substrate surface reaches the end of heat shrink tube near flat cleaved end (opposite to the cone tip). Optrode is then rotated such that position of deposited microelectrodes corresponds to the position of copper layers on FR4 substrate. When properly positioned, in respect to FR4 substrate, optrode part containing heat Figure 1. (a) Schematic representation of optrode consisting of tapered optical fiber with two deposited electrodes on fiber surface. After electrode deposition, PTFE heat shrink tube is applied to increase mechanical stability and enable control of exposed electrode area. Optrode assembly consists of establishment of electrical contacts with copper layer deposited on FR4 substrate fixed perpendicularly to longitudinal fiber axis. Light coupling is achieved by butt coupling between flat cleaved optrode end placed in ceramic ferrule and LED-coupled optical fiber using mating sleeve. (b) Photograph of fabricated fiber-based optrode. (c) Magnified microscope image of tapered optrode tip with cured heat shrink tube and exposed electrode area (exposed electrode is 485 × 60 µm 2 ). shrink tube is fixed to uncoated FR4 substrate surface using UV-curable epoxy. Next, conductive silver epoxy is applied on the interface between microelectrodes and copper layer on FR4 substrate. After curing of conductive silver epoxy, UV-curable epoxy is applied over the conductive silver epoxy, and ceramic ferrule is placed around bare part of optical fiber. UV exposure leads to fixture of ceramic ferrule to FR4 substrate. The rest of the bare fiber sticking out of the ceramic ferrule is then cut and polished. Buttcoupling between flat cleaved optrode end placed in ceramic ferrule and light emitting diode (LED)coupled patch fiber is achieved using mating sleeve.

Fiber tip processing
The custom-made setup for fiber tip formation by grinding is represented on figure 2(a). Our grinding approach employs a rotating optical fiber, which is fixed in a fiber rotator and mounted on an angular stage, and a horizontal grinding disc. In order to produce fiber tips with a rotational symmetrical cone, two motors are used to drive the grinding disc and the fiber rotator, respectively. We control the rotational speed of both motors independently. In our approach the rotational speed of the grinding disc is 40 rounds per minute, while the optical fiber rotates 20 rounds per minute. For lubrication purposes a drop of water is added periodically onto the grinding film. In order to achieve best surface quality, four polishing discs with decreasing grit sizes are used subsequently. After processing the fiber tip using silicon carbide film of 2500 and 4000 grit size, each for 10 min respectively, polishing disc with an aluminum oxide film of 1 µm and 0.3 µm grit size are used each for another 10 min respectively to achieve a smoother surface of the tapered fiber tip. Although, an increase in rotating speed of grinding disc and/or fiber rotator can accelerate fiber tip formation and save time, more importance needs to be given to a tip surface quality. For example, doubling the rotation speed of grinding disc (from 40 to 80 rpm), and reducing a processing time to half (from 10 to 5 min for each grinding disc) resulted in a fiber tip with noticeable surface roughness, which would affect its optical performance. The setup consists of multiple angular stages that can be easily exchanged. This allows for the adjustment of the angle between the fiber axis and the surface of grinding disc (polishing angle) and enables the control of the cone angle. The polishing angle (θ) determines the final cone angle (β): (1)

