You are on page 1of 7

Analytica Chimica Acta 584 (2007) 129135

Characterization of electrokinetic gating valve in


microfluidic channels
Guiseng Zhang a , Wei Du a , Bi-Feng Liu a, , Hideaki Hisamoto b , Shigeru Terabe b
a

The Key Laboratory of Biomedical Photonics of MOE, Hubei Bioinformatics & Molecular Imaging Key Laboratory, Department of Systems Biology,
College of Life Science & Technology, Huazhong University of Science and Technology, Wuhan 430074, China
b Graduate School of Material Science, University of Hyogo, Hyogo 678-1297, Japan
Received 28 August 2006; received in revised form 18 October 2006; accepted 29 October 2006
Available online 7 November 2006

Abstract
Electrokinetic gating, functioning as a micro-valve, has been widely employed in microfluidic chips for sample injection and flow switch.
Investigating its valving performance is fundamentally vital for microfluidics and microfluidics-based chemical analysis. In this paper, electrokinetic
gating valve in microchannels was evaluated using optical imaging technique. Microflow profiles at channels junction were examined, revealing
that molecular diffusion played a significant role in the valving disable; which could cause analyte leakage in sample injection. Due to diffusion,
the analyte crossed the interface of the analyte flow and gating flow, and then formed a cometic tail-like diffusion area at channels junction. From
theoretical calculation and some experimental evidences, the size of the area was related to the diffusion coefficient and the velocity of analytes.
Additionally, molecular diffusion was also believed to be another reason of sampling bias in gated injection.
2006 Elsevier B.V. All rights reserved.
Keywords: Electrokinetic gating; Microfluidic chip; Valving; Leakage

1. Introduction
Over the past decade, analysis on a microfluidic device has
attracted great attention, which leads to currently fast-growing
community of miniaturized total analysis system (TAS) [14].
Due to high-throughput potential, capabilities of miniaturization, integration and automation that represent the development
direction of next-generation instrumentation, TAS has been
recognized as a powerful tool in far-reaching application fields
[47], such as chemistry, biology, medicine, environment, and
pharmaceutics as well.
In TAS, there are several key modules or elements that functionally achieve microfluidic operations including actuating,
pumping, valving or switching, mixing, sensing and separating etc., which form the solid basis of so-called lab-on-a-chip.
Thus, to give a deep insight into these elements is fundamentally vital, as it will greatly help us to understand the mechanism
of microfluidics and further to construct the state-of-the-art

Corresponding author. Tel.: +86 27 8779 2170; fax: +86 27 8779 2203.
E-mail address: bfliu@mail.hust.edu.cn (B.-F. Liu).

0003-2670/$ see front matter 2006 Elsevier B.V. All rights reserved.
doi:10.1016/j.aca.2006.10.046

microfluidic platform for analysis. Valve plays an important


role in microfluidics, e.g., flow control in microchannels. Much
attention has been paid to valving technology in microfluidic
chip. Basically, there are several major kinds of micro-valves
in category, such as mechanical valve [812], magnetic valve
[13], thermo valve [1417], and electrokinetic valve [1830],
etc. Mechanical valving is accomplished with mechanical design
to open or close the entries of a microchannel.
Electrokinetic valving [1830] is a microfluidic scheme that
functions through manipulating the distribution of electric fields
in microchannels to prevent a certain flow from entering the
target channel. Although it is not a physical module, in the
viewpoint of function it is also categorized as a valve. Due to
its simplicity and easy-to-manipulate, electrokinetic valving has
been widely used for sample injection on a chip, e.g., pinching
[1823] that has been detailedly characterized in both theoretical
simulations [22,23] and experiments [1821]. Gating [2430] is
another widely used electrokinetic valve for sample injection
in microchip-based electro-separations such as capillary electrophoresis [26,27] and capillary electrochromatography [29].
It can conveniently achieve repeated sample introductions with
high reproducibility [28]. Furthermore, gating valve shows great

