Assessment of Submicron Particle Zeta Potential in Simple Electrokinetic

Electrophoresis 2019, 40, 1395–1399 1395

Samuel Hidalgo-Caballero1,2 Short Communication Cody Justice Lentz1 Blanca H. Lapizco- Encinas1 Assessment of submicron particle zeta

1Microscale Bioseparations potential in simple electrokinetic Laboratory and Biomedical Engineering Department, microdevices Rochester Institute of Technology, Rochester, NY, USA The present communication illustrates the use of simple electrokinetic devices for the assessment of the zeta potential of submicron polystyrene particles. A combination of 2Facultad de Ciencias Fısico´ Matematicas,´ Benemerita´ manual and automatic particle tracking was employed. This approach allows for charac- Universidad Autonoma´ de terizing particles in the same conditions and devices in which they can be separated, e.g. Puebla, Puebla, Mexico´ dielectrophoretic separations; making the resulting data readily applicable.

Received October 9, 2018 Keywords: Revised November 8, 2018 Electrical charge / Electrokinetics / Electrophoresis / Submicron particles / Zeta Accepted November 21, 2018 potential DOI 10.1002/elps.201800425 Additional supporting information may be found online in the Supporting Infor- mation section at the end of the article.

Particle migration is an important research area in microﬂu- White et al. [8] employed CE experimentation to determine

idic devices, in particular, when working with electric ﬁeld the ␨p of polystyrene particles in order to assess particle con- driven techniques, one crucial property is the particle zeta ductivity and predict dielectrophoretic behavior. Other stud-

potential (␨p ). This parameter accounts for the electrical ies have been focused on characterizing the relationship be- charge present on a particle, as it characterizes the electrical tween electrophoretic migration and particle size [7, 9] with

double layer (EDL) around the particle [1]. Particle zeta simultaneous determination of ␨p and zeta potential of the potential determines the electrophoretic mobility (␮EP)and channel surface (␨w) [1, 10]. Recognized research groups in the electrophoretic migration of a particle. Differences in the ﬁeld of CE have dedicated considerable attention to the

particle electrophoretic migration are widely exploited in measurement and prediction of ␮EP.TheGasˇ group devel- analytical electrokinetic separations, such as CE, capillary oped a sophisticated software package called PeakMaster for

electrochromatography, isotachophoresis, isoelectric focus- the prediction of ␮EP of analytes of interest [11]. The Kasiˇ ckaˇ ing, and dielectrophoresis [2–4]. Moreover, most of the group has studied the determination of ␮EP for a wide array of techniques mentioned above can be enhanced by employing analytes [12]. Our research group has analyzed the electroki- liquid metal electrodes [5, 6] that provide a better control of netic migration of micron-sized particles and cells by using the electric ﬁeld while avoiding the divergence that usually particle image velocimetry (PIV) [13,14]. The methodology re- appears in the case of thin electrode surfaces. ported here allows extending this approach, that only requires

Characterizing the ␨p and ␮EP allows assessing particle simple devices, to the assessment of the ␨p of nanoparticles. surface and morphological properties, which can be related This study presents the experimental analysis of ␨p for to important biological attributes [7]. Furthermore, knowing 12 distinct types of submicron polystyrene particles (diam-

the ␨p enables a better selection of the proper separation tech- eter from 100–500 nm). A combination of PIV and current nique to be used for a given particle sample or application. monitoring was employed to characterize particle electroki- Signiﬁcant efforts have been devoted to the characterization netic migration in microchannels made from PDMS. Particle

of ␨p and ␮EP for microparticles and nanoparticles [1, 7–10]. tracking was an essential step that had to be modified when assessing the smaller particles (100–200 nm) in our study, due to the diffraction limit for an optical microscope [15, 16]. ␨ Correspondence: Professor Blanca H. Lapizco-Encinas, Mi- This study demonstrates that the characterization of the p of croscale Bioseparations Laboratory, Rochester Institute of Tech- submicron particles is possible in simple microfluidic chan- nology, Institute Hall (Bldg. 73), Room 3103, 160 Lomb Memorial nels without the need of specialized equipment, such as a zeta Drive, Rochester, NY 14623, USA analyzer. The data generated in this study provides a valuable E-mail: [email protected] tool for designing new electric-field driven microfluidic sys-

Abbreviations: EDL, electrical double layer; EK, electrokinetic; EO, electroosmosis; EP, electrophoresis; iDEP, insulator- based dielectrophoresis; PIV, particle image velocimetry Color online: See article online to view Figs. 1 and 2 in color.

