Miniaturized Electroosmotic Pump Capable of Generating Pressures

Oct 12, 2012 - The pressure output of a pump cannot be increased simply by connecting several of them in series. This barrier is eliminated with the m...
0 downloads 9 Views 2MB Size
Download PDF
Article pubs.acs.org/ac

Miniaturized Electroosmotic Pump Capable of Generating Pressures of More than 1200 Bar Congying Gu,† Zhijian Jia,‡ Zaifang Zhu,† Chiyang He,§ Wei Wang,† Aaron Morgan,† Joann J. Lu,† and Shaorong Liu*,† †

Department of Chemistry and Biochemistry, University of Oklahoma, Norman, Oklahoma 73019, United States College of Chemical Engineering, Ningbo University of Technology, Ningbo, Zhejiang 315016, P. R. China § College of Chemistry and Chemical Engineering, Wuhan Textile University, Wuhan, Hubei 430073, P. R. China ‡

ABSTRACT: The pressure output of a pump cannot be increased simply by connecting several of them in series. This barrier is eliminated with the micropump developed in this work. The pump is actually an assembly of a number of fundamental pump units connected in series. The maximum pressure output of this pump assembly is directly proportional to the number of serially connected pump units. Theoretically, one can always enhance the pressure output by adding more pump units in the assembly, but in reality the upper pressure is constrained by the microtees or microunions joining the pump components. With commercially available microtees and microunions, pressures of more than 1200 bar have been achieved. We have recently experimented using open capillaries to build this pump, but many capillaries have to be utilized in parallel to produce an adequate flow to drive HPLC separations. In this paper, we synthesize polymer monoliths inside 75 μm i.d. capillaries, use these monoliths to assemble miniaturized pumps, characterize the performance of these pumps, and employ these pumps for HPLC separations of intact proteins. By tuning the experimental parameters for monolith preparations, we obtain both negatively and positively charged submicrometer capillary channels conveniently. Each monolith in a 75 μm i.d. capillary is equivalent to several thousands of open capillaries.

T

readily integratable with lab-on-a-chip devices) and has no moving parts. The first EOP was developed in early 1970s when Pretorius et al.14 used a 5 cm long by 1 mm inner diameter (i.d.) glass columns packed with 1−20 μm silica particles to create the electroosmotic flow (EOF) to drive HPLC separations. Based on Yao and Santiago’s analysis,15 the working pressure of this pump was ∼40 bar. More recently, Paul and Rakestraw10 used a capillary packed with 1−3 μm silica beads to generate the EOF, and the maximum pressure created was ∼350 bar. Chen et al.16,17 used parallel columns packed with 2−3 μm silica beads and produced a maximum pressure of ∼150 bar. Polymer- and silica-based monolithic18,19 and open capillary20−23 EOPs have also been developed, but pressures produced by these pumps were rarely over 100 bar. We have recently developed an innovative configuration of EOP using open capillaries.20,21 The basic pump unit consists of one EOP made from capillaries coated with a positively charged polymer, and another EOP made from capillaries coated with a negatively charged polymer. We have demonstrated that we can join these pump units in series to enhance the pumping pressure, and a maximum pressure of ∼200 bar has been achieved. In this work, we describe an electroosmotic pump made from positive and

he ability of a micropump to generate high pressures is appealing because it enables integration of a complete high-performance liquid chromatographic (HPLC) system on a lab-on-a-chip (LOC) platform. HPLC is arguably the most commonly utilized separation technique for chemical analysis. HPLC can analyze 80% of all known compounds,1 and it is widely used: from drug compound screening to medicine production, from clinical diagnosis to food quality examination, from environmental protection to space exploration, etc. Miniaturizing an HPLC system is of great significance because these systems also allow samples to be analyzed at the point of need rather than a centralized laboratory.2 LOC devices provide a new class of research tools for the investigation of life processes and drug discovery.3,4 Currently, few micropumps can produce adequate pumping pressure for LOC HPLC applications. Here, we describe a miniaturized electroosmotic pump (EOP) capable of generating thousands of bars to drive HPLC separations. Numerous micropumps have been developed since the early 1980s,5 and several articles have reviewed the progress in this area.6−9 The EOP is one of the most promising candidates that has shown potential for practical HPLC separations.10−13 It has several inherent advantages over other types of micropumps. It is capable of generating pulse-free flows; its flow magnitude and direction are convenient to control, and it can be fabricated using standard microfabrication technologies (and thus it is © 2012 American Chemical Society