130

G. Zhang et al. / Analytica Chimica Acta 584 (2007) 129135

convenience for flow switch between channels in multidimensional separation on a microchip [31]. However, it has seldom
been experimentally investigated in detail, for example, sampling bias or leakage. Ermakov et al. [30] demonstrated a
computational simulation of gating valving. Processes of sample
loading and dispensing were theoretically calculated. Recently
Slentz et al. [29] discussed sample bias at microchannel junctions in gating injection. They found a new type of sampling bias
beyond electrokinetic discrimination, called transradial electrokinetic selection (TREKS). Based on this phenomenon, the
authors further suggested a diffusion injection method to correct such a sampling bias by TREKS. But it was still unclear the
underlying mechanism TREKS occurred.
Leakage is a significant criteria to evaluate the quality of
a valve. Zero leakage is required for a reliable valve. Since
electrokinetic gating functions based on microfluidic interaction, not actual physic barrier, leakage shows different features
from mechanical or magnetic valves simulated by Ermakov et
al. [30]. To reveal the underlying mechanism, it is fundamentally
important to examine the transportation behaviors of molecules
at microchannels junction.
For this consideration, electrokinetic gating valve in
microfluidic channels was characterized in this study. The
microflow profiles at channels junction were monitored and analyzed using fluorescence imaging technique. It was suggested
that molecule diffusion was the key factor of the leakage of
electrokinetic valve, and also cause molecular discrimination in
gated injection.

Japan). The microchip was fabricated using standard photolithograph, wet chemical etching, and bonding techniques as
previously described [32]. The channel was 80 m in width on
the top and 40 m in depth. For new microchip, channels were
washed sequentially with methanol, water, 1 M NaOH, 0.1 M
HCl, 0.1 M NaOH, water and running buffer at room temperature. Between analysis, channels were rinsed in an order of 0.1 M
NaOH, water and running buffer to ensure the reproducibility.
2.4. Electrokinetic gating
In gated injection or flow switch in two-dimensional separation on a microchip, a microflow initially carried the analytes
from Channel S to Channel SW. To prevent the analytes from
leaking into Channel BW, another microflow from Channel B
was introduced into Channel SW and Channel BW simultaneously, as shown in Fig. 1A. When the microflow from Channel B
was stopped, the analytes were then injected from Channel S into
Channel BW (Fig. 1B). Afterwards, the microflow from Channel B was recovered (Fig. 1C). The analytes were gated again.
Microfluidic program above could be conveniently achieved by

2. Experimental
2.1. Materials
Fluorescein isothiocyanate (FITC) was obtained from
SigmaAldrich (MO, US). Rhodamine series dyes including
rhodamine 6G, rhodamine B and sulforhodamine were purchased from TCI Chemicals (Tokyo, Japan). All other reagents
were of analytical grade. Water purified by the Milli-Q system
(Millipore, US) was used for the preparations of all solutions.
2.2. Instrument
Visualization and imaging of flow profiles at the microchannel junction were carried out on an inverted fluorescence
microscope (Eclipse TS 100-F, Nikon, Tokyo, Japan) equipped
with a 3CCD digital camera (HV-D28S, Hitachi Kokusai Electric Inc., Tokyo, Japan). Acquired fluorescent images were
further analyzed using Scion Image software package (NIH,
US). Power supply (Matsusada Precision Devices, Japan) for
flow pumping was computer-controlled with terminal emulation
software KTX (freeware) connected to a D/A converter (Nippon
Filcon, Inagi, Japan) via an RS-232C serial interface.
2.3. Experimental procedures
The glass microchip with a channel pattern of a cross was
kindly obtained from Kitamori group (University of Tokyo,

Fig. 1. Schematic diagram of electrokinetic gating valve for sample injection.


(A) Gating, the sample was transported from Channel S to Channel SW, and
was gated by a buffer from Channel B. (B) Injection, valve was open for sample
injection when the buffer flow from Channel B was hold. (C) Gating, sample
stream was gated again for preventing sample from leakage into Channel BW.