C 2018 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.electrophoresis-journal.com 1396 S. Hidalgo-Caballero et al. Electrophoresis 2019, 40, 1395–1399

Figure 1. (A) Schematic representation of the microchannel including the win- dow of study. For illustration purposes, a particle with negative charge is depicted in this image; therefore the EP migration is toward the inlet (left). Sample is introduced at the inlet, where the posi- tive electrode is located, the ground electrode is at the outlet. (B) Picture of a 500 nm red particle sample before application of an electric potential. (C) Trace lines of 500 nm red particles under an applied potential of 25 V. (D) Particle velocity as function of the electric field for particles 300–500 nm diameter. Legend: 320 nm particles (continuous line) and 500 nm particles (dashed line). tems for particle analysis and separation. In particular, our EP is the movement of particles relative to the suspension research group works with insulator-based dielectrophoresis medium, under the influence of an electric field. (iDEP) that combines electrophoresis (EP), electroosmosis According to the Helmholtz-Smoluchowski Equation [18] (EO), and DEP for the manipulation of nano and microparti- the electroosmotic velocity is given by: cles, including bioparticles such as DNA, proteins, virus and ε ␨ v = ␮ =− m w cells. Additionally, the results from this study are essential for EO EO E ␩ E (1) any type of computational modeling or applications used for ␮ ε predicting the behavior of particles in microscale electroki- where EO is the electroosmotic mobility, m is the media per- mittivity, ␨w is the wall zeta potential, ␩ is the media viscosity netic (EK) systems, such as iDEP devices. It is expected that similar assessments will be carried out for the characteriza- and E is the local electric field. For particles, the Helmholtz- tion of the zeta potential of biological submicron particles. Smoluchowski Equation is valid when the particle radius is

Electrokinetic phenomena are a consequence of a polar- much greater than the Debye length (␭D). The other limit ization mechanism at the interface between a solid and a is the Huckel¨ approximation which is valid when the parti- liquid which creates an EDL [17]. The ﬁrst layer (Stern layer) cle radius is equal to or smaller than ␭D. None of these two comprises the surface charge generated by ion adsorption on limits are the case for the particles in this study. Taking into the solid surface itself. The second layer (Helmholtz layer) account the low ionic strength of the suspension medium, is composed of mobile counterions diffusely attracted to the the Debye length under these conditions was estimated as ␭ = . ␬ = ␭−1 surface charge that give rise to two important phenomena: D 6 1 nm ( D ); considering a as the particle radius, EO and EP. Electroosmosis is the motion of a ﬂuid, while this produces ␬a values between 8 and 41 for the particles

C 2018 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.electrophoresis-journal.com Electrophoresis 2019, 40, 1395–1399 Miniaturization 1397

Ta b l e 1 . Properties of the 12 fluorescent polystyrene particles analyzed in this study, including mobility (␮EK and ␮EP), particle zeta 2 potential (␨p), and fitting correlation (R ) values for the vEK versus electric field data

−10 2 Sample Particle Color Brand Surface ␮EK ±1 × 10 ␮EP ±1 × ␨p ±2 R values 2 −1 −1 -9 2 -1 -1 size (nm) functionality (m V s ) 10 (m V s ) (mV) vEK vs. E

1 500 Red Invitrogen Carboxyl 1.6 ±0.01 × 10−8 –6.0 × 10−8 –81 0.9984 2 500 Green Magsphere Carboxyl 1.1 × 10−8 –6.5 × 10−8 –87 0.9940 3 500 Red Magsphere Carboxyl 1.8 × 10−8 –5.8 × 10−8 –78 0.9775 4 320 Green Magsphere Carboxyl 1.2 × 10−8 –6.3 × 10−8 –88 0.9997 5 200 Green Invitrogen Carboxyl –4.4 × 10−9 –8.0 × 10−8 –116 0.9414 6 200 Orange Invitrogen Carboxyl –1.3 × 10−8 –8.9 × 10−8 –128 0.9868 7 200 Red Invitrogen Amine 2.6 × 10−8 –0.9 × 10−8 –72 0.9916 8 200 Red Magsphere Amine 2.4 × 10−8 –5.1 × 10−8 –72 0.9960 9 100 Red Invitrogen Carboxyl 7.8 × 10−9 –6.8 × 10−8 –106 0.9862 10 100 Green Invitrogen Carboxyl 9.4 × 10−9 –6.6 × 10−8 –104 0.9952 11 100 Green Magsphere Carboxyl 4.7 × 10−9 –7.1 × 10−8 –111 0.9921 12 100 Green Magsphere Amine 8.3 × 10−9 –6.8 × 10−8 –106 0.9979