Received: September 5, 2012 Accepted: October 12, 2012 Published: October 12, 2012 9609

dx.doi.org/10.1021/ac3025703 | Anal. Chem. 2012, 84, 9609−9614

Analytical Chemistry

Article

Scheme 1. Preparation of monolith and bubbleless electrode. (a) Vinylization of the capillary inner wall, (b) preparation of the positive monolith, (c) preparation of the negative monolith, and (d) preparation of the bubbleless electrode. Detailed conditions are described in Experimental Section.

16.39% 1,4-butanediol, and 6% water were mixed ultrasonically into a homogeneous solution and purged with helium for 5 min. This solution was then introduced into the vinylized capillary. After both ends of the capillary were sealed with two septa, the solution was allowed to polymerize at 60 °C in a water bath for 20 h, forming a positive monolith (see Scheme 1b) immobilized inside the capillary. The monolith was then washed with methanol at 100 bar for 12 h using an HPLC pump to remove the unreacted monomers and porogens. The monolith needed to be equilibrated with pump solution (PS) (3 mM sodium acetate, pH 5.0) right before the pump was assembled. The ends of the monolith were cut off to obtain the desired length for experimentation. The same procedure was followed for preparing the negative monolith, but the composition of the polymerization mixture was changed to 1% AMPS, 22.91% BMA, 15.94% EDMA, 0.40% AIBN, 41.36% 1-propanol, 12.41% 1,4-butanediol, and 5.98% water.25 The monolith structure is presented in Scheme 1c. Measurement of Electroosmotic Mobility (μeo). A capillary electrophoresis system with an ultraviolet (UV) absorbance detector (214 nm) and thiourea as an unretained neutral marker were used for this test. Briefly, a detection window was prepared at a distance of 20 cm to the injection end of the EOP monolith (25 cm total length) by removing part of the polyimide coating using a razor blade. The monolith column with the detection window was mounted onto the UV detector, and the monolith was equilibrated with PS by pumping PS through the monolith using an HPLC pump. Then, the injection end of the monolith was inserted into the anode reservoir (for a negative monolith column) or the cathode reservoir (for a positive monolith column), both containing PS. After 5 mM thiourea was electrokinetically injected into the monolith column (at 5 kV for 3 s), 10 kV was applied between the reservoirs. The migration time, t, of thiourea was recorded, and μeo was computed by the velocity of

negative monoliths and demonstrate it for generations of very high pressures (>1000 bar). We also employ such a miniaturized pump to develop a single-pump HPLC system for the gradient elution of proteins.

■

EXPERIMENTAL SECTION Reagents and Materials. Ethylene glycol dimethacrylate (EDMA, 98%) and butyl methacrylate (BMA) were purchased from Alfa Aesar (Ward Hill, MA). 2-Acrylamido-2-methylpropane sulfonic acid (AMPS, 99%) and [2-(methacryloyloxy)ethyl]- trimethylammonium chloride (META, 75 wt % in water) were obtained from Sigma-Aldrich (St. Louis, MO). Stearyl methacrylate, 1-propanol, and 2,2′-azobisiso-butyronitrile (AIBN, 98%) were obtained from Aldrich (Steinheim, Germany). Methacryloyloxypropyl-trimethoxysilane (γ-MAPS, 98%) was purchased from Acros (Fairlawn, NJ). Cyclohexanol was obtained from J.T. Baker (Phillipsburg, NJ). Acrylamide, N,N′-methylene bisacrylamide (bis), N,N,N′,N′-tetramethylethylenediamine (TEMED), and ammonium persulfate (APS) were obtained from Bio-Rad Laboratories (Hercules, CA). 1,4Butanediol (99%) was supplied by Emerald BioSystems (Bainbridge Island, WA). Ribonuclease A, cytochrome C, insulin, lysozyme, transferrin, trypsin inhibitor, and myoglobin were obtained from Sigma (St. Louis, MO). Fused silica capillaries were purchased from Polymicro Technologies Inc. (Phoenix, AZ). All solutions were prepared with ultrapure water purified by a NANO pure infinity ultrapure water system (Barnstead, Newton, WA). Preparation of EOP Monolith. Referring to Scheme 1, the inner walls of the capillaries (17−30 cm length × 75 μm i.d. × 360 μm o.d.) were vinylized with γ-MAPS (Scheme 1a) following the procedures we reported previously.24 A solution containing 0.72% (w/w, hereafter the % indicates w/w concentrations unless otherwise indicated) META, 23.19% BMA, 15.94% EDMA, 0.39% AIBN, 37.39% 1-propanol, 9610