G. Zhang et al. / Analytica Chimica Acta 584 (2007) 129135

131

analytes. However, it was not the case in practice. Even when


g 1, the analytes could leak into Channel BW. Fig. 2 illustrated
the leakage condition of g = 1.2. In Fig. 2A, the analyte (FITC)
flow was delivered from Channel S to Channel SW, which was
electrokinetically gated by a gating flow (rhodamine B solution)
from Channel B as shown in Fig. 2B. Two sets of optical filters
were employed to match the fluorescence excitation and emis-

Fig. 2. Analyte leakage at g = 1.2. FITC was employed as the analyte dissolved
in buffer of 20 mM PBS (pH 8.5), and rhodamine B was added into buffer
solution for fluorescent visualization. (A) Micrograph using filters for FITC.
(B) Micrograph using filters for rhodamine B.

modulating the electric field strengths (EFS) in microchannels.


The currents in horizontal or vertical vector were kept constant,
that was, Is = Ibw and Ib = Isw . The parameter Ib /Is representing
the degree of segmenting the analytes from Channel BW, was
defined as gating factor (g).
3. Results and discussion
Theoretically, gating factor g 1 meant that the gating valve
was totally shut off, and no analyte would enter Channel BW.
And on the contrary if g < 1, Channel BW would be open for

Fig. 3. Electropherogram of four fluorescent species using capillary electrophoresis on a microchip. Buffer, 20 mM PBS (pH 8.5). Electric field strength
for separation, 330 V cm1 . Gated injection, 330 V cm1 for 0.2 s. Peak identity,
1, rhodamine 6G; 2, rhodamine B; 3, sulforhodamine; 4, FITC.

Fig. 4. Profiling of fluorescence intensity at the entrance of Channel SW. (A)


Intensity was measured along the dashed line. (B) Intensity profiling was corrected to eliminate the influence of channel geometry. (C) Calibrated intensity
profiling of fluorescent species at g = 3.0. (D) Calibrated intensity profiling of
fluorescent species at g = 2.0. Profiling curves identity: 1, rhodamine 6G; 2,
rhodamine B; 3, sulforhodamine; 4, FITC.

132

G. Zhang et al. / Analytica Chimica Acta 584 (2007) 129135

sion of FITC and rhodamine B so that only one flow could be


visualized in each micrograph. Since g 1.2, FITC should not
enter into Channel BW according to the theoretical calculation.
However, a large amount of FITC molecules actually leaked
into Channel BW (Fig. 2A), while the part of the gating flow
functioning for gating was irrigated into Channel SW as the
theoretical calculation predicted (Fig. 2B).
We thought that it might be concerned with molecular behaviors at the channels junction, where the analyte flow and the
gating flow interacted. Since the electroosmotic flow rate was
relatively low, the Reynolds number was usually less than 1.
That meant the two flows were typically parallel streams, known
as laminar flow. In such situation, molecular diffusion would
dominate the interaction of the two flows at the interface. To
investigate the underlying mechanism that led to molecular leakage, optical imaging was employed to analyze the microflow
profiles at the channels junction. Four fluorescent species including rhodamine 6G, rhodamine B, sulforhodamine and FITC were
used as the analytes, which were chose due to their properties of molecular diffusion and electrophoretic mobility. They
showed similar molecular diffusion coefficients. However, their
observed electrophoretic mobility in microchannel had significant differences at given conditions as shown in Fig. 3. Because
the optical filter set was optimized for excitation and emission

Fig. 5. Linear regression of (1/) vs. d2 . Curve 1: g = 2.0; Curve 2: g = 3.0.

Fig. 6. Influence of electric field strength on analyte profiling at the channels junction. (A) 500 V cm1 ; (B) 400 V cm1 ; (C) 300 V cm1 ; (D) 200 V cm1 ; (E)
100 V cm1 ; (F) schematic description of analyte profiling. Curves identity, 0, interface of analyte flow and gating flow; 15, boundaries the analyte could diffused
to in different electric field strengths at the channels junction (AE), respectively.