An analysis of error estimation for the ␨pvalues, which was calculated as ±2 mV, is included in Section 4 of the Supporting Information.

in this study (100–500 nm diameter). For these intermediate particles with diameters from 100–500 nm were studied (Ta- conditions [8], the function proposed by Henry [19], for which ble 1). Particle suspensions ranged in concentration from Ohshima developed an approximation [20] applies: 3.4–7.3 × 108 particles/mL. The suspending media was made ε from DI water with 0.05% v/v Tween 20, to reduce particle 2 m␨ p ␮EP = f (␬a) (2) adhesion, and addition of 0.1 M KOH, to produce a pH of 3␩ 6.0–6.5 and a conductivity of 25.3 ± 0.1 ␮S/cm. Microparti- ⎡ ⎤ cle behavior was observed with a Leica DMi8 (Wetzlar, Ger- ⎢ 1 ⎥ many) inverted microscope and direct current (DC) electric f (␬a) = ⎣1 + ⎦ (3) . 3 potentials were applied with a high voltage supply (Model 2 5 2 1 + −␬ ␬a 1+2e a v ( ) HVS6000D, LabSmith, Livermore, CA). Briefly, EK for each Thus, the expression for the electrophoretic velocity for particle type was measured with PIV at different electric fields ␮ these particles becomes: and this data was used to obtain the EK values for all particles. EOF was characterized independently [22] in terms ε 2 m␨ p ␮ ␨ , ␮ ␨ v = ␮ = (␬ ) of EO and w then, EP and p were estimated by using EP EP E ␩ f a E (4) 3 Eqs. (2)–(4). Particles suspended in a medium inside a microchannel For the larger particles in our study, 300–500 nm di- (Fig. 1A) experience EP and EO, thus, their total EK velocity ameter (Table 1, samples 1–4), the computational software v is: TrackMate [24] was utilized for assessing EK. The imaging of particles in this larger size range resulted in easily distin- v = ␮ = ␮ + ␮ EK EK E ( EO EP) E (5) guishable particles with reasonable particle–particle spacing Particle image velocimetry is a computational tool that ap- (Fig. 1B). To ensure accurate particle tracking, it was verified plies image processing methods to assess the motion of par- that the trace lines generated by TrackMate had a consis- ticles contained in a region of interest. Classical algorithms tent direction and ran parallel to the electric field (Fig. 1C). for this purpose are well described in [21]; however, most Three to four videos were produced for each one of the volt- of them require good visualization of individual particles. A ages studied. To verify accuracy, one of the videos for each manual velocimetry approach is required when the particle voltage was analyzed using manual tracking and compared diameter is close to the diffraction limit for an optical micro- to results from TrackMate, and if the difference between scope [16]. Current monitoring is a technique developed to average velocities was less than 5%, TrackMate was used ␮ ␨ for the rest of the analysis. These results are included in estimate EO of a liquid and the w of a substrate. The zeta potential of the PDMS channels employed was characterized Fig. 1D as a graph of vEK versus E. A trend line was added for as ␨w = -97±1 mV employing a refined current monitoring each particle and the resulting slope corresponds to the ␮EK approach [22]. (Table 1). The R2 values obtained are all above 0.97, demon- Experiments were conducted in microchannels made strating the accuracy of this technique. The y-intercepts of from PDMS employing standard soft lithography techniques the trend lines were not set to zero (Supporting Information [23]. The microchannels were 10.16 mm long, 0.88 mm wide, Table S1) as they account for any undesired pressure driven 40 ␮m deep, and contained inlet and outlet liquid reservoirs flow that can develop during experimentation; as EOF causes (Fig. 1A), all internal channel surfaces are PDMS and have the a higher pressure at the exit reservoir and therefore a slower same ␨w. A total of 12 distinct types of fluorescent polystyrene particle velocity.