dx.doi.org/10.1021/ac3025703 | Anal. Chem. 2012, 84, 9609−9614

Analytical Chemistry

Article

Figure 1. Schematic configuration of pump assembly. (A) A fundamental pump unit. (B). A pump assembly.

thiourea divided by the field strength [μeo (cm2 s−1 V−1) = 0.05/t]. Measurement of Overall EOF (Q′eo). An open capillary (150 μm i.d.) was attached to the outlet of a pump assembly. After some PS was pumped into the open capillary, a microscope was used to monitor the movement of the meniscus. The overall EOF was estimated by the PS flow rate inside the open capillary; Q′eo (μL s−1) = 0.0177L′/t′, where L′ (mm) is the distance the meniscus had moved, and t′ (s) is the period of the test. Preparation of Bubbleless Electrode. A 20 cm long by 100 μm i.d. (360 μm o.d.) capillary was cut, and its inner wall was cleaned by sequentially flushing the capillary with 1.0 M NaOH for 45 min, DI water for 15 min, and acetonitrile for 15 min. After the capillary was dried under N2, the inner wall was reacted with a solution containing 30% v/v γ-MAPS in acetone at 50 °C for 14 h. Then, the capillary was flushed with acetonitrile and dried with N2. A degassed solution containing 15%T (%T is the total weight concentration of acrylamide and bis in the solution), 2% C (% C represents bis concentration relative to acrylamide), 0.2% v/v TEMED, and 0.1% APS was pressurized into the capillary and remained inside the capillary at 0 °C overnight and then at 4 °C for another 24 h. The polymer structure is presented in Scheme 1d. After 1 cm of the capillaries at both ends were cut off and discarded, 50 mM sodium tetraborate was electrophoretically driven through the polyacrylamide inside the remaining capillary until a stable current was obtained. The above capillary was cut into ∼2 cm pieces, and each piece serves as a bubbleless electrode. Construction of Pump Assembly. The configurations of the basic pump unit and pump assembly are presented in Figure 1, and an image of it is shown in Figure 5B. Each bubbleless electrode had a length of 2 cm and an inner diameter of 100 μm (360 μm o.d.). The positive monoliths have a length of 15 cm, while the negative monoliths have a length of 11.9 cm. All monoliths were prepared inside 75 μm i.d. capillaries. Microtees from Valco Instruments were used to connect the bubbleless electrode, positive monolith, and negative monolith. The inlet of the pump assembly and the free ends of the bubbleless electrodes were inserted into appropriate reservoirs containing PS. The HV was applied to the appropriate reservoirs via platinum electrodes.

Measurement of Pumping Pressure. A flow-through pressure sensor (model 602192-2, Senso-Metrics Inc., Simi Valley, CA) was used for the ΔPmax measurements. The sensor channel was flushed with methanol at a flow rate of 1 mL/min for 10 min first under no pressure, and then under pressure (∼100 bar) for 10 min to remove possible trapped air bubbles. After the sensor was rinsed with PS, the outlet of the pump assembly was connected to the sensor inlet while the sensor outlet was blocked using a stainless steel plug from IDEX (Lake Forest, IL, USA). The HV was then applied to the pump assembly, and the pressure number on the sensor indicator increased gradually. A ΔPmax point was recorded as the pressure data became stabilized. Preparation of Monolith Separation Column. The protocol was similar to that for the monolith EOP preparation. After a 32 cm long by 75 μm i.d. (360 μm o.d.) fused silica capillary was cleaned and vinylized, a solution containing 17.8% stearyl methacrylate, 11.9% EDMA, 0.30% AIBN, 42% 1propanol, and 28% 1,4-butanediol was introduced into the capillary to fill a length of up to ∼22 cm, leaving ∼10 cm of the capillary unfilled. The capillary ends were sealed with septa, and the entire capillary was placed in a water bath at 60 °C for 20 h for monolith formation. A 1−2 mm detection window was created on the empty-capillary end but close to the monolith by removing the polyimide coating with a razor blade. The column was cut to 30 cm with an effective length of 20 cm.

