Nano-Enhanced Electric-Field Treatment Harnessing Lightning-Rod Effect for 1 Rapid Bacteria Inactivation

14 The growth of undesired bacteria can cause numerous problems. Seeking effective and sustainable 15 bacteria inactivation approaches is an everlasting effort. Here, we show that nano-enhanced 16 electric field treatment (NEEFT) can cause rapid bacteria inactivation with a lower applied voltage 17 than bulk EFT. A lab-on-a-chip with nanowedge-modified electrodes is developed, and the 18 bacteria inactivation in NEEFT is visualized and studied in real-time at a single-cell level. Rapid 19 bacteria inactivation (~ 1 ms) occurs specifically at nanowedge tips where the electric field is 20 enhanced due to the lightning-rod effect. Nanowedges with a high aspect ratio are critical for 21 bacteria inactivation. NEEFT works for both immobilized and free-moving cells, where the free- 22 moving cells will be first attracted to the nanowedge tips followed by rapid inactivation. The 23 mechanism study shows that the bacteria inactivation is caused by electroporation induced by the 24 nano-enhanced electric field. The bacteria inactivation performance depends on the strength of the 25 enhanced electric field instead of the applied voltage. Quick pore closure and membrane recovery 26 under moderate NEEFT indicate that electroporation is the predominant mechanism. NEEFT only 27 requires facile treatment to achieve bacteria inactivation, which is safe for treating delicate samples 28 and energy-efficient for large scale applications. It is also expected to find applications for targeted 29 cell inactivation.

moving cells will be first attracted to the nanowedge tips followed by rapid inactivation. The 23 mechanism study shows that the bacteria inactivation is caused by electroporation induced by the 24 nano-enhanced electric field. The bacteria inactivation performance depends on the strength of the 25 enhanced electric field instead of the applied voltage. Quick pore closure and membrane recovery 26 under moderate NEEFT indicate that electroporation is the predominant mechanism. NEEFT only 27 requires facile treatment to achieve bacteria inactivation, which is safe for treating delicate samples 28 and energy-efficient for large scale applications. It is also expected to find applications for targeted 29 cell inactivation. 30 31 32 33 34 Bacteria are indispensable for both ecological systems and human bodies, but the growth of 35 undesired bacteria can also cause serious problems. Seeking approaches for bacteria inactivation 36 is an everlasting effort. Most of our current practices for bacteria inactivation highly rely on the 37 uses of chemicals, such as antibiotics for infection treatment, chlorine for water disinfection, 38 antiseptics for food preservation, and chemical anti-fouling agents. They have been effectively 39 inactivating bacteria, but caused new problems: overusing of antibiotics has already raised the 40 concern of antibiotic resistance, 1 chlorination generates disinfection by-products (DBPs) that can 41 be carcinogenic; 2 food antiseptics and anti-fouling agents themselves may be harmful to human 42 health or the environment. 43 Effective physical processes, such as thermo/ultraviolet radiation, 3,4 acoustic vibration, 5,6 44 microwave, 7 and electric-field treatment (EFT), 8 can be superior alternatives to chemical 45 approaches for bacteria inactivation, although many of them suffer from high capital cost or energy 46 consumption. Among these processes, the EFT has been found increasing interest for food 47 preservation and water disinfection. 9-12 The EFT aims to inactivate bacteria by electroporation: 48 when a cell is exposed to a strong electric field, an induced transmembrane voltage (TMV) will 49 cause pore formation on the lipid bilayer membrane, 13-15 and when this external electric field is 50 strong enough, the membrane damage, i.e., the pores, will become lethal to the bacterial cells. 10 51 The lethal electroporation threshold was found to be between 10 ~ 35 kV/cm. 16 Typically, in order 52 to achieve the strong enough electric field, the EFT processes will require high applied voltages 53 (e.g., ~ 23 kV to achieve 35 kV/cm on the electrodes with 0.65 cm distance), 17 which leads to 54 safety issues, side reactions, and high energy consumption. 55 A strategy to realize the high electric-field strength with lower voltages is to decorate the 56 electrodes with sharp objects, such as nanowires or nanowedges. Attributed to the lightning-rod 57 effect, the electric field near the tips could be largely enhanced depending on the aspect ratio of 58 the electrode decorations. 18 As a result, even with relatively low applied voltages, the nano-59 enhanced electric field can still build up the transmembrane voltage that is sufficient to cause 60 irreversible electroporation and bacteria inactivation. Although this concept has been claimed as 61 the predominant mechanism for bench-scale EFT water disinfection devices equipped with 62 nanowire-modified electrodes, 19-27 direct demonstration of lightning-rod effect for bacteria 63 inactivation, especially at the single-cell level, is not yet done. Here, we conduct nano-enhanced 64 EFT (NEEFT) on lab-on-a-chip devices with nanowedge-modified electrodes and characterize the 65 microbial inactivation process in-situ and in real-time. Results show that bacteria located at the 66 tips of nanowedges on both positive and negative electrodes are rapidly inactivated at the voltages 67 which are not sufficient to kill bacteria in bulk. Electroporation induced by the nano-enhanced 68 electric field attributed to the lightning-rod effect is demonstrated to be the predominant 69 mechanism for this bacteria inactivation. 70 71 72 Results 73 Visualization of bacteria inactivation by the NEEFT. We develop a lab-on-chip device with 74 gold nanowedges fabricated on both positive and negative electrodes ( Fig. 1a & Fig. S1). The gap 75 between the two electrodes is 50 μm. The length and thickness of the nanowedge are 8 µm and 76 200 nm, respectively. The width of the nanowedge tip is 200 nm, and it gradually increases to 1 77 μm to allow a steadier connection to the bulk electrode. This is the default chip design for our 78 experiments unless otherwise stated. When an 18 V voltage is applied to the two electrodes, the 79 electric field near the nanowedge tips will be enhanced due to the lightning-rod effect, which is 80 simulated using COMSOL Multiphysics (Fig. 1b).