Electrode deposition
After tapered tip has been fabricated, pre-cleaning of the fiber surface is conducted with acetone using ultrasonic support and rinsed with deionized water. Next, plasma oven (Femto, version D, Diener electronic Plasma-Surface-Technology GmbH) is used to activate the surface and promote adhesion by removing organic contaminants. Oxygen plasma surface treatment is performed for 10 min under following conditions: pressure p = 0.3 mbar, power 90%, chamber temperature T = 100 • C. Adhesion is much improved after plasma surface treatment, resulting in a homogeneous thin film without delamination. Following plasma surface treatment, optical fiber with cone tip is placed in a rectangular groove of custom fiber holder, and a 50 µm thick stainless steel shadow mask, featuring a 50 µm wide and 7 cm long laser cut slit opening, was aligned and fixed on the fiber holder with temperature resistant polyimide tape ( figure 3(a)). Precise alignment of the shadow mask to the center of fiber groove was done using optical microscope. Fiber holder contains a rectangular 7.5 cm long groove with the opening cross section 230 × 230 µm 2 . When placed inside the groove, optical fiber is unlikely to move or rotate during the sputtering process, which leads to deposition of the straight electrode with uniform width. Since the fiber has a conical tip, micromanipulators are used together with microscope to achieve precise alignment of the fiber tip to the mask opening, which enables electrode deposition on the fiber tip. Since the electrode is not supposed to be deposited on the light emitting segment of tapered fiber tip, fiber tips are first optically tested to determine position in tapered region where light exits fiber tip. Next, fiber holder is mounted on a substrate holder and placed in a substrate transport device in a lock chamber of magnetron sputtering equipment (Von Ardenne LA 320S). Substrate is then moved from a lock into evacuated sputtering chamber, with a titanium and polycrystalline gold sputtering targets (99.99% purity), where the high vacuum (low pressure of 5·10 −6 mbar) is generated in order to minimize the partial pressures of all background gasses and possible contaminants. After high vacuum has been established, sputtering gas (Argon), which comprise the plasma, is flowed into the chamber and total pressure is regulated using a pressure control system. Prior to gold sputtering deposition, a thin titanium film (10 nm) is deposited on the fiber glass surface as an intermediate layer to enhance gold adhesion. Deposition time of 600 s and a DC power of 200 Watts are used to deposit a first gold microelectrode. When the sputtering process is finished, fiber holder together with the optical fiber is removed from sputtering equipment. To sputter second electrode on same optical fiber, precise rotation and alignment setup, consisting of micromanipulators and fiber rotator is used. Since the goal is to achieve two separate equidistant gold microelectrodes on same optical fiber, a rotation angle of 180 • after first sputtering is required. After rotation for 180 • , fiber holder is again placed in the sputtering equipment and sputtering process is repeated. To ensure a very precise alignment, especially of fiber tip, whole alignment process is observed with digital microscope.

Optical characterization
To perform measurements of the transmitted light power through optical fibers with tapered tips, created using previously described grinding setup, integrating sphere photodiode power sensor (S140C, ThorLabs) was used. Measurements were performed in case flat cleaved optical fiber and four tapered optical fibers, each fabricated with different polishing angle (7 • , 15 • , 30 • , and 45 • ) using previously described grinding setup. In case of each cone angle, as well as in case of flat cleaved fiber tip, measurements are performed on three different samples. To enable comparison between obtained results in case of different cone angles, each fiber tip is positioned the same way in respect to the integrating sphere opening. Furthermore, identical butt-coupling procedure between the investigated fibers and fiber connected to LED source was ensured by microscope observation. Experimental demonstration of correlation between cone angles and emission profiles is achieved by immersing tapered fibers in a non-scattering medium, Rhodamine-6G dye water solution (R6G, Sigma Aldrich) with concentration of 1 mg ml −1 . The fluorescence excitation is achieved by coupling green light (530 nm) into investigated optical fibers immersed into Rhodamine 6 G medium. Prior to experiment, light power on LED source is set to 1 mW and kept constant throughout the whole experimental procedure. This experiment is done in dark room atmosphere to avoid any effect of ambient light. Liquid surface of Rhodamine solution is observed with microscope camera with optimal settings required for this experiment. Microscope images of induced fluorescence that corresponds to the tip emission profile are recorded in case of flat cleaved optical fiber, as well as in case of four tapered optical fibers fabricated by applying various polishing angles (7 • , 15 • , 30 • , and 45 • ). However, an information about spatial intensity distribution, obtained from such emission profiles, is not reliable since they are recorded using microscope camera positioned above the container filled with Rhodamine 6G water solution containing full immersed optical fiber tip. In this case, two-dimensional image represents intensity distribution on a plane parallel to fiber optical axis observed from a top view of light emitted from optical fiber, and it contains information about intensity distribution on multiple lower planes along the diameter of optical fiber. Thus, peak intensity at defined distance from tip end on such image does not correspond to peak intensity on beam profile at same distance. More detailed information about spatial intensity distribution are obtained from beam profiles at various distances from fiber tip end. Such profiles at defined distances from tip end are recorded with microscope camera as a light pattern on glass diffuser and later analyzed by image processing in Matlab. To correlate the size of cone angle to emission profile in case of tapered optical fibers, analytical study of light propagation in tapered fiber region is applied [28]. According to an in-depth geometrical analysis presented in [28] (figure 4), there is a connection between semi-cone angle (α) and number of total internal reflections (TIRs) taking place in tapered fiber tip (m): where n co and n cl represent refractive index of fiber core and cladding, respectively, while n stands for refractive index of surrounding medium.