■

RESULTS AND DISCUSSION Figure 1A presents a schematic configuration of the fundamental unit of this pump. It consists of one positive monolith and one negative monolith connected together via a piece of capillary tubing. An external high voltage (HV) is applied across the monoliths via a conducting gel immobilized inside a capillary (the so-called bubbleless electrode). Because the EOF goes from ground to +HV in the positive monolith and from +HV to ground in the negative monolith, the pump solution can move from the inlet to the outlet smoothly. An excellent feature of this pump is that if the positive monolith generates a pressure of P1 (the pump solution in the connecting capillary tubing will be under a pressure of P1) and the negative monolith generates a pressure of P2, the pressure at the outlet will have a pressure of P1 + P2. Since both the inlet and outlet of the pump unit are electrically grounded, we can further connect 9611

dx.doi.org/10.1021/ac3025703 | Anal. Chem. 2012, 84, 9609−9614

Analytical Chemistry

Article

the pump assembly: by attaching more pump units to the EOP or by elevating the electric field across the monoliths. The former option is preferred if high voltages are undesirable. A large number of pump units can be constructed on an LOC device conveniently. As the above EOP was practically utilized, the actual pump pressure (ΔPactual) was smaller than ΔPmax, because the actual pump rate (Qactual) will not be zero. Figure 3 presents the

the outlet of one unit to the inlet of another unit, and so on (see Figure 1B). The overall pressure output of this pump assembly will be n(P1 + P2), where n is the number of pump units connected. That is, we can theoretically achieve any pressure we want by incorporating an appropriate number of pump units into the pump assembly. To achieve the best performance, we should make flow rates in all monolith segments identical. When the electroosmotic mobilities in two monoliths were different, we usually compensated the differences by reducing the length of (and therefore increasing the electric field strength on) the monolith having the lower electroosmotic mobility. Since the ratio of the flow rate in the positive monolith to that in the negative monolith was ∼115:145 in this experiment, the length of the positive monolith was shortened to 11.9 cm, while the negative monolith was 15 cm. Even if the flow rates in two monolith segments were different, the pump would still work properly as long as all monolith segments were pushing the pump solution forward (the flow rate from a later monolith was lower than that from a proceeding monolith). If any of the monoliths were pulling the pump solution forward, a negative pressure must have been created before the monoliths, which could lead to bubble formation and electricity disconnection inside the pump. Figure 2 presents the maximum pressure output of a pump assembly as a function of the number of serially connected

Figure 3. Relationship between output pressure and pump rate. Linear fitting results using eq 1 for curves from the bottom to the top: ΔPmax = 80.4 bar, Q′eo = 145 nL min−1, and R2 = 0.996; ΔPmax = 135 bar, Q′eo = 145 nL min−1, and R2 = 0.993; ΔPmax = 210 bar, Q′eo = 145 nL min−1, and R2 = 0.979; ΔPmax = 397 bar, Q′eo = 145 nL min−1, and R2 = 0.998; ΔPmax= 551 bar, Q′eo = 143 nL min−1, and R2 = 0.995; ΔPmax = 1108 bar, Q′eo = 287 nL min−1, and R2 = 0.995. The error bars were obtained from 3−5 repetitive measurements. Other conditions are the same as stated in Figure 2.

relationship between these two parameters; the actual pressure output of an EOP decreases linearly with the increasing pump rate. In general ⎛ ⎞ Q ΔPmax ΔPactual = ⎜1 − actual ⎟ΔPmax = ΔPmax − × Q actual Q ′eo ⎠ Q ′eo ⎝ (1)

Figure 2. Relationship between the pressure output and the number of pump units. Both the positive and negative monoliths were prepared inside 75 μm i.d. capillaries. The positive monolith had a length of 11.9 cm, while the negative monolith had a length of 15 cm. The pumping solution was 3 mM sodium acetate (pH 5.0).