G. Zhang et al. / Analytica Chimica Acta 584 (2007) 129135

of the fluorescence of rhodamine species, FITC showed a very


small peak in this electropherogram.
In the investigations, the analyte was electrokinetically gated
as described in Section 2. The microflow conditions at the channels junction were fluorescently imaged, and recorded. The
fluorescent intensity along the dashed line (Fig. 4A) at the
entrance of Channel SW was profiled. Before analysis, a twostep data process procedure was carried out. Firstly, to avoid the
bias caused by nonuniform geometry of microchannel, the data
was calibrated as illustrated in Fig. 4B. Secondly, the data was
further normalized since four analytes exhibited different fluorescence intensities at the same concentration. Fig. 4C and D
showed the results of g = 3.0 and 2.0, respectively. It was found
that the fluorescence intensity rapidly dropped down from the
dashed Line 0 to Line 1, 2, 3 and 4. Line 0 symbolized the
interface of the analyte flow and the gating flow. Line x (1, 2, 3
and 4) accordingly represented the limit that no fluorescence of
the analytes (rhodamine 6G, rhodamine B, sulforhodamine and
FITC) could be detected out. The distances (d) between Line 0
and Line x were measured, and showed significant difference. If
such a difference was caused by molecular diffusion, the square
of distance d2 should be linearly proportional to molecular diffusion coefficient (D) and diffusion time (t), according to Giddings
zone spreading theory [33]:
d 2 Dt

133

a series of images at g = 3.0 with FITC as the analyte, varying


EFS at the channels junction from 500 V cm1 to 100 V cm1
(Fig. 6AE). A schematic description about the boundaries of
FITC was given in Fig. 6F. Line 0 represented the interface of
the two flows. Line x (1, 2, 3, 4 and 5) indicated the boundaries of FITC at channels junction in different EFS in channels
junction, respectively. The tendency was quite clear. At EFS of
500 V cm1 , FITC flow was tightly confined, and then delivered into Channel SW (Fig. 6A). No analyte leakage was found.
Nevertheless, with the decrease of EFS, the boundaries gradually
approached to Channel BW (Fig. 6B and C). And eventually, the

(1)

Because the diffusion coefficients of the four fluorescent species


were similar, the vital factor was the diffusion time that was
actually determined by the velocity (v) of the analyte across the
channels junction area. Thus the distance d would be inversely
proportional to the mobility of the analytes (), because:
v = E

(2)

where E denoted the EFS at the channels junction. Fig. 5 revealed


the relationship as predicted. The curves 1 and 2 in Fig. 5
described the linear relationships between d2 and 1/ when the
gating factor g = 2.0 and 3.0, respectively. It was easily understood that the analytes of lower electrophoretic mobility would
have more diffusion time at the channels junction. The gating
flow from Channel B could be divided into two portions. One of
portions entered into Channel BW. The other one was deemed
to enter into Channel SW, which was the real part of stream
for valving. If the analyte only diffused into the latter portion,
no leakage would be found. But in case that the analyte diffused further into the portion dispatched into Channel BW, the
leakage would definitely occur. Theoretically, the gating factor
determined the width of the stream for gating. The greater the
value of gating factor was, the lower the possibility of leakage
was. The analyte needed more time to cross from the gating part
to the part that entered Channel BW with the increase of gating
factor g. It is the reason why all the four species could safely
reach Channel SW without any leakage in Fig. 4C at g = 3.0, and
the slowest species FITC leaked in Fig. 4D at g = 2.0.
If the above analysis was right, reducing the EFS at the channels junction could also lead to analyte leakage according to Eq.
(2). Our experiments validated this hypothesis. Fig. 6 showed

Fig. 7. Influence of analyte concentration on analyte profiling at the channels


junction. Analyte, FITC. (A) 1 103 M; (B) 5 104 M; (C) 1 104 M; (D)
schematic description of analyte profiling. Curves identity, 0, interface of analyte
flow and gating flow; 1, boundary FITC in different concentrations at the channel
junction (AC).