C 2018 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.electrophoresis-journal.com 1398 S. Hidalgo-Caballero et al. Electrophoresis 2019, 40, 1395–1399

Figure 2. (A) Schematic illustration of the diffraction limit of submicron particles, which applies to all particles with diameters less than 300 nm in this study. (B) Picture of 200 nm red particle mixture before application of an electric field. (C) Trace lines generated by manual tracking of 200 nm red particles under a potential of 30 V. (D) Particle velocity as function of the electric field for particles 100–200 nm diameter. Legend: 100 nm particles (continuous line), 200 nm particles moving opposite to the field (dashed line), and 200 nm particles moving with the field (dotted line).

Optical microscopy is widely used in biological research; nonimmersed objectives NA≤ 1and␭ =∼ 500 nm, then however, its resolution is limited by a physical phenomenon d ≥ 250 nm [16]. This means that particles with a diame- called diffraction, which is the spreading of a wave front when ter less than 250 nm are difficult to distinguish and only a it encounters an aperture, resulting in secondary wave fronts cloudy visualization is achieved (Fig. 2B). Therefore, a manual that interfere. A cartoon illustrating the diffraction limit is in- tracking approach was applied to particles 100–200 nm diam- cluded in Fig. 2A. If the aperture is circular, an Airy’s pattern eter (samples 5–12 in Table 1); by following the migration is formed and it is the origin of the diffraction limit for an op- paths (Fig. 2C) between each frame of the video recordings at tical microscope, which was first studied by Abbey [25]. The rates between 6 and 25 fps depending on the exposure time. light intensity is also deformed by diffraction as depicted in Figure 2D illustrates these results of vEK versus E, where two the bottom part of Fig. 2A. The minimum resolution (Dmin)is types of particles (samples 5 and 6, Table 1) moved in the op- obtained from the Fraunhofer Equation as [26]: posite direction of the electric field because both have negative −8 2 −1 −1 −8 2 −1 −1 ␮EP (−8.0 × 10 m V s and -8.9 × 10 m V s ,re- Dmin = ␭/(2 · NAOBJ )(6) spectively) that are greater in magnitude than the ␮EO (7.6 × −8 2 −1 −1 ␮ where ␭ is the wavelength of the exciting light and NAOBJ 10 m V s ). Thus, their resulting EK are both negative is the numerical aperture of the microscope objective. For (Table 1), i.e., toward the inlet.

C 2018 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.electrophoresis-journal.com Electrophoresis 2019, 40, 1395–1399 Miniaturization 1399