Electrical characterization
EIS is broadly used as a method for determining the quality of recording electrodes. Due to practical reasons, many researchers perform impedance characterization of microelectrodes only at 1 kHz [10,26,29,30], claiming that the main information content in an action potential resides around this frequency. Even though this statement is correct, commenting on cut-off frequency of the electrode ( f cut−off ) is more relevant than just mentioning impedance magnitude at 1 kHz (Z 1kHz = |Z 1kHz | ) [31]. Electrode transitioning from predominantly resistive to capacitive behavior, estimated from magnitude of impedance, is often used to determine f cut -off [31]. Signals above cut-off frequency can be reliably recorded by microelectrode, while signals recorded below f cut -off are substantially phase shifted experiencing nonlinear distortion. The size of electrode-electrolyte interface area directly determines the impedance at 1 kHz, as well as cut-off frequency. Larger electrode area is associated with lower f cut−off , and also lower impedance at 1 kHz [31][32][33].
We performed EIS using Gamry Potentiostat (the Reference 600). To avoid any artifacts due to impedance of counter electrode, three-electrode setup is used. Measurements are performed in frequency range from 100 kHz to 10 Hz, with 10 points per decade at a voltage of 10 mV against the reference electrode. To obtain results close to those expected in brain environment, EIS measurements are performed in Phosphate-buffered saline. Measurements are performed on optrodes containing heat shrink tube, such that only a small controlled portion of electrode area in fiber tip region is exposed. Same measurement procedure is repeated in case of three different electrode areas (60 × 50 µm 2 , 60 × 500 µm 2 , and 100 × 500 µm 2 ). For each electrode, EIS is repeated several times until reproducible curve appeared.

Results
Tapered fiber tips fabricated according to previously described grinding procedure are presented below together with comments on the effect of fabrication procedure on tip shape and its surface quality. Furthermore, results of electrode deposition process are presented in terms of electrode shape, thickness, and adhesion. Optical and electrical optrode performances are commented based on results of previously described characterization techniques.

Tapered fiber tips
Application of angular stages with different inclination (polishing angle) in previously described grinding setup enables fabrication of tapered fiber tips with various cone angles. In this research we used angular stages with an inclination of 7 • , 15 • , 30 • and 45 • , which resulted in fabrication of tapered fiber tips with four different cone angles ( figure 5(a)). In order to demonstrate fabrication reproducibility, four fiber tips are processed using the proposed method with a polishing angle of 15 • ( figure 5(b)). Expected correlation between the polishing angle and cone angles is confirmed from SEM images of tapered fiber tips fabricated with various polishing angles (table 1). Cone angles of fabricated tapered fiber tips were obtained from SEM images assuming that there is no curvature in tapered region. An example of such procedure in case of tapered fiber tip fabricated with 45 • polishing angle is presented on figure 6.

Electrode deposition
Microscopic cross-section image of optical fiber with two deposited gold electrodes is presented on figure 7(a). For a better understanding of electrode shape and thickness, magnified SEM image of a single gold electrode deposited on a fiber glass surface is presented on figure 7(b). Precise alignment of fiber cone tip relative to the slit opening on shadow mask enables sputtering deposition of electrode on the cone tip ( figure 7(c)). Reliable and accurate thickness measurement, based on step height measurements of electrodes fabricated in this work are conducted using surface profiler (DektakXT stylus, Bruker) ( figure 7(d)). Since such measurements are limited to flat substrates, gold microelectrodes are sputtered on flat glass surface after plasma surface treatment, using the same shadow mask and process parameters (600 s processing time, 200 W DC power) as in case of electrode deposition on optical fiber surface. To ensure a direct contact between the shadow mask and a glass substrate, mask is fixed on a glass substrate using temperature resistant polyimide tape. Thickness profile of such 60 mm long gold microelectrode deposited on a flat glass substrate is presented on figure 7(d).