134

G. Zhang et al. / Analytica Chimica Acta 584 (2007) 129135

the analyte flow was electrokinetic gated by the gating flow, the
analyte would diffuse through the interface of the two flows, and
formed a cometic tail-like diffusion area at the channels junction. Because the size of the diffusion area was dependent of
molecular diffusion coefficient and electrophoretic mobility, it
was not surprised why sampling bias could happen by TREKS
[29].
4. Conclusion

Fig. 8. Gaussian fit of data points from intensity profiling in Line 1/Fig. 4C.
Data was symmetrically extended. Hollow and solid circles symbolized original
data and extended data, respectively.

analyte leaked into Channel BW (Fig. 6D). Further decreasing


EFS would make the leakage more serious (Fig. 6E).
Another indirect proof was that the leakage condition should
have nothing to do with analyte concentration if the leakage
was really caused by diffusion, since molecular diffusion is
independent of analyte concentration in diluted solution. Fig. 7
showed the evidence. Changing the analyte concentration had
no influence on the boundaries of analyte.
All evidences above led to the same conclusion that molecular
diffusion played the key role in electrokinetic valving leakage.
As such, the fluorescence intensity curve at the entrance of
Channel SW could be described by the following equation in
mathematics [33]:
c
2 c
=D 2
t
y

(3)

where c, t, D and y designate analyte concentration, diffusion


time, diffusion coefficient and the distance away from the interface of the two microflows, respectively. Note that in Eq. (3), we
used concentration instead of intensity. It was equivalent since
these two parameters were linearly correlated based on Beer-law.
The solution of Eq. (3) was given below:
 2
1
y
c=
(4)
exp
0.5
4Dt
(4Dt)
Hence, the concentration or intensity profiling would exhibit a
Gaussian curve. We randomly selected some data points along
one of the profiling curves in Fig. 4C, and made a symmetric
extension for supplement considering that the diffusion occurred
only in one side in Fig. 4C. It was interesting that these points
matched a Gaussian fit well (R2 = 0.9926) as shown in Fig. 8.
Therefore, molecular diffusion seemed to be a reasonable
explanation accounting for analyte leakage in electrokinetic gating valving based on above investigations. That was to say, when

To reveal underlying mechanism of molecular leakage in


electrokinetic gating valve, optical imaging was utilized to analyze the microflow profile at the channels junction. Molecular
diffusion was believed to be one of the keys responsible for
analyte leakage, which was proven by theoretical calculation
and experiments. Due to diffusion, the analyte would cross the
interface of the analyte flow and gating flow, and then form a
cometic tail-like diffusion area at the channels junction. The size
of the area was determined by the diffusion coefficient and the
velocity of analytes. In case that the size was large enough, the
electrokinetic gating would be disabled for a given gating factor. Additionally, molecular diffusion was the underlying reason
that caused TREKS.
Acknowledgements
The authors gratefully acknowledge the fellowship and the
Grant-in-Aid for Scientific Research (No. P01082) supported by
Japan Society for the Promotion of Science, financial supports
from National Natural Science Foundation of China (Grant Nos:
20405006, 30570468), Program for New Century Excellent Talents in University (NCET), and Program for Distinguish Young
Scientist of Hubei Province (2004ABB004).
References
[1] A. Manz, N. Graber, H.M. Widmer, Sens. Actuators B 1 (1990) 244.
[2] D.J. Harrison, K. Fluri, K. Seiler, Z. Fan, C.S. Effenhauser, A. Manz,
Science 261 (1993) 895.
[3] S.C. Jacobson, R. Hergenroder, L.B. Koutny, J.M. Ramsey, Anal. Chem.
66 (1994) 1114.
[4] P.S. Dittrich, K. Tachikawa, A. Manz, Anal. Chem. 78 (2006) 3887.
[5] C.S. Effenhauser, in: A. Manz, H. Becker (Eds.), Microsystem Technology
in Chemistry and Life Science, Springer, Berlin, Heidelberg, 1999.
[6] V. Dolk, S. Liu, S. Jovanivich, Electrophoresis 21 (2000) 41.
[7] J. Khandurina, A. Guttman, J. Chromatogr. A 943 (2002) 159.
[8] J.S. Go, S. Shoji, Sens. Actuators A 114 (2004) 438.
[9] K. Hosokawa, R. Maeda, J. Micromech. Microeng. 10 (2000) 415.
[10] M.A. Unger, H.-P. Chou, T. Thorsen, A. Scherer, S.R. Quake, Science 288
(2000) 113.
[11] E.T. Lagally, I. Medintz, R.A. Mathies, Anal. Chem. 73 (2001) 565.
[12] R.F. Ismagilov, D. Rosmarin, P.J.A. Kenis, D.T. Chiu, W. Zhang, H.A.
Stone, G.M. Whitesides, Anal. Chem. 73 (2001) 4682.
[13] H. Hartshorne, C.J. Backhouse, W.E. Lee, Sens. Actuators B 99 (2004)
592.
[14] P. Selvaganapathy, E.T. Carlen, C.H. Mastrangelo, Sens. Actuators A 104
(2003) 275.
[15] L. Klintberg, M. Svedberg, F. Nikolajeff, G. Thornell, Sens. Actuators A
103 (2003) 307.
[16] R.H. Liu, J. Bonanno, J. Yang, R. Lenigk, P. Grodzinski, Sens. Actuators
B 98 (2004) 328.

