**1. Introduction**

**\***

Lab-on-a-chip systems, microfluidic bioreactors and organ-on-chip platforms with integrated sensors and actuators for the monitoring of crucial parameters (e.g., flow rate, temperature and pH) are of grea<sup>t</sup> interest to maintain micro-environmental conditions [1–3]. On the other hand, inducing perturbations of these parameters leads to new perceptions of such systems, e.g., by changing the extracellular pH during cell culturing [4–7]. Often, due to geometrical restrictions inside microfluidic channels, the flexible integration of "conventional" sensing devices is not easy to accomplish [1]. The sensor information is mostly obtained at a fixed position, predefined during fabrication of the usually rigid sensor geometries. At the same time, a spatially resolved mapping of the molecular species functionalized areas are defined inside the microstructure. In addition, actuation functionalities should be flexible as well, enabling manipulation of e.g., local pH changes without affecting neighboring elements. Therefore, a precise addressability of both the actuator (here, a light-addressable electrode, LAE) and the sensor (here, a light-addressable potentiometric sensor, LAPS) is required.

**Citation:** Welden, R.; Jablonski, M.; Wege, C.; Keusgen, M.; Wagner, P.H.; Wagner, T.; Schöning, M.J. Light-Addressable Actuator-Sensor Platform for Monitoring and Manipulation of pH Gradients in Microfluidics: A Case Study with the Enzyme Penicillinase. *Biosensors* **2021**, *11*, 171. https://doi.org/10.3390/ bios11060171

Received: 30 April 2021 Accepted: 25 May 2021 Published: 27 May 2021

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2021 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

LAPS is a semiconductor-based chemical sensor which was first proposed by Hafeman et al., in 1988 [8]. LAPS belongs to the group of field-effect-based electrochemical sensors with an electrolyte-insulator-semiconductor (EIS) structure [9]. LAPS offers (depending on its transducer layer) the spatially resolved monitoring of concentration-dependent surface-potential changes, e.g., induced by (bio)chemical/biological molecules or living cells in the analyte solution. LAPS can be designed as multiwell- and multianalyte-sensor devices and can serve for chemical imaging, where the distribution of the analyte concentration is visualized on its chip surface [10–14]. Moreover, LAPS provides a broad range of possible applications and has been utilized for various (bio)chemical and biotechnological approaches, such as monitoring the metabolic activity of bacteria in fermentation broth [14], for on-sensor cryopreservation of cells [15], multi-ion and penicillin detection [16,17], and DNA sensing [18].

In contrast to LAPS, a LAE exhibits no insulating layer, having a direct charge transfer with the analyte. By spatially resolved illumination, conductive areas inside the semiconducting chip can be defined resulting in, e.g., photoelectrocatalytical water oxidation. Furthermore, the LAE can be used for photoelectrochemical material deposition, adjustment of a pH gradient or cell stimulation [19–21]. Advantageously, the LAE offers a high flexibility without the need for patterning complicated electrode arrays [22,23].

The combination of LAPS and LAE as a sensor-actuator system would enable the simultaneous, spatially resolved pH manipulation and monitoring inside a microfluidic system. This way, reaction processes taking place inside microchannels could be further analyzed and optimized, which is helpful for e.g., studying the response characteristics of immobilized bioreceptors in a microfluidic channel, such as enzymes.

The proposed experimental approach of a light-addressable actuator-sensor platform (consisting of a LAE and a LAPS) elaborates for the first time the mutual reaction of local pH gradients inside a microfluidic channel and the enzyme-triggered sensor signal. Penicillinase has been selected as a model enzyme to detect penicillin, since it is a robust enzyme making it advantageous for experimental application. The enzymatic reaction induces a pH change, which can be detected with the LAPS. Si3N4 was used as pH-sensitive transducer layer for the LAPS, and TiO2 as photoelectrocatalytical material inducing pHvalue manipulations.

The surface morphology of the LAE and the LAPS surface where physically characterized by means of scanning electron microscopy (SEM). The penicillin sensitivity of the LAPS and the effect of pH changes on the enzyme activity, provoked by the LAE, were evaluated by photocurrent-voltage and chemical-image measurements. Dynamic pH variations induced by the LAE can be further used to control the enzymatic reaction rate and to adjust the biosensor response.