Optical characterization
Results obtained from measurements of transmitted light power using an integrating sphere are presented on figure 8(a). In addition to optrodes transmission properties, another important information, is the light emission profile. Depending on cone angle, light emission profile can be such to enable either large volume, or very focused stimulation. In either case, a detailed characterization of emitted light power distribution is required to enable accurate optogenetic experiments. Fluorescence profiles generated by light emitted from tapered fiber tips fabricated with previously presented grinding setup using various polishing angles (θ), as well as from flat cleaved optical fiber immersed in Rhodamine 6G solution are presented on figure 8(b). Figure 8    Range of cone angles (β) associated with different number of TIRs in tapered region is calculated according to equation (2), and presented in table 2. To gain a better understanding of the effect of cone angle on shape and intensity distribution of light delivered by optrode, microscope images of output beam profiles are obtained at various distances from the tip end ( figure 10(b)). Estimated values of beam width at various distances from tip end in case of flat cleaved optical fiber and fiber tips with various cone angles are graphically presented on figure 8(a). It is important to mention that microscope images presented on figure 10(b) correspond to beam profiles at different distances from tip end placed in air. To gain a better understanding of spatial illumination from various tapered fiber tips, estimated values of beam width at various distances from tip end in case of flat cleaved optical fiber and fiber tips with various cone angles in air are presented in table 3.
Absorption and scattering in brain tissue lead to changes in beam divergence and intensity distribution. However relative comparison between flat cleaved fiber tip and tapered tips with different cone angles positioned in the same medium should not be much affected by the characteristics of the surrounding medium.

Electrical characterization
Bode plot containing information about impedance magnitude (logarithmic) against frequency (logarithmic) for gold microelectrodes of different sizes with a thickness distribution presented on figure 7(d) is presented on figure 11. As expected, impedance magnitude at 1 kHz, and cut-off frequency both decrease with increase of electrode area.

Tapered fiber tips
Profiles of tapered fiber tips fabricated using our grinding setup with various angular stages ( figure 5(a)) suggest that the polishing angle not only determines the cone angle, but also affects the shape of the tapered conical fiber tip. It can be noted that in case of large polishing angles (30 • and 45 • ) fiber tip radius decreases linearly along the axial direction of the fiber towards the tip, whereas the convex shape, as well as increased radius reduction rate are characteristic for fiber tips fabricated with reduced polishing angles (7 • , and 15 • ). The curvature of the convex shape of tapered fiber tips is getting larger with the reduction of polishing angle. The taper length increases with decrease of polishing angle, which means that the contact surface between the optical fiber and grinding disc becomes larger with the reduction of polishing angle. Pressure applied on optical fiber by positioning it on grinding disc can lead to a small fiber bending close to the tip during the grinding process, which becomes more pronounced as the contact area between optical fiber and grinding disc increases. Our grinding setup requires manual positioning of the optical fiber on grinding disc, and at this point we are not able to quantify the pressure on optical fiber during grinding process.
We fabricated four tapered fiber tips with each polishing angle (7 • , 15 • , 30 • , and 45 • ) applying identical earlier described process parameters. From SEM images we were able to estimate cone angle of fabricated tapered fiber tips (table 1). Based on these results, we can state that correlation between polishing angle and resulting cone angle (equation (1)) is approximately confirmed. However, there is a slight deviation from the expected value of cone angle in case of each tapered fiber tip. At this point we cannot perform statistical analysis regarding fabrication reproducibility due to insufficient number of samples. Manual positioning of optical fiber on grinding disc, as well as subsequent exchange of grinding discs during the tapering process is probably leading to deviation of cone angles of fabricated tapered fiber tips from expected values (according to equation (1). This is currently the major limitation of our tapering approach, which is planned to be overcomed by modification of grinding setup in terms of implementation of controlled motorized stage for precise and reproducible fiber positioning on grinding disc.