The pump assembly can generate both high pumping pressure and relatively high flow rates, compared to a pump assembly using open capillaries as pumping elements.26,27 The reason is that monoliths are porous materials; by tuning the experimental parameters, their pore sizes can be controlled to be small to enhance the pumping pressure, and their porosities can be controlled to be high to boost the pumping rate. The effective pore diameter of a monolith can be estimated from its maximum pressure output according to the Hagen−Poiseuille equation

pump units. As expected, an excellent linear relationship exists between the two variables, confirming that any high pressure can be theoretically achieved by incorporating a correct number of pump units into the pump assembly. In reality, however, the upper pressure limit is constrained by the microtees and microunions used to join the bubbleless electrode to the monoliths. The currently available microtees and microunions fail to hold the bubbleless electrode or the monolith (a capillary component) together as the pumping pressure goes beyond a certain pressure (often ∼1000 bar). Using selected microtees from Valco Instruments (Houston, Texas, USA), the highest pressure we have achieved so far is 1208 bar. The maximum pressure output also increases linearly with the voltage applied on the pump (see the inset of Figure 2). Therefore, we have two options to raise the pressure output of

ΔPmax =

⎛ 32ηV ⎞ ⎛ 128ηL ⎞ ⎜ ⎟(mQ ) = ⎜ ⎟μ 4 eo ⎝ mπ d ⎠ ⎝ d 2 ⎠ eo

(2)

where ΔPmax represents the maximum output pressure (a pressure produced as the pump is completely blocked); η, the viscosity of the pump solution; L, the length of the monolith; d, the effective pore diameter; m, the number of equivalent pores; Qeo, the electroosmotic flow in a single pore; μeo, the electroosmotic mobility; and V, the voltage applied across the monolith. Since ΔPmax and μeo can be experimentally measured, 9612

dx.doi.org/10.1021/ac3025703 | Anal. Chem. 2012, 84, 9609−9614

Analytical Chemistry

Article

Table 1. Effective Pore Diameters and Porosities of Positive and Negative Monoliths μeo2 1

positive monolith negative monolith

3.98 ± 0.05 2.85 ± 0.04

6

Qeo3

ΔPmax4

d5

m

porosity

115 ± 4 145 ± 4

130 ± 1.9 76 ± 4.9

0.70 ± 0.01 0.77 ± 0.02

3370 ± 254 5440 ± 424

33% ± 1.4% 58% ± 2.3%

1 Both positive and negative monoliths were prepared inside 75 μm i.d. capillaries and had a length of 15 cm. The external voltage applied across the monoliths was 5 kV. 2The numbers have a unit of 10−4 cm2 V−1 s−1. 3The numbers have a unit of nL min−1. 4The numbers have a unit of bar. 5The numbers have a unit of μm. 6All the numbers after ± are standard deviations.

and η and V are known experimental parameters, d can be calculated using eq 2. Once d is known, m can be calculated from the overall EOF (Q′eo). The porosity of the monolith can be computed by ϕ=

md 2 d′2

(3)

where d′ is the i.d. of the capillary housing the monolith. Table 1 lists the effective pore diameter and porosity. For positive monoliths, their effective pore diameters are around 700 nm, and porosities are around 33%. The negative monoliths have slightly larger pore diameters (∼770 nm) but much higher porosities (58%). Porosities at these levels are pretty common in polymer monoliths.28 We have experimented using open capillaries to assemble the EOP.25 Owing to the challenges to derivatize narrow capillaries, the narrowest capillaries we tested had an i.d. of 5 μm. With n = 10 and V = 20 kV, we achieved a pressure of ∼200 bar. In order to obtain adequate flow to drive an HPLC separation, 52 capillaries (m = 52) were used in parallel. On the basis of the data presented in Table 1, we can produce positive and negative nanometer pores easily, and one 75 μm diameter monolith is equivalent to several thousands of capillaries in parallel. Figure 4 presents scanning electron micrograph (SEM) images of two typical monoliths; the largest pores in both