A slight inverse correlation between particle size and ␨p [5] So, J.-H., Dickey, M. D., Lab Chip 2011, 11, 905–911. can be seen for carboxylated and aminated particles (Table 1, [6] Tang, S.-Y., Zhu, J., Sivan, V., Gol, B., Soffe, R., Zhang, Fig. 1D and Fig. 2D), with the exception of particles 5–6. W., Mitchell, A., Khoshmanesh, K., Adv. Funct. Mater. The inverse correlation is most likely due to a difference 2015, 25, 4445–4452. in the manufacturing of the particles and is in contrast to [7] Palanisami, A., Miller, J. H., Electrophoresis 2010, 31, Vorwerg et al. [9] simply because the surface charge den- 3613–3618. sity was not held constant for this study. Other studies with [8] White, C. M., Holland, L. A., Famouri, P., Electrophoresis particles of similar sizes have shown no correlation between 2010, 31, 2664–2671. particle size and ␨p [8]. Another remarkable point is that the [9] Vorwerg, L., Antonietti, M., Tauer, K., Colloids Surf. A resulting values of the ␨p for particles with same size and 1999, 150, 129–135. same type of functionalized group, but different manufac- [10] Sureda, M., Miller, A., Diez, F. J., Electrophoresis 2012, turer, have a variation of only ten percent from its average 33, 2759–2768. (Table 1). [11] Jaros,ˇ M., Soga, T.,van de Goor, T., Gas,ˇ B., Electrophore- The results from this work demonstrate that the zeta sis 2005, 26, 1948–1953. potential of submicron particles can be accurately measured [12] Solˇ ınov´ a,´ V., Kasiˇ cka,ˇ V., Sazelov´ a,´ P., Barth, T., Miksˇık,´ I., using simple microfluidic devices without the need of spe- J. Chromatogr. A 2007, 1155, 146–153. cialized equipment. By employing this approach, particles are [13] Martınez-L´ opez,´ J. I., Moncada-Hernandez,´ H., Baylon- characterized under the same conditions and type of devices Cardiel, J. L., Martınez-Chapa,´ S. O., Rito-Palomares, M., in which they will be separated, e.g. dielectrophoretic separa- Lapizco-Encinas, B. H., Anal. Bioanal. Chem. 2009, 394, 293–302. tions; so this data is readily applicable. The results from this study have many potential applications for predicting EK be- [14] Garza-Garcıa,´ L. D., Perez-Gonz´ alez,´ V. H., Perez-´ Sanchez,´ O. A., Lapizco-Encinas, B. H., Chem. Eng. Tech- havior and are currently being used in our group for the math- nol. 2011, 34, 371–378. ematical modeling of particle behavior in iDEP devices, where [15] Born, M., Wolf, E., Principles of Optics: Electromagnetic knowing a priori the particle zeta potential values is essential. Theory of Propagation, Interference and Diffraction of Similar work can be done on submicron biological particles Light, Elsevier Science, Elmsford, NY, USA 2013. to create a library that is useful for selecting the appropri- [16] Meinhart, C. D., Wereley, S. T., Measurement Sci. Tech- ate separation technique for samples containing virus, cells nol. 2003, 14, 1047. organelles, etc. Depending on the experimental conditions [17] Jones, T. B., Electromechanics of Particles, Cambridge and particle characteristics, the Helmholtz–Smoluchowski or University Press, New York, USA 1995. Huckel approximations could also be applied, modifying only ␨ [18] Probstein, R. F., Physicochemical Hydrodynamics: An In- Eqs. (2) and (3) for the determination of p , and employing troduction, John Wiley & Sons, Hoboken, New Jersey, the same methodology. Nonetheless, if the particle volume USA 2005. fraction is more than 1%, the Stokes limit does not produce [19] Henry, D. C., Proc. R. Soc. Lond. Ser. A 1931, 133, accurate results and other models for the drag force must be 106–129. used [27–29]. [20] Ohshima, H., J. Colloid Interface Sci. 1994, 168, 269–271. [21] DuBose, J., Lu, X., Patel, S., Qian, S., Woo Joo, S., Xuan, The authors acknowledge the financial support provided by X., Biomicrofluidics 2014, 8, 014101. the National Science Foundation (CBET-1705895), and the [22] Saucedo-Espinosa, M. A., Lapizco-Encinas, B. H., Biomi- Mexican National Council on Science and Technology (CONA- crofluidics 2016, 10, 033104. CyT) in the form of a scholarship (REG-849369). [23] Duffy, D. C., McDonald, J. C., Schueller, O. J. A., White- sides, G. M., Anal. Chem. 1998, 70, 4974–4984. The authors have declared no conflict of interest. [24] Tinevez, J.-Y., Perry, N., Schindelin, J., Hoopes, G. M., Reynolds, G. D., Laplantine, E., Bednarek, S. Y., Shorte, References S. L., Eliceiri, K. W., Methods 2017, 115, 80–90. [25] Mccutchen, C. W., JOSA 1967, 57, 1190–1192. [1] Yan, D., Yang, C., Nguyen, N.-T.,Huang, X., Electrophore- [26] Tierney, J., Rader, C. M., Gold, B., IEEE Trans. Audio sis 2006, 27, 620–627. Electronic. 1971, AU-19, 48–57. [2] Breadmore, M. C., Electrophoresis 2007, 28, 254–281. [27] Richardson, J. F., Zaki, W. N., Chem. Eng. Sci. 1954, 3, [3] Polniak, D. V., Goodrich, E., Hill, N., Lapizco-Encinas, B. 65–73. H., J. Chromatogr. A 2018, 1545, 84–92. [28] Kim, S., Russel, W. B., J. Fluid Mech. 1985, 154, 269–286. [4] Sautrey, G., Duval, R. E., Chevalley, A., Fontanay, S., [29] Van Der Hoef, M. A., Beetstra, R., Kuipers, J. A. M., J. Clarot, I., Electrophoresis 2015, 36, 2630–2633. Fluid Mech. 2005, 528, 233–254.

C 2018 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.electrophoresis-journal.com