Cell preparation, experimental, and analysis procedures

Cells were immersed in buffers at different conductivities (presented in Table ​Table2).2). The procedure was as follows: 1: Washing cells for 3 times with HCB (osmolarity was adjusted to 290 mOsm/L with sucrose and dextrose³⁴), where each cycle includes centrifugation (at 300×g for 5 min), supernatant removal, resuspension in HCB, consecutively. 2: Trypan blue (1:1, Interlab/ dist. Sigma Aldrich, TR) staining for cell viability check (> 98%) and the determination of final concentration (5 × 10⁵ cells/mL) by an automated cell counter (TC20, Bio-Rad Laboratories Inc.) (Supplementary material, Sect. 5.1, Fig. ^S3).

The experimental setup is presented in Fig. ^S4 (Supplementary material, Sect. 5.2). The microchannel was flushed with ethanol and DI water, respectively. To decrease the nonspecific attachment of cells to the parylene surface, poly(l-lysine)-[g]-poly(ethylene glycol) (PLL(20)-g[3.5]-PEG(2), 100 µg/mL in DI water) (SuSoS, Switzerland) coating was applied by treating the device for 1 h at room temperature³⁵. Next, the channel was washed with DI water and filled with buffer. Washing were performed at a flow rate of 7.5 µL/min while PLL(20)-g[3.5]-PEG(2) coating and cell analysis were carried out at 1 µL/min by using a coupled pressure generator/mass flow controller with a flow rate sensor (OB1 and BFS, Elveflow, France). DEP-D unit was energized with a signal generator (Agilent, 81150A). Two of its four outputs were utilized at 20V_pp with a 180° phase difference. Impedance measurement at IM-C units was performed using an impedance spectroscope (HF2IS, Zurich Instruments, Switzerland) with the application of 2V_PP sinusoidal voltage at 100 kHz. Measurement duration was 2 min for each analysis. Nonspecific attachment of cells was determined at the beginning of each experiment while keeping DEP-D unit inactive (i.e., no voltage was applied in DEP-D unit).

Time-based impedance data were collected at IM-C units using HF2IS software (ziControl, Zurich Instruments). We applied baseline removal and peak detection functions in MATLAB to extract cell count information from these data. 100Ω was chosen as the threshold of impedance magnitude for cell detection. In combination with the cell detection algorithm, a cross correlation function, XCORR, was also run since there were two different time-dependent sequences: impedance magnitude values obtained in IM-C units at the entrance and outlet of LOC. XCORR detected the similarity between IM-C 1 and IM-C 2 and determined the delay in time between them by examining the correlation between two impedance sequences and estimating the time at which maximum correlation occurs³⁶. Employing XCORR, the time delay between two IM-C units, inherently formed due to the duration that cells spend in the DEP-D unit between IM-C units (Supplementary material, Sect. 5.3, Fig. ^S5), was identified and eliminated. , This method prevents misinterpretation in the quantification of incoming and outgoing cell counts.

DEP responses of cells were quantified by defining a trapping ratio, which was calculated by the following formula: