Langmuir 2007, 23, 1483-1488
1483
Variably Elastic Hydrogel Patterned via Capillary Action in Microchannels Rui Dong,†,| Tor W. Jensen,‡,| Kristin Engberg,§,⊥ Ralph G. Nuzzo,† and Deborah E. Leckband*,†,‡,§ Department of Chemistry, Institute for Genomic Biology, and Department of Chemical and Biomolecular Engineering, UniVersity of Illinois, Urbana-Champaign, Illinois 61801 ReceiVed September 19, 2006 Agarose hydrogels of varied elastic modulus can be patterned into 100-µm-wide channels with wall heights of 60 µm. After modifying the hydrogels with chloroacetic acid (acid gels), they are amenable to modification with aminecontaining ligands using EDC-NHS chemistry. Using both rheometry and atomic force microscopy (AFM) nanoindentation measurements, the elastic modulus of unmodified hydrogels increases linearly from 3.6 ( 0.5 kPa to 45.2 ( 5.5 kPa for 0.5 to 2.0 wt/vol % hydrogel, respectively. The elastic modulus of acid gels is 2.2 ( 0.3 kPa to 16.2 ( 1.6 kPa for 0.5 to 2.0wt/vol %, respectively. No further changes were measured after further modifying the acid gels with fibronectin. Confocal images of rhodamine-modified acid gels show that the optimal filling viscosity of the agarose solutions is between 1 and 4 cP. This new method of patterning allows for the creation of substrates that take advantage of both micron-scale patterns and variably elastic hydrogels.
I. Introduction Micron-scale patterning is of enormous potential use to the cell research community because of its ability to create bioactive substrates patterned on length scales relevant to individual cells. Micropatterning exploiting the rapid prototyping of poly(dimethylsiloxane) (PDMS) structures1 has been used to create immobilized patterns and gradients of specific cellular ligands on glass substrates.2-5 In addition, soft material substrates are able to mimic in vivo tissues and provide another variable, elastic modulus, for the in vitro examination of cells. The impact of substrate rigidity on cell growth and differentiation is an important new area of research for understanding both normal and cancerous cell responses to immobilized ligands.6-9 Recently published reports have described methods for patterning soft materials on * To whom correspondence should be addressed:
[email protected]. † Department of Chemistry. ‡ Institute for Genomic Biology. § Department of Chemical & Biomolecular Engineering. | Rui Dong and Tor Jensen have contributed equally to this work. ⊥ Current address: Department of Chemical Engineering, Stanford University, Palo Alto, CA. (1) Duffy, D. C.; McDonald, J. C.; Schueller, O. J. A.; Whitesides, G. M. Rapid prototyping of microfluidic systems in poly(dimethylsiloxane). Anal. Chem. 1998, 70 (23), 4974-4984. (2) Jeon, N. L.; Dertinger, S. K. W.; Chiu, D. T.; Choi, I. S.; Stroock, A. D.; Whitesides, G. M. Generation of Solution and Surface Gradients Using Microfluidic Systems. Langmuir 2000, 16 (22), 8311-8316. (3) Dertinger, S. K. W.; Jiang, X.; Li, Z.; Murthy, V. N.; Whitesides, G. M. Gradients of substrate-bound laminin orient axonal specification of neurons. Proc. Natl. Acad. Sci. U.S.A. 2002, 99 (20), 12542-12547. (4) Fosser, K. A.; Nuzzo, R. G. Fabrication of patterned multicomponent protein gradients and gradient arrays using microfluidic depletion. Anal. Chem. 2003, 75 (21), 5775-5782. (5) Gunawan, R. C.; Choban, E. R.; Conour, J. E.; Silvestre, J.; Schook, L. B.; Gaskins, H. R.; Leckband, D. E.; Kenis, P. J. A. Regiospecific Control of Protein Expression in Cells Cultured on Two-Component Counter Gradients of Extracellular Matrix Proteins. Langmuir 2005, 21 (7), 3061-3068. (6) Engler, A.; Bacakova, L.; Newman, C.; Hategan, A.; Griffin, M.; Discher, D. Substrate compliance versus ligand density in cell on gel responses. Biophys. J. 2004, 86 (1, Pt. 1), 617-628. (7) Engler, A. J.; Griffin, M. A.; Sen, S.; Boennemann, C. G.; Sweeney, H. L.; Discher, D. E. Myotubes differentiate optimally on substrates with tissue-like stiffness: Pathological implications for soft or stiff microenvironments. J. Cell Biol. 2004, 166 (6), 877-887. (8) Ra, H. J.; Picart, C.; Feng, H. S.; Sweeney, H. L.; Discher, D. E. Muscle cell peeling from micropatterned collagen: direct probing of focal and molecular properties of matrix adhesion. J. Cell Sci. 1999, 112 (10), 1425-1436.
length scales similar to individual cells. Gradients of either ligand density10 or elastic modulus11 were created with, respectively, polyethylene glycol (PEG) or polyacrylamide (PAAM) by using micromixers.2 The polymerization of both PEG and PAAM systems was initiated by ultraviolet (UV) light exposure followed by removal of the channel system and subsequent exposure of the polymer to target cells. These systems have proven to be very useful for creating variably elastic modulus substrates for ligand presentation. Because of the use of prepolymer solutions and UV initiation while patterning, it would be useful to find alternative substrates that may be more biocompatible during the patterning procedure. Polysaccharide hydropolymers, for instance, are commonly used for cell encapsulation and have a wide range of achievable elastic moduli.12-14 These biologically inert gels can be chemically modified with ligands in order to elicit specific cellular responses.14-16 These methods were expanded by Luo et al.16 using a laser-activated immobilization to pattern biomolecules in uniform agarose hydrogels. While useful, the latter method is limited to the immobilization of a single ligand at uniform mass coverage and is not amenable to cell encapsulation. The goal of this work is to use high-throughput micron-scale patterning techniques combined with variably elastic (9) Discher, D. E.; Janmey, P.; Wang, Y. L. Tissue cells feel and respond to the stiffness of their substrate. Science 2005, 310 (5751), 1139-1143. (10) Burdick, J. A.; Khademhosseini, A.; Langer, R. Fabrication of Gradient Hydrogels Using a Microfluidics/Photopolymerization Process. Langmuir 2004, 20 (13), 5153-5156. (11) Zaari, N.; Rajagopalan, P.; Kim, S. K.; Engler, A. J.; Wong, J. Y. Photopolymerization in microfluidic gradient generators: Microscale control of substrate compliance to manipulate cell response. AdV. Mater. (Weinheim, Ger.) 2004, 16 (23-24), 2133-2137. (12) Drury, J. L.; Dennis, R. G.; Mooney, D. J. The tensile properties of alginate hydrogels. Biomaterials 2004, 25 (16), 3187-3199. (13) Kong, H. J.; Wong, E.; Mooney, D. J. Independent Control of Rigidity and Toughness of Polymeric Hydrogels. Macromolecules 2003, 36 (12), 45824588. (14) Balgude, A. P.; Yu, X.; Szymanski, A.; Bellamkonda, R. V. Agarose gel stiffness determines rate of DRG neurite extension in 3D cultures. Biomaterials 2001, 22 (10), 1077-1084. (15) Dhoot, N. O.; Tobias, C. A.; Fischer, I.; Wheatley, M. A. Peptide-modified alginate surfaces as a growth permissive substrate for neurite outgrowth. J. Biomed. Mater. Res., Part A 2004, 71A (2), 191-200. (16) Luo, Y.; Shoichet, M. S. Light-activated immobilization of biomolecules to agarose hydrogels for controlled cellular response. Biomacromolecules 2004, 5 (6), 2315-2323.