Electrode deposition
Cross section image of optical fiber with deposited microelectrode obtained from SEM ( figure 7(b)) shows non-uniform thickness distribution of deposited microelectrode. It can be observed that electrode thickness decreases towards both ends in the same manner. This effect is probably due to a curved fiber surface, and it is correlated to an increase in spacing between the fiber surface and the slit opening starting from the center towards both ends of slit opening width. Thickness of deposited microelectrode can  Table 2. Calculated range of cone angles (β = 2α) for different number of total internal reflections (TIRs) in tapered region. Calculations are performed using equation (2) based on analytical study presented in [28].

Number of TIRs in taper region
Range of cone angles (calculated using equation (2)) be controlled by controlling parameters of sputtering deposition process (DC power and sputtering time). However, some processes during deposition affect thickness uniformity of the resulting layer in magnetron sputtering. The effect of geometric parameters, such as distance between target and substrate, ionic energy, target erosion area, gas pressure, and substrate temperature influence thickness uniformity of deposited thin films. Furthermore, surface homogeneity of the target plays significant role in film thickness distribution. An increased electrode width in tapered region ( figure 7(c)) is due to a larger under sputtering effect caused by increased distance between fiber tip surface and slit opening of the shadow mask.

Optical characterization
Compared to the flat cleaved optical fiber, transmitted light power decreases with cone angle reduction. According to data presented on figure 7(a), relationship between cone angle and transmitted light power can be approximated by exponential function with 2nd order polynomial exponent: where x and y values represent polishing angle and transmitted light power, respectively. Fitting parameters (a = −1.23, b = 0.01, c = −5.09 · 10 −6 ) were obtained in origin using nonlinear curve fitting method for average values of transmitted light power in case of flat and tapered fiber tips with different cone angles. Average value of transmitted light power related to a certain cone angle was obtained from measurements performed on three tapered optical fibers fabricated with identical polishing angle. Reduction of cone angle leads to increased number of TIRs inside tapered fiber tip region [28,34,35], which leads to shift in starting point of light emission along the fiber axis and may contribute to reduced power of emitted light. Other factors   0  244  656  625  1111  694  150  269  775  831  1553  1019  300  319  944  1162  1935  1295  450  369  1126  1550  2050  1418  600  413  1335  1896  1494  1468  750  469  1500  2230  1537  1565  900  506  1700  2625  1768  1706  1050  567  1900  2974  2000  1846 that may influence overall transmitted light power through optical fiber with cone tip are Fresnel reflection, surface scattering, and back reflection. Contrast between refractive index of Silica glass fiber and air results in Fresnel reflection at the optrode backside, as well as at the cone tip. Successful reduction of Fresnel reflection at the optrode back side is achieved by filling the air gap between the optrode and fiber coupled to light source with index matching gel. However, this procedure cannot be applied at the interface between cone tip and air. Since the cone angle affects tapered region and influences the size of tip-air interface, it may differently contribute to losses due to Fresnel reflection depending on size of cone angle. Surface scattering may also contribute to transmission losses within the cone tip [34]. Even though cone tips are polished to obtain a smooth surface with good optical quality, sidewall roughness cannot be completely eliminated. Amplitude and spatial periodicity of sidewall roughness in tapered region influence total integrated scatter (TIS) that represents the amount of light scattered by a surface from a single reflection. Depending on total number of reflections in tapered region, TIS accumulates and contributes to overall scattering loss. Smaller cone angle leads to longer tapered region, and thus higher number of reflections inside it, which may explain the effect of cone angle reduction on transmitted light power. To achieve light stimulation of genetically modified neurons in optogenetic experiment, one needs to deliver required light power to the stimulation site. Therefore, it is crucial to have an information about optrode's transmission capabilities. When planning an optogenetic experiment, beside power loss that arises from the coupling between the optrode and light source, optrode transmission properties need to be considered in order to set the sufficient power of the light source, and thus to deliver sufficient light power to the stimulation site. Light rays are guided through fiber core by TIR until they reach tapered region. Each ray reflection in tapered region leads to increase in its propagation angle, with respect to the fiber optical axis, by an amount equal to the value of cone angle (β) [19]. When the TIR is lost, light ray leaves the cone tip and radiates into the surrounding medium. Compared to flat cleaved fiber tip, fiber tips fabricated