81
Model bacteria Staphylococcus epidermidis (S. epidermidis) cells immobilized on the poly-L-82 lysine coated chip are uniformly distributed between the positive and negative electrodes (Fig. 1c). 83 Live-and-dead cell distinguishing stain propidium iodide (PI) was added in the deionized water 84 (DI water) medium before treatment (See experimental setup in Fig. S2a). After 500,000 electrical 85 pulses at 18 V with 2 μs pulse width and 100 μs period are applied (denoted as 18 V/2 μs/100 86 μs/500,000 pulses, see the waveform in Fig. S3), the bacteria at the tips of nanowedges on both 87 positive and negative electrodes show red fluorescence of the PI stain, indicating cell inactivation, 88 while cells anywhere else are intact (Fig. 1d). The zoom-in image clearly shows that only the cells 89 located very close to the nanowedge tips are inactivated, which is consistent with the electric field 90 enhancement pattern (Fig. 1b). By comparison, for the electrodes that have no nanowedge 91 modification but a smaller gap of 34 μm, hardly any cells are inactivated (Fig. S4), suggesting that 92 this treatment is not sufficient to kill bacteria in bulk. Therefore, NEEFT can cause bacteria 93 inactivation with lower applied voltages than in bulk-EFT. To the best of our knowledge, this is 94 also the first time that the bacteria inactivation in NEEFT is visualized at the single-cell level.

95
The bacteria inactivation process is observed in real-time. The onset position of PI fluorescence 96 indicates that the cell membrane damage takes place at the position adjacent to the nanowedge tip, 97 where the nano-enhanced electric field has the highest strength. The circled bacteria cells are at 98 the nanowedge tips on the negative electrode ( Fig. 1e) and positive electrode (Fig. 1f), and they

112
The observed bacteria inactivation process also shows that bacteria inactivation in NEEFT is 113 very quick (Video S1). To figure out how fast this inactivation occurs, different effective treatment 114 time (i.e., the total time that the applied voltage is not zero, equals to pulse width × pulse number) 115 is tested by applying different pulse numbers of 2 μs/100 μs pulses. Under 30 V and 18 V applied 116 voltage, 0.1 ms and 1 ms of effective treatment times are long enough to achieve >80% bacteria 117 inactivation (represented as the percentage of nanowedges inducing bacteria inactivation at tips), 118 indicating that bacteria inactivation in NEEFT is a very rapid process (Fig. 2a). Under relatively 119 lower applied voltages (14 V and 10 V), bacteria inactivation stays at low percentages up to 1000 120 ms of effective treatment time, suggesting that the limiting factor of the lower bacteria inactivation 121 is the applied voltage rather than treatment time (Fig. 2a). Therefore, the bacteria inactivation at 122 different applied voltages with 1000 ms effective treatment time are tested. The bacteria 123 inactivation shows a positive correlation with the applied voltage (Fig. 2b). The inactivation starts 124 at a low voltage of 10 V, and 20 V is already high enough to achieve bacteria inactivation for 125 almost all nanowedges, and there is no significant difference between the positive and negative