#### **2. Materials and Methods**

#### *2.1. Fabrication Process of Light-Addressable Electrodes (LAEs)*

The LAE employed in this study consists of a glass/SnO2:F/TiO2 heterostructure. The SnO2:F glass substrate (7 <sup>Ω</sup>·sq<sup>−</sup>1) was purchased from Sigma Aldrich (Darmstadt, Germany). The SnO2:F glass substrate was cleaned in an ultrasonic bath with acetone, 2-isopropanol and deionized water for 5 min, respectively, and dried with nitrogen. Afterwards, the TiO2 layer was deposited by pulsed laser deposition (PLD). During the PLD process, a TiO2 target (MaTeck Material-Technologie and Kristalle GmbH, Jülich, Germany) was vaporized with a KrF-excimer laser (λ = 248 nm) using a power density of 5.0 J·cm<sup>−</sup><sup>2</sup> with a repetition frequency of 10 Hz at a pressure of 2.0 × 10−<sup>2</sup> hPa O2 for 700 s. Hereby, the SnO2:F glass substrate was heated during the PLD process to 400 ◦C to achieve a rutil crystal structure. To achieve an inlet and outlet for the later prepared microfluidic structure, two holes with a diameter of 1.2 mm were drilled in the LAE.

#### *2.2. Preparation of Light-Addressable Potentiometric Sensor (LAPS) Chips*

The utilized LAPS chips, consisting of a n-Si/SiO2/Si3N4-multilayer structure, were acquired from SEIREN KST Corp. (Fukui, Japan). The thickness of the n-Si, SiO2 and Si3N4 layer was 100 μm, 50 nm and 50 nm, respectively. To remove the surface-oxide layer, the rear side was treated by wet-chemical etching with hydrofluoric acid (HF). Afterwards, a 300 nm thick aluminum (Al) film was deposited by electron-beam evaporation with a deposition rate of 2 nm·s<sup>−</sup><sup>1</sup> to contact the n-Si substrate electrically. The wafer was diced into 20 × 20 mm<sup>2</sup> chips and an optical window (ca. 15 × 15 mm2) was made by etching an inner rectangle of the Al layer with 5% HF, leaving an outer Al frame.

#### *2.3. Enzyme Immoblization with Tobacco Mosaic Virus Particles as Enzyme Nanocarriers on LAPS Chips*

For enzyme immobilization, *tobacco mosaic virus* (TMV) particles were utilized as enzyme nanocarriers. A TMV variant (S3C) that exposes a cystein residue on each coat protein was used [24,25]. To functionalize the TMV particles, bifunctional biotin-linker molecules (EZ-Link Maleimide-PEG11-Biotin, ThermoScientific, Rockford, IL, USA) were covalently bound to the thiol groups located on the surface of each coat protein, as described in [26–28]. The biotinylated TMV particles were suspended in 10 mM sodium-potassium-phosphate (SPP) buffer (pH 7.0) and stored at 4 ◦C until use. As a model enzyme, penicillinase from *Bacillus cereus* (Sigma-Aldrich, Darmstadt, Germany) was utilized. For specific enzyme immobilization to the biotin linkers on the surface of the TMV nanotubes, the enzyme was conjugated with streptavidin molecules using a commercial streptavidin conjugation kit (LNK162STR, Bio-Rad, Feldkirchen, Germany) [26,27]. The streptavidin-conjugated penicillinase ([SA]-penicillinase) was stored in 10 mM phosphate-buffered saline (PBS) buffer (1000 Units·mL−1, pH 7.0) at 4 ◦C until further use.

The LAPS chips were cleaned in an ultrasonic bath for 5 min in acetone, 2-isopropanol, ethanol and deionized water, respectively. After drying with nitrogen, 10 μL TMV solution (0.1 mg·mL−1) were drop-coated on the Si3N4 surface of the later-on microfluidic channel, and incubated for 1 h at room temperature (RT) in a humid chamber. Afterwards, the solution with unbound TMV was washed away with deionized water and the sensor chip was dried with nitrogen. In the next step, 5 μL [SA]-penicillinase solution were drop-coated on the immobilized TMV particles and incubated for 1.5 h at RT in a humid chamber. After the incubation time, the sensor chip was rinsed again with deionized water to remove unbound enzyme molecules and dried with nitrogen.

#### *2.4. LAPS-LAE Microfluidic Assembly*

To prepare the LAPS/microfluidic foil/LAE sandwich structure (schematically depicted in Figure 1a, a ~86 μm thick double-sided adhesive microfluidic tape (3MTM, St. Paul, MN, USA) was patterned by laser cutting using a ProtoLaser U3 (LPKS Laser and Electronics AG, Garbsen, Germany). A 20 × 20 mm<sup>2</sup> rectangle with a 1.0 mm wide channel was cut out of the tape. The LAE was cleaned in an ultrasonic bath with acetone, 2-isopropanol, ethanol and deionized water and finally dried with nitrogen. Afterwards, the laser-cut microfluidic foil was stuck onto the TiO2 surface of the LAE, positioning the drilled holes of the LAE at the inlet and outlet of the microfluidic channel. In the final step, the TMVand penicillinase-functionalized LAPS chip was placed below the LAE-microfluidic structure, immobilizing the enzyme-loaded TMV particles at the bottom of the microfluidic channel. For tube connection, ferrules were attached to the inlet and outlet with the help of a photopolymer.