10.1021/la062738l CCC: $37.00 © 2007 American Chemical Society Published on Web 12/05/2006
1484 Langmuir, Vol. 23, No. 3, 2007
culture materials to allow the examination a wide range of tissuemimetic substrates in parallel. We introduce here a new technique for patterning hydrogels. The use of thermoreversible materials allows the control of reversible gelation with temperature as opposed to chemical methods. This method utilizes capillary-induced flow to fill micron-scale channels with hydrogels above the melt transition temperature followed by cooling induced gelation. These channels are created using decal transfer lithography.17 This procedure generates defined micropatterns of agarose hydrogels, which are easily modified with amine-containing bioactive groups. We describe the fabrication of microchannels on glass substrates, the chemistry used to modify the agarose substrates, and the measured impact of gel density and modification on the elastic modulus of the hydrogel. In addition, we explore the impact of viscosity on channel filling and present an example of the use of these patterns as cell culture substrates. II. Experimental Section Materials. Poly(dimethylsiloxane) was obtained from Dow Corning (Sylgard 184). Silicon wafers (p-type, 100 orientation) were obtained from Silicon Sense. The photoresist, SU8(50), and developer were obtained from Microchem. Glass microscope slides were obtained from Fisher. Tridecafluoro-1,1,2,2-tetrahydrooctyl-1trichlorosilane was obtained from Gelest. 1-Ethyl-3-[3-dimethylaminopropyl]carbodiimide hydrochloride (EDC) was purchased from Sigma, and N-hydroxysulfosuccinimide (Sulfo-NHS) was purchased from Pierce. Lissamine rhodamine B ethylenediamine (Rhodamine) was from Molecular Probes. Agarose (1) (molecular biology grade) with a gel strength of g950 g/cm was from Fisher. Silicon Masters. Pattern masks were designed using Freehand MX software. A Si wafer was first heated to 250 °C for 10 min to remove water. The wafers were then spin-coated with a negative photoresist SU8(50) at 1300 rpm for 30 s followed by a “soft bake” at 120 °C for 15 min. Photoresist-coated wafers were UV-exposed (365 nm) through the pattern mask for 1 min and postbaked at 120 °C for 5 min. This was followed by immersion in a developer bath. Agarose Gel Modification. Agarose hydrogels were prepared at 2 wt/vol % by adding powdered agarose to preheated (95 °C) Dulbecco’s phosphate buffered saline (DPBS) followed by repeated heating in a microwave and shaking until no signs of aggregated agarose were visible. Agarose suspensions were cooled in 150mm-diameter Petri dishes at 4 °C, cut into cubes approximately 2 mm square, and stored at 4 °C. On the basis of a method by Inman,18 a solution of 375 mL of 6 M NaOH, 125 mL of 0.1 M NaCl, and 70.5 g of chloroacetic acid (2) was added to 500 g of 2 wt/vol % agarose gel pieces stirred at room temperature for 24 h followed by thorough rinsing (as monitored by pH of the elute) through a glass frit with 3-4 L of 0.1 M NaCl, then 0.5 L DPBS, over a period of 3-5 h. Gels were further modified with lissamine rhodamine B ethylenediamine using EDC-NHS. Briefly, 20 g of 2 wt/vol % acidmodified gel was rinsed with a solution of 0.1 M MES and 0.1 M NaCl (MES buffer, pH 5.5) twice for 30 min each. After rinsing, 16 mg EDC and 44 mg sulfo-NHS were added to the gels in 10 mL of MES buffer. To this, 0.4 mL of 10 mg/mL rhodamine in DMSO was added. The mixture was shaken for 16 h and rinsed. Atomic Force Microscopy (AFM). Samples were characterized with a silicon nitride cantilever with a spring constant of 0.03 N/m (Veeco) using an MFP 3D AFM (Asylum Research). A conical tip approximation for the AFM tip as described by Sneddon19 and Domke and Radmacher20 is used (17) Childs, W. R.; Nuzzo, R. G. Decal transfer microlithography: A new soft-lithographic patterning method. J. Am. Chem. Soc. 2002, 124 (45), 1358313596. (18) Inman, J. K. Functionalisation of Agarose Beads Via Carboxymethylation and Aminoethylamide Formation. In Affinity Chromatography: A Pratical Approach, Dean, P. D. G., Johnson, W. S., Middle, F. A., Eds. IRL Press Limited: Oxford, 1985, 53-59.