with polishing angles equal or less than 30 • offer the possibility of spatially extended illumination ( figure 8(b)). Furthermore, tapered fiber tip fabricated with polishing angle of 45 • enables spatially restricted illumination close to fiber tip with a focusing feature. While light beam diverges immediately after leaving flat cleaved fiber tip, onset of beam divergence is located much further from tip end in case of tapered fiber tip fabricated with polishing angle of 45 • . Quantitative analysis of beam divergence from tapered fiber tips is presented on figure 10(a). Beam shape in case of tapered fiber tip fabricated with 45 • polishing angle is comparable to that in case of flat cleaved optical fiber. However, compared to flat cleaved optical fiber, beam width at defined distance from fiber tip end is approximately three times larger in case of tapered fiber tip fabricated with polishing angle of 45 • . Starting from fiber tip end up to distance of approximately 700 µm, beam divergence significantly differs among tapered fiber tips fabricated with different polishing angles. However, at a distance equal or larger than 700 µm from tip end, difference in width of light beams emitted in air by tapered fiber tips fabricated with polishing angles of 45 • , 15 • and 7 • becomes negligible. Compared to flat cleaved optical fiber, fibers with tapered tips offer much larger illumination volume near tip end, but with different intensity distribution.
Illuminated volume depends on beam width near cone tip, as well as on beam divergence and tissue optical properties. According to beam profiles at various distances from tip end, tapered fiber tip fabricated with polishing angle of 45 • provides most uniform illumination volume through tissue depth compared to other tapered fiber tips fabricated in this study. Decrease of cone angle of tapered fibers leads to wider illumination volume. Based on microscope images of beam profiles ( figure 10(b)), it can be noted that the cone angle of tapered fiber tip affects not only the size of illumination volume, but also its shape. The shape of illumination volume in case of optical fiber fabricated with 45 • polishing angle does not differ much from that of flat cleaved optical fiber. However, optical fiber with tapered tip fabricated with 45 • polishing angle offers much wider illumination volume, which increases with distance from fiber tip end. Illumination volume obtained from tapered fiber tips fabricated with polishing angles smaller than 45 • is less uniform with most of intensity located in the center of emission profile.
Intensity distribution of induced fluorescence suggest that increase of cone angle leads to larger light emitting segment of the taper, as well as to larger taper emitting length (figure 9). In case of tapered fiber tip fabricated with 45 • polishing angle light is emitted from entire-tip length, with estimated taper emitting length of approximately 95% of entire taper length. Decreasing polishing angle to 30 • does not affect light emitting segment of the taper, but leads to reduction of taper emitting length to approximately 72% of entire taper length. Further reduction of polishing angle to 15 • and 7 • is followed by decrease in both, light emitting segment and taper emitting length. Taper emitting length decreases to approximately 27% and 13% of entire taper length in case of fiber tips fabricated with polishing angle of 15 • and 7 • , respectively. Reduction of polishing angle, and thus cone the angle, followed by reduction of output zone along the length of a specific taper can be explained by mechanism of TIR in tapered region.
Results presented in table 2 agree with previously observed fluorescence profiles ( figure 8(b)), and indicate the absence of TIR in tapered region of tapered fiber tips fabricated with polishing angles of 45 • and 30 • . This explains light emission along entire taper of these fiber tips. Furthermore, reduction of light emitting segment to 53% and 30% of entire taper length in case of tapered fiber tips fabricated with polishing angles of 15 • and 7 • , respectively, can be explained by increased number of TIRs taking place in taper region. It is important to mention that equation (2) assumes linear taper, whereas the real structure has a modestly parabolic shape, which can slightly affect the results.
Compared to flat cleaved optical fiber, tapered fibers fabricated with polishing angles of 45 • and 30 • show similar behavior of intensity reduction along fiber axis (figure 8(c)), except a focusing feature observed in case of tapered fiber tip fabricated with polishing angle of 45 • . However, tapered fibers fabricated with polishing angles of 15 • and 7 • show a significantly different behavior of intensity distribution along fiber axis. Starting from the tip end, intensity decreases as soon as the light leaves fiber tip, and reduces to half of its maximum value after approximately 430 µm and 170 µm in case of tapered tips fabricated with 15 • and 7 • cone angles, respectively. Intensity drop to half of its maximum value is observed approximately one millimeter from tapered tip fabricated with 30 • cone angle, whereas the same intensity reduction is reached approximately two millimeters from flat cleaved fiber tip, as well as from the tapered tip fabricated with 45 • cone angle.