**Figure 1.** (**a**) Schematic of the microfluidic setup with a light-addressable potentiometric sensor (LAPS)/microfluidic foil/ light-addressable electrode (LAE)-sandwich structure. *Tobacco mosaic virus* (TMV) particles functionalized with the enzyme penicillinase are immobilized inside the microchannel. (**b**) Typical shape of a photocurrent-voltage curve for a n-type LAPS with characteristic regions of inversion, depletion and accumulation.

#### *2.5. Measurement Setup and Characterization Methods*

The inlet tube of the microfluidic setup was connected to a syringe-driven pump system (neMESYS 290N, Cetoni GmbH, Korbussen, Germany) to control the flow inside the channel. In the outlet tube, a Pt-counter electrode and a reference electrode (DRIREF-2SH, World Precision Instruments, Sarasota, FL, USA) for the electrical connection of the LAE and LAPS were inserted. A clamp connected the Al rear side contact of the LAPS to a transimpedance amplifier (gain = 10<sup>7</sup> <sup>V</sup>·A−1, AMP100, Thorlabs GmbH, Bergkirchen, Germany) to convert the alternating photocurrent into a measurable voltage. The voltage was recorded by a measurement card (USB 7855R, NI, Austin, TX, USA). The same card also provided the bias voltage to the LAPS with respect to the counter electrode. A potential was directly applied to the LAE and Pt-counter electrode by a source measurement unit (2600b, Keithley Instruments, Solon, OH, USA).

The LAE and LAPS rear side were illuminated by a digital light processing (DLP) projector (STAR-07, ViALUX Messtechnik + Bildverarbeitung GmbH, Chemnitz, Germany) with a 405 nm and a 613 nm light-emitting diode (LED) light source. Both DLPs are modified with a lens system to focus each of the 1024 × 768 micromirrors to a size of 10 × 10 μm2. All measurements were performed in a dark Faraday cage at room temperature (RT). For all experiments, 0.33 mM PBS buffer was used. The pH was adjusted by titration with NaOH and HCl. For penicillin detection, varying concentrations of penicillin G (Sigma Aldrich, Darmstadt, Germany) were added to the measurement solution.

For photoelectrocatalytically induced pH changes, a constant potential of 300 mV was applied to the LAE with respect to the Pt-counter electrode. After reaching steady- state conditions, the rear side was illuminated with spots of various sizes.

For LAPS characterization, three measurement modes were applied: photocurrentvoltage (I-V), chemical-image and photocurrent-time mode. The illumination of the LAPS-DLP projector was modulated with a frequency of 512 Hz to achieve an alternating photocurrent for all measurements. The voltage from the transimpedance amplifier was sampled with a frequency of 50 kHz and further processed. For I-V curves, the measurement time for each bias voltage was 400 ms. Figure 1b depicts a theoretical I-V curve of a n-type silicon LAPS. In the I-V mode, the applied voltage was swept from 0 to −3.0 V while measuring the photocurrent for a fixed illumination spot. The typical output curve shows the three characteristic regions of inversion, depletion and accumulation. A pH increase or decrease shifts the I-V curve to more positive or negative voltage, respectively,

which is particularly evident in the depletion region. As slight changes of the photocurrent amplitude can occur when replacing the measurement solutions, the I-V curves were normalized with respect to the inversion region.

To obtain spatially resolved images from the microfluidic channel, the chemical image mode was utilized. A constant bias voltage, chosen close to the inflection point of the I-V curve, was applied and the rear side was scanned sequentially with a moving light beam and a sampling time of 200 ms for each spot (250 × 250 μm2). In the results section, differential chemical images are visualized. For that, the chemical image after the enzymatic reaction was subtracted from the initial reference chemical image (before enzymatic catalysis of penicillin by penicillinase). On the basis of the exemplary I-V curve in Figure 1b, at a fixed bias voltage, a pH decrease results in a photocurrent drop, while it increases with rising pH values.

Similarly, during the photocurrent-time mode, the temporal change of the photocurrent is measured for a fixed illumination spot and a fixed bias voltage (from the I-V inflection point), enabling the dynamic detection of pH changes. The sampling time was 1 s.

With all three measurement modes, changes of the pH value in the solution can be monitored. In this work, the resulting pH changes are caused by two different proton generation mechanisms, being photoelectrocatalysis and enzymatic conversion. The induced pH change from the LAE originates mostly from the water oxidation reaction,

$$2\,\mathrm{H}\_{2}\mathrm{O} \to 4\,\mathrm{e}^{-} + 4\,\mathrm{H}^{+} + \mathrm{O}\_{2} \tag{1}$$

where water is split into H+ ions and oxygen. The second pH change is due to the immobilized enzyme penicillinase, where penicillin is converted to penicilloic acid and H+ ions through the enzyme's ß-lactamase activity,

$$\text{pericillin} + \text{H}\_2\text{O} \rightarrow \text{pericillloic acid} + \text{H}^+.\tag{2}$$