Figure 5. Miniaturized high-pressure EOP for HPLC separations of proteins. (A) A capillary HPLC system with a 3-unit EOP. Pump solution (PS), 3 mM sodium acetate (pH 5.0); applied voltage (V), 10 kV; MA, 0.1% trifluoroacetic acid in 30% v/v acetonitrile aqueous solution; MB, 0.1% trifluoroacetic acid in 64% v/v acetonitrile aqueous solution; V1, a 4 nL injection valve; V2, a 10 port valve with an 11 cm long by 75 μm i.d. capillary loop for MA and a 25 cm long by 320 μm i.d. capillary loop for MB; M, a 10 cm long by 75 μm i.d. capillary; C, a 20 cm long by 75 μm i.d. acrylate monolith column; D, a Linear UVIS 200 absorbance detector (210 nm); RC, a 40 cm long by 75 μm i.d. acrylate monolith column. (B) Images of the 3-unit EOP. (C) Chromatograms of protein separations. The dashed lines exhibit the profiles of the acetonitrile (indicated by ACN) concentration in the eluent. Peak Identifications: 1, ribonuclease A; 2, insulin; 3, cytochrome C; 4, lysozyme; 5, transferrin; 6, trypsin inhibitor; and 7, myoglobin.

used for protein separation with gradient elution. The gradient eluent is produced using a ten-port valve (V2, Valco Instruments). When V2 is set at the MA position (all ports are connected via the dashed lines), mobile phase A (MA, a weaker eluent) is pumped to condition the separation column (C). At this time, mobile phase B (MB) is loaded into the MBloop, and a sample is loaded into the injector (V1). When V2 is switched to the MB position (all ports are connected via the solid lines), V1 is also switched to inject the sample to C. As MA in tubing M and MB in the MB-loop are driven forward, they are mixed, forming a gradient eluent for analyte separation, while MA in the MA-loop is replenished. The gradient profile can be controlled by changing the compositions and/or concentrations of MA and MB and the geometry and dimensions of M. In general, the MA-loop and MB-loop should be sufficiently long so that all analytes can be eluted out and the separation column can be reconditioned thoroughly. A restriction coil (RC) was employed in this work to prevent bubble formations inside the system. Figure 5B presents an image of a three-unit pump assembly. A coin was located on the left of the pump as a reference to

Figure 4. SEM images of typical positive and negative monoliths. The red arrows and marks indicate the pore sizes (μm), while the blue arrows and numbers indicate the granule sizes (μm). The scales on both images are 2 μm.

monoliths seem to have comparable sizes (∼2 μm). The high porosities of the negative monoliths are likely caused by the greater granules composing the monoliths. Due to the differences in porosities, one 75 μm diameter positive monolith is equivalent to ∼3400 of 700 nm diameter open capillaries, while the 75 μm diameter negative monolith is equivalent to ∼5400 of 770 nm diameter open capillaries. In order for a pump to produce a high pumping pressure and a large flow rate, we should optimize the monoliths to have high porosities and small pore diameters. The pump assembly is readily utilized for HPLC separations. Figure 5A presents an HPLC setup using an assembly with three EOP units serially connected, and the system has been 9613

dx.doi.org/10.1021/ac3025703 | Anal. Chem. 2012, 84, 9609−9614

Analytical Chemistry

Article

(14) Pretorius, V.; Hopkins, B. J.; Schieke, J. D. J. Chromatogr., A 1974, 99, 23. (15) Yao, S.; Santiago, J. G. J. Colloid Interface Sci. 2003, 268, 133. (16) Chen, L.; Ma, J.; Guan, Y. Microchem. J. 2003, 75, 15. (17) Chen, L.; Ma, J.; Guan, Y. J. Chromatogr., A 2004, 1028, 219. (18) Tripp, J. A.; Svec, F.; Fréchet, J. M. J.; Zeng, S.; Mikkelsen, J. C.; Santiago, J. G. Sens. Actuators, B 2004, 99, 66. (19) Nie, F. Q.; Macka, M.; Barron, L.; Connolly, D.; Kent, N.; Paull, B. Analyst 2007, 132, 417. (20) Dasgupta, P. K.; Liu, S. Anal. Chem. 1994, 66, 1792. (21) Dasgupta, P. K.; Liu, S. Anal. Chem. 1994, 66, 3060. (22) Liu, S.; Dasgupta, P. K. Talanta 1994, 41, 1903. (23) Liu, S.; Dasgupta, P. K. Anal. Chim. Acta 1995, 308, 281. (24) Gu, C.; Lin, L.; Chen, X.; Jia, J.; Ren, J.; Fang, N. J. Chromatogr., A 2007, 1170, 15. (25) Peters, E. C.; Petro, M.; Svec, F.; Fréchet, J. M. J. Anal. Chem. 1998, 70, 2288. (26) He, C.; Lu, J. J.; Jia, Z.; Wang, W.; Wang, X.; Dasgupta, P. K.; Liu, S. Anal. Chem. 2011, 83, 2430. (27) He, C.; Zhu, Z.; Gu, C.; Lu, J.; Liu, S. J. Chromatogr., A. 2012, 1227, 253. (28) Xu, Y.; Cao, Q.; Svec, F.; Fréchet, J. M. J. Anal. Chem. 2010, 82, 3352.