Electrical characterization
EIS represents an important step before application of deposited microelectrodes for recording purpose, and helps to determine electrode area that should be exposed to obtain the most satisfying electrophysiological recording. Smallest tested, 60 µm wide and 50 µm long, electrode has a very high impedance magnitude (455 kΩ) at 1 kHz ( figure 11).
According to previous research [29], gold electrodes coated on optical fiber with such impedance value are considered to be applicable for electrophysiological recording in optogenetic experiments. However, cutoff frequency (f cut−off ) of such small electrode is located outside measurement frequency range (10 Hz-100 kHz).
Some studies showed that signals above cut-off frequency can be reliably recorded by microelectrode, while signals recorded below f cut−off are expected to be phase shifted. In this particular case, it means that a gold microelectrode with cut-off frequency higher than 100 kHz would not be suitable for reliable recordings (due to low signal to noise ratio) at frequency of 1 kHz, where main information content in an action potential resides. One way to address this issue is the use of rough electrode materials or various roughening methods providing a large electrochemically active surface area on minimal geometrical dimensions [30,36,37]. Even though these methods enable to maintain small electrode size while achieving electrochemical properties of large electrodes, most roughening methods require rather complex processes which usually result in limited mechanical stability [38] or delamination [39]. Comparably easier method to achieve rough electrode surface is electrolyte-based electrochemically deposition of platinum coating [39], which leads to 28 times lower average impedance compared to non-coated platinum electrodes. Increasing of the electrode area, by increasing exposed electrode length from 50 µm to 500 µm, resulted in decrease of Z 1kHz to 98 kΩ followed by shifting of f cut−off to 12 kHz (figure 10). Further enlargement of 500 µm long electrode area, by increasing electrode width from 60 µm to 100 µm, have led to further reduction of Z 1 kHz to 82 kΩ and f cut−off to 8 kHz. These values agree well with previously published EIS data for gold electrodes [40], and prove already established method of decreasing Z 1 kHz and f cut−off by increasing geometrical electrode area. However, application of large electrodes is limited to recording of local field potentials (LFPs) due to spatial averaging of recorded signals. Since there is no significant difference in signal amplitudes detected by differently sized electrodes far from the peak signal source [33], these electrodes (60 × 500 µm 2 and 100 × 500 µm 2 ) are expected to be suitable for extracellular distant recording of LFPs.

Conclusion
In this paper, we presented a miniaturized fiber-based optrode for simultaneous light delivery and electrophysiological recording in optogenetics. This research advanced current fiber-based optrode designs by introducing conical fiber tips with various angles for custom light delivery. Compared to existing fiberbased optrodes, the device presented here differs in terms of its novel approach to fiber tip formation which also provides reduced invasiveness. Related work only integrates a single electrode fabricated as partially or fully coated fiber. Furthermore, we deposited two separate microelectrodes on the fiber surface, which increases density of recording sites, and thus enables electrical recording from tissue sections on opposite sides of the fiber tip.
A custom grinding setup enables fabrication of tapered fiber tips with good optical quality. Results presented in this work suggest that tapered fiber tips with correct cone angle are suitable to define the illuminated tissue volume, which is beneficial in optogenetic experiments. As the cone angle affects the total transmitted optical power, adjusting the power of the light source is essential to deliver sufficient power to the stimulation site. Our custom electrode deposition technique, enables deposition of multiple microelectrodes on the surface of the optical fiber with defined electrode thickness and angular spacing. In addition to increased mechanical stability, application of heat shrink tube also allows control of exposed electrode area, which strongly determines contact impedance in electrophysiological recording. Beyond the presented tapered fiber tips, our custom grinding setup can be extended to allow for fabrication of tip geometries which yield well-controlled illumination profiles.

Data availability statement
The data that support the findings of this study are available upon reasonable request from the authors.