155
Mechanism of bacteria inactivation in nano-enhanced EFT. 156 In NEEFT, only the bacteria located near nanowedge tips are inactivated, while bacteria in bulk 157 are not affected. This pattern is consistent with the electric field enhancement of nanowedges due 158 to the lightning-rod effect. Therefore, irreversible electroporation induced by the enhanced electric 159 field is considered as the predominant mechanism for bacteria inactivation in NEEFT. Here, we 160 investigate this mechanism and the evidence collected is discussed below. In these studies, model 161 bacteria S. epidermidis are all immobilized on the chip for more precise characterization.

163
The bacteria inactivation depends on the strength of the nano-enhanced electric field. 164 To verify whether the observed bacteria inactivation is directly due to the nano-enhanced 165 electric field instead of the applied voltage, chips of different positive/negative electrode gaps (25 166 μm, 50 μm, 100 μm) and with nanowedges of different intervals (0.8 μm, 4 μm, 40 μm) are tested 167 for bacteria inactivation. The strength of the nano-enhanced electric field is reversely proportional 168 to the gap between the two electrodes ( Figs. S5a & b). Therefore, with the same applied voltage, 169 chips with a smaller gap achieve a higher percentage of bacteria inactivation (Fig. 4a & Fig. S5c).

170
Similarly, because of the stronger lightning-rod effect for electric-field enhancement (Figs. S6a &  171 b), the nanowedges with a larger interval in between could achieve higher bacteria inactivation 172 under the same applied voltage (Fig. 4b & Fig. S6c). When all the results are analyzed, the 173 percentage of bacteria inactivation at the tips of nanowedges shows a positive correlation with the 174 electric field strength ( Fig. 4c left), but not with the applied voltage (Fig. 4c right). This result 175 indicates that the bacteria inactivation is attributed to the nano-enhanced electric field. 176

182
The bacteria inactivation is not attributed to reactive oxygen species (ROS pulses treatment, DCFH-DA stained cells show no fluorescence (Fig. 5a), suggesting no ROS 187 generation. Meanwhile, >90% bacteria inactivation is achieved (Fig. 5b & experiment group, no 188 DMSO in Fig. 5g), indicating that this bacteria inactivation is not due to ROS damage. To confirm 189 this ROS detection method is valid, we intentionally induce ROS generation with a much longer 190 pulse width in the positive control (20 V/200 μs/10 ms/1000 pulses). The significant green 191 fluorescence of DCFH-DA stained cells shows that ROS is generated near the positive electrode 192 (Fig. 5d). The positive electrode shows more inactivated bacteria at each nanowedge tip than the 193 experiment group and negative electrodes (Fig. 5e & positive control, no DMSO in Fig. 5g), 194 which could be attributed to the ROS damage.

195
To further confirm that the bacteria inactivation at 30 V/2 μs/100 μs is not due to ROS 196 generation, a ROS scavenger, DMSO, is added to the medium at 15% (w/w) to quench ROS and 197 protect bacteria from ROS damage. 32 In the positive control group, the bacteria at the positive 198 electrodes are largely protected by DMSO ( Fig. 5f & positive control in Fig. 5g), proving that 199 15% DMSO is able to protect bacteria from ROS damage. In the experiment group, even with the 200 ROS scavenger DMSO, the bacteria inactivation percentage and inactivated cell number are not 201 affected ( Fig. 5c & experiment group in Fig. 5g), which further confirms that the bacteria 202 inactivation is not due to ROS damage.

210
Quick cell membrane recovery supports electroporation as the main bacteria inactivation 211 mechanism.

212
Reversible electroporation is a phenomenon that pores formed on the lipid bilayer membrane 213 will reseal automatically after the electric field is removed. It occurs when the cell is exposed to a 214 relatively weaker electric field than irreversible electroporation. 10 The PI fluorescence intensity of 215 four cells under 14 V/2 μs/100 μs intermittent treatment shows that when the treatment is on (red 216 zoon, 1 s), the fluorescence increases, which means pore formation and PI dye inflow (Fig. 6a).