show the overall pump size. Figure 5C presents the separation results from the above system. The top chromatogram was obtained using isocratic elution (MA only); only three out of seven proteins were eluted out. The middle chromatogram was obtained also using isocratic elution (MB only); seven proteins were eluted out unresolved. The bottom chromatogram was obtained from gradient elution (the dashed line indicated the gradient profile); seven proteins were baseline resolved in less than 5 min. All these separations were carried out at a flow rate of 210 nL min−1 and a pressure of 310 bar (∼100 bar on the separation column and ∼200 bar on the restriction column). In this application, only a fraction of the maximum pumping power of the 3-unit pump assembly was utilized.

■

CONCLUSIONS We have developed a miniaturized high-pressure electroosmotic pump and have demonstrated that the pump generates pressures >1200 bar. The pump is capable of generating higher pressures, but the highest pressure it can generate is currently limited by the lack of high-pressure microtees used to join capillary components. This achievement is attributed to the use of monolithic materials which provide arrays of dense, narrow pores for electroosmosis. At this point, we have not tested the lifetime of the monolith, but we would not anticipate its lifetime to be longer than or even comparable to that of a commercial reciprocating pump. We will report this data in the future. We have used this pump for HPLC separations of proteins, but only a small portion of the pumping power (310 bar out of 1200 bar) was utilized. We are looking actively for applications that can use its full power and hope the publication of this work will facilitate identification of some practical applications.

■

AUTHOR INFORMATION

Corresponding Author

*Email: [email protected]. Fax: (405) 325-6111. Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS This work is partially sponsored by the Department of Energy Grant DE-SC0006351 and the Oklahoma Center for the Advancement of Science and Technology Grant AR11-003.

■

REFERENCES

(1) Church, W. H. Journal of Undergraduate Neuroscience Education 2005, 3, A36. (2) Daw, R.; Finkelstein, J. Nature 2006, 442, 367. (3) Craighead, H. Nature 2006, 442, 387. (4) Neužil, P.; Giselbrecht, S.; Länge, K.; Huang, T. J.; Manz, A. Nat. Rev. Drug Discovery 2012, 11, 620. (5) Smits, J. G. Sens. Actuators, A 1990, 21, 203. (6) Laser, D. J.; Santiago, J. G. J. Micromech. Microeng. 2004, 14, R35. (7) Nguyen, N. T.; Huang, X. Y.; Chuan, T. K. Journal of Fluids Engineering-Transactions of the ASME 2002, 124, 384. (8) Wang, X.; Cheng, C.; Wang, S.; Liu, S. Microfluid. Nanofluid. 2009, 6, 145. (9) Wang, X.; Wang, S; Gendhar, B.; Cheng, C.; Byun, C.;K.; Li, G.; Zhao, M.; Liu, S. TrAC, Trends Anal. Chem. 2009, 28, 64. (10) Paul, P. H.; Rakestraw, D. J. U. S. Patent 6019882, 2000. (11) Fuentes, H. V.; Woolley, A. T. Lab Chip 2007, 7, 1524. (12) Toh, G. M.; Corcoran, R. C.; Dutta, D. J. Chromatogr., A 2010, 1217, 5004. (13) Yanagisawa, N.; Dutta, D. Electrophoresis 2010, 31, 2080. 9614

dx.doi.org/10.1021/ac3025703 | Anal. Chem. 2012, 84, 9609−9614