G. Zhang et al. / Analytica Chimica Acta 584 (2007) 129135


[17] R. Pal, M. Yang, B.N. Johnson, D.T. Burke, M.A. Burns, Anal. Chem. 76
(2004) 3740.
[18] S.C. Jacobson, J.M. Ramsey, Anal. Chem. 69 (1997) 3212.
[19] J.P. Alarie, S.C. Jacobson, J.M. Ramsey, Electrophoresis 22 (2001) 312.
[20] C.D. Thomas, S.C. Jacobson, J.M. Ramsey, Anal. Chem. 66 (2004) 6053.
[21] B.-F. Liu, H. Hisamoto, S. Terabe, J. Chromatogr. A 1201 (2003) 201.
[22] S.V. Ermakov, S.C. Jacobson, J.M. Ramsey, Anal. Chem. 70 (1998) 4494.
[23] L.-M. Fu, R.-J. Yang, G.-B. Lee, H.-H. Liu, Anal. Chem. 74 (2002) 5084.
[24] S.C. Jacobson, R. Hergenroder, A.W. Moore Jr., J.M. Ramsey, Anal. Chem.
66 (1994) 4127.
[25] S.C. Jacobson, S.V. Ermakov, J.M. Ramsey, Anal. Chem. 71 (1999) 3273.
[26] B.-F. Liu, M. Ozaki, Y. Ustumi, T. Hattori, S. Terabe, Anal. Chem. 75
(2003) 36.

135

[27] B.-F. Liu, M. Ozaki, H. Hisamoto, Y. Ustumi, T. Hattori, S. Terabe, Anal.


Chem. 77 (2005) 573.
[28] B.-F. Liu, M. Ozaki, H. Hisamoto, Y. Ustumi, T. Hattori, S. Terabe, Lab
Chip 4 (2004) 368.
[29] B.E. Slentz, N.A. Penner, F. Regnier, Anal. Chem. 74 (2002) 4835.
[30] S.V. Ermakov, S.C. Jacobson, J.M. Ramsey, Anal. Chem. 72 (2000)
3512.
[31] R.D. Rocklin, R.S. Ramsey, J.M. Ramsey, Anal. Chem. 72 (2000)
5244.
[32] A. Hibara, M. Tokeshi, K. Uchiyama, H. Hisamoto, T. Kitamori, Anal. Sci.
17 (2001) 89.
[33] J.C. Giddings, Unified Separation Science, Wiley-Interscience Publication,
New York, 1991.