217
When the treatment is removed (gray zoon, 5 s), the fluorescence stops rising immediately, 218 suggesting that the pores close and the membrane regains its integrity after the treatment stops 219 (Fig. 6a). This kind of quick cell membrane recovery is a common phenomenon in reversible 220 electroporation, 13,14 but is hard to find in other kinds of membrane damages, such as direct 221 oxidation. Therefore, quick pore reseal is strong evidence for reversible electroporation. 222 Reversible electroporation is also tested using a double staining method with SYTOX Green 223 and PI, which are both cell impermeable stains that can only enter cells with compromised 224 membrane. 33 SYTOX Green is first added to the medium (Time point 1, Fig. 6b). After the NEEFT 225 is applied, perforated cells are stained with SYTOX Green and show green fluorescence ( Time  226  point 2, Fig. 6b). After 10 minutes, PI is added, which could only stain the cells that still have 227 compromised membrane. Thus, the cells that are not stained with PI are considered as having 228 reversible pores (Time point 3, Fig. 6b). With a relatively low applied voltage at 14 V (2 μs/100 229 μs/20,000 pulses), some already perforated cells could not then be stained with PI, indicating the 230 pores formed on the cell membrane are reversible (Fig. 6c). While under a high applied voltage at 231 80 V (1 μs/1 ms/10 pulses), almost all cell perforation is irreversible, since cells stained with 232 SYTOX Green are also stained with PI (Fig. S7). This phenomenon conforms to the feature of 233 electroporation, indicating that electroporation is the predominant mechanism for bacteria 234 inactivation in NEEFT.

246
Discussion on mechanisms 247 In this work, we for the first time show the bacteria inactivation by NEEFT at the single-cell 248 level and demonstrate the mechanism to be electroporation induced by the lightning-rod effect of 249 the nanowedges. Due to the lightning-rod effect, the electric field at the tips of metal rods with a 250 high aspect ratio will be greatly enhanced compared to that in bulk. Therefore, this strong electric 251 field could be sufficient to charge cell membrane, cause irreversible electroporation, and kill 252 bacteria even under lower applied voltages. 253 Although bench-scale NEEFT for water disinfection was developed based on this concept, the 254 mechanism was only supported by control experiments done with electrodes with/without 255 nanowire modifications. 19,23 There was no direct evidence confirming that the bacteria were 256 inactivated due to the nano-enhanced electric field and electroporation. The results achieved in this 257 study provide important evidence on the mechanisms. Firstly, only the bacteria in the area of the 258 nano-enhanced electric field are inactivated while others in bulk are intact (Fig. 1b & d). The 259 inactivation percentage shows a positive correlation with the strength of the nano-enhanced 260 electric field instead of the applied voltage (Fig. 4). Furthermore, when >90% bacteria inactivation 261 is achieved with NEEFT at 30 V/2 μs/100 μs, there is no significant ROS generation (Fig. 5a), 262 indicating this bacteria inactivation is not due to ROS damage. Reversible electroporation is 263 detected under relatively low applied voltage (Fig. 6), indicating that NEEFT could induce 264 electroporation, and irreversible electroporation causing bacteria inactivation will be dominant at 265 higher voltages. The rapidness of the cell damage (< 1 ms) also conforms with the property of 266 electroporation (Fig. 2a).

267
It is worth noticing that electric field enhancement by nanowedges is the same for both positive 268 and negative electrodes (Fig. S5 & S6). Consistently, all the bacteria inactivation phenomena 269 discussed above do not show a significant difference between positive and negative electrodes. In 270 an electrochemical disinfection study, significantly higher bacteria inactivation efficiency on the 271 anode was found compared to the cathode, suggesting that electrical reduction should not cause 272 the same level of cell damage as electrical oxidation. 34 Our positive control group for ROS 273 detection also confirms that (Fig. 5e). Therefore, the same phenomenon on both electrodes found 274 in this work indicates that electrical oxidation/reduction should not be the mechanism causing 275 bacteria inactivation. Therefore, electroporation is demonstrated as the predominant mechanism 276 causing bacteria inactivation in NEEFT.