Dong et al. F)
2 Eδ2 tan(R) π (1 - ν2)
(1)
where F is the penetration force, E is the elastic modulus of the agarose gel, δ is the tip penetration depth, ν is the Poisson ratio (assumed to be 0.520), and R is the half-opening tip angle. The force F is determined by the cantilever deflection and spring constant described by the equation F ) kd ) k(z - δ)
(2)
where d is the tip deflection and z is the height of the piezo. Combining eqs 1 and 2 results in z - zo ) d - do +
x
k(d - do)
(2/π)[E/(1 - ν2)] tan(R)
(3)
Individual force curves were modeled with eq 3 using a least-squares fit by adjusting E. A thermal power spectral density (PSD) graph was fit to a simple harmonic oscillator function for each individual cantilever to calculate the spring constant k. The data shown represents three independent samples, each measured at three locations. A minimum of six force curves were analyzed at each location. Rheology. Rheology experiments were performed on the Bohlin C-VOR rheometer with a C-14 cup and bob geometry. For elastic modulus measurements, agarose was poured into the cup and allowed to gel with the bob in place. The elastic modulus was measured for each sample across the range of 0.1 to 10 Hz. The reported moduli represent values averaged across the measured frequency range. The viscosity of melted agarose samples was measured at 55 °C using the same geometry with a range of shear stresses from 1 to 20 Pa. The reported viscosities represent values averaged across the measured shear stress range. Confocal Microscopy. Confocal images were taken with a Leica SP2 confocal microscope with computer control. Multiple images were combined using Amira (Advanced 3D Visualization and Volume Modeling) software, and individual samples were colored artificially. Optical and Fluorescence Microscopy. Optical and fluorescence images were taken with the Zeiss Axiovert 100 inverted researchgrade microscope configured for fluorescence as well as differentialinterference contrast (Nomarski) and phase-contrast microscopy. The entire system is driven using MCID software (Imaging Research, Inc.), which facilitates image collection as well as subsequent analysis and quantification.
III. Results and Discussion A. Channel Formation. Figure 1 illustrates the method for creating open channel patterns by decal transfer lithography. The details for this method have been described previously for transferred patterns with feature heights on the order of 5-10 µm.4,17 A brief description for creating transferred patterns with heights up to 60 µm is presented here. Pattern masters with feature heights of 100 ( 10 µm were created with SU8(50) photoresist on silicon wafers via standard soft lithography methods. The masters were modified by vapor deposition of tridecafluoro-1,1,2,2-tetrahydrooctyl-1-tichlorosilane (FluoroTCS) in a vacuum for 3 h. All master feature heights were measured with a Sloan Dektak 3ST profilometer (Materials Research Laboratory). Masters were then coated with PDMS prepolymer by pouring PDMS onto the master, degassing under vacuum for 5 min to fill microchannels, and then spin-coating to remove excess prepolymer (Figure 1A). (19) Sneddon, I. N. The relation between load and penetration in axisymmetric boussinesq problem for a punch of arbitrary profile. Int. J. Eng. Sci. (Oxford, U.K.) 1965, 3, 47-57. (20) Domke, J.; Radmacher, M. Measuring the Elastic Properties of Thin Polymer Films with the Atomic Force Microscope. Langmuir 1998, 14 (12), 3320-3325.
Variably Elastic Hydrogel
Langmuir, Vol. 23, No. 3, 2007 1485 Scheme 1. Chemical Addition of Acid Moieties to Agarose Gels
Figure 1. Schematic for the construction of open-channeled PDMS microchannels on glass substrates. (A) Filled PDMS channels in a silicon master. (B) Fluoro-TCS release layer on PDMS (red). (C) PDMS capped master. (D) Pattern transferred to glass substrate. (E) PDMS cap removed to expose microchannels. (F) Microchannels filled with agarose gels by capillary force. (G) Image of PDMS microchannels on a glass substrate. The pattern shown has channels 100-µm wide with 300-µm spacing and channel heights of 60 µm. Scale bar represents 800 µm.
For all work presented here, we spin-coat at 1000 rpm for 1 min and at 2000 rpm for 15 s, because this enabled us to achieve well-controlled square-featured PDMS deposition in the master channels with final wall heights of 60 µm. For the 100-µm-tall master features used here, a final wall height of 60 µm transferred to the glass substrates represents a master filled to 40 µm below the tops of the master features. Wall heights ranging from 7 to 60 µm can be obtained using masters with 100-µm features (see Supporting Information). After spin-coating, the master was heated for