278
Theoretical analysis: NEEFT versus Bulk EFT 279 Since electroporation is the predominant mechanism for bacteria inactivation in NEEFT, the 280 induced transmembrane voltage (TMV), which is the increased potential difference across the cell 281 membrane resulting from exposure to an external electric field, is analyzed theoretically using 282 finite element method to compare NEEFT and bulk EFT. Both the on-chip system like the one 283 used in this work (Fig. 7) and a 3D system with standing nanowire (Fig. S8) are simulated. Two 284 cells in NEEFT and bulk EFT respectively are compared, which is cell No. 1 located at the 285 nanowedge tip (Fig. 7a left), and cell No. 2 located between two electrodes without nanowedge 286 (Fig. 7a right). Two concentric spheres are built to represent the inner and outer surface of the 287 bacteria cell wall. 35 The diameter of the cell is 1 μm and cell wall thickness is 50 nm. 36 The 288 simulation results show that the voltage drop across the membrane, i.e., the electric field, is greatly 289 enhanced at cell No. 1 near the nanowedge tip (Fig. 7b left) compared to cell No. 2 (Fig. 7b right). 290 The maximum TMVs of the two cells show that with the same applied voltage, the cell No. 1 in 291 NEEFT located at the nanowire tip can achieve around 7.5 times higher TMV than cell No. 2 in 292 bulk EFT (Fig. 7c), indicating that much lower voltage could be applied to achieve the same level 293 of TMV on cells in NEEFT than bulk EFT. Potential applications and future studies of NEEFT 302 NEEFT effectively kill bacteria with mild treatment conditions without causing electrochemical 303 reactions or other side effects, making it suitable for high-quality sample processing, such as liquid 304 food or blood sample. The mild treatment conditions also make it a safe process, which is expected 305 to find medical applications, such as for wound healing. Furthermore, it is an energy-efficient 306 approach that is applicable for large-scale treatment processes, such as drinking water treatment. 307 Since it is a highly localized bacteria inactivation process, it is perfect for biofilm control, which 308 has potential applications in biofouling prevention. On the other hand, although NEEFT is not a 309 homogeneous process, we have shown that in a dynamic system, bacteria could be attracted to the 310 nanowedge tips where they then get inactivated. This targeted cell transportation occurred in 311 NEEFT makes it possible for broad applications in continuous flow systems. The as-shown rapid 312 cell damage and the effectiveness of both electrodes further improve its efficiency. Last but not 313 least, NEEFT should also work on other kinds of cells in addition to bacteria since electroporation 314 targets the lipid bilayer membrane. Therefore, it has much broader potential applications, including 315 intracellular molecule delivery and cell lysing for a broad range of cell types. 316 To enhance the applications of NEEFT in dynamic systems treating free-moving cells, it is 317 critical to transport the targeted cells to the effective zone, which is the tips of nanowedges or 318 nanowires. Therefore, a future study is to investigate cell transportation in NEEFT system, 319 including using electrical pulses with different parameters to control cell transport, using devices 320 or reactors with specific designs to increase the transport of targeted cells to the effective zoon, or 321 introducing baffles or increasing flow mixing to increase the probability of transporting cells to 322 the effective zoon. 323 324

325
Chip fabrication and pre-treatment 326 Glass wafer was used as the substrate for electrode deposition. Gold nanowedges were first 327 defined by electron beam lithography. Then, 200 nm gold layer was deposited using electron beam 328 evaporation and lift-off method. The gold bulk contact pads of 300 nm thickness were defined by 329 photolithography and fabricated by lift-off method ( Figure S1). There are 330 nanowedges on one 330 chip in total. The interval between nanowedges are 7 μm, which is to deploy a large number of 331 nanowedges without interference between each other on showing bacteria inactivation 332 phenomenon. The default nanowedge is 200 nm wide at the tip and 1 μm wide at the base in default 333 chips. Note that the chips designed for interval experiments (Fig. 4b and S6) have nanowedges of 334 200 nm width tip and 400 nm width base in order to achieve 800 nm interval. To achieve bacteria 335 immobilization on the chip, the chip was first washed and coated with poly-L-lysine (0.01%, mw 336 150,000-300,000). The detailed methods are stated in supplementary information.

338
Cell culture and harvest 339 S. epidermidis or B. subtilis were cultured in nutrient broth for 15 hours, and E. coli was cultured 340 in LB broth for 7 hours before use. For the immobilized cells, 4 mL bacterial solution was 341 centrifuged at 4000 rpm for 5 min. The supernatant was discarded, and the bacterial pallet was 342 resuspended in 1 mL 10 mM phosphate buffer. After 3 times of washing, the cell pallet was 343 resuspended in 0.5 mL 10 mM phosphate buffer to achieve the bacteria suspension with a higher 344 cell concentration. For the experiment with free-moving cells, bacteria were washed with DI water 345 for 3 times instead of phosphate buffer.

347
Experimental setup for immobilized bacteria cells 348 To conduct NEEFT with immobilized bacteria cells, add 40 μl of prepared bacteria suspension 349 onto a poly-L-lysine coated chip to cover the gap between two electrodes, then let the cells settle 350 down for 50 mins in room temperature. During this time, a layer of cells will be immobilized on 351 the chip surface. Then, the bacterial solution on the chip was gently washed away with 4 ml DI 352 water containing 15 μΜ Propidium iodide (PI) using a pipette, to remove the non-immobilized 353 cells and leave the immobilized bacteria in a drop of PI staining DI water. The chip was then 354 flipped, put onto a coverslip, loaded onto an inverted microscope for in situ observation (see 355 experimental setup in Fig. S2).

357
Experimental setup for free-moving bacteria cells 358 To visualize the bacteria inactivation process with free-moving bacteria, stain bacteria in the 359 prepared bacteria suspension with 15 μM PI and 5 μΜ SYTO 9 (Thermo Fisher Scientific) for 5 360 min before use. Αdd 40 μl of stained bacteria suspension onto the chip to cover the gap between 361 two electrodes. Then the chip was flipped, put onto a coverslip, and loaded onto an inverted 362 microscope for in situ observation (see experimental setup in Fig. S2).

364
NEEFT procedures 365 The pulsed voltages were applied to the chip using a pulse generator (Avtech Electrosystems 366 LTd.) which is triggered with a waveform generator (Keysight 33509B). The typical pulses used 367 in this work have 2 μs pulse width, 100 μs period (10 kHz), 500,000 pulses, corresponding to 1 s 368 effective treatment time and 50 s total time, unless stated otherwise. The effective treatment time 369 is the total time when the applied voltage is not 0, which equals pulse width × pulse number. The 370 pulse width of 2 μs is used to minimize electrochemical reactions. The pulse waveform was 371 measured using an oscilloscope (Keysight InfiniiVision 6000 X-series). 372 373

ROS detection 374
For ROS detection experiments, the ROS indicator DCFH-DA was added to stain the bacteria 375 at 0.2 mM during the bacteria immobilizing process for 50 min. After staining, DCFH-DA was 376 washed away with DI water. To ensure that this method is able to detect ROS generation, pulses 377 with longer pulse width (200 μs/10 ms) at 20 V were tested as positive control. Significant 378 fluorescence near the positive electrode was observed, indicating this method is valid for ROS 379 detection. 380 In DMSO test, DI water containing 15% DMSO and 15 μΜ PI was used as the medium to 381 quench ROS and protect bacteria form ROS damage. 382 383 Reversible electroporation tests 384 For reversible electroporation tests, 5 μM SYTOX Green was first added to the medium before 385 treatment. Electrical pulses of 14 V/2 μs/100 μs/20,000 pulses were applied, and the microscopy 386 images were collected. Ten minutes after the electrical treatment, 15 μM PI was then added, and 387 the images were collected again.

389
Microscope observation 390 The NEEFT treatment process was observed and recorded in situ using an inverted fluorescence 391 microscope (Zeiss Axio Observer 7). Cell and nanowedge images were captured via DIC channel.

392
PI was excited at 488 nm. Syto 9, and SYTOX Green, and DCFH-DA were excited at 555 nm, 393 respectively. In supporting video 2, 3 and 4, the electrode and nanowedges are visualized using 394 reflection channel with 555 nm incident light. All fluorescent signals are filtered with a 90 HS 395 filter. The video taking was triggered by the Keysight 33509B waveform generator.

397
Image processing and data statistics 398 The microscopy images were processed using MATLAB. The fluorescence image of PI signal 399 before treatment was subtracted from the image after treatment, which only keeps the changing of 400 PI signal. The subtracted image was then processed for inactivation and counting analyzing. 401 The bacteria inactivation percentage is represented as the percentage of nanowedges that have 402 inactivated cell at the tip, which is 403 There are 330 nanowedges on one chip for the default design. Each treatment experiment was 405 repeated with three chips, and the error bars show the standard deviation of the three repeated 406 experiments. Note that for the chips that have 0.8 um interval nanowedges, cell number at the 407 nanowedge tip is less than nanowedge number due to the small interval. Bacteria inactivation is 408 represented as the percentage of dead cells, which is 409 (%) = . 410

411
Electric field and TMV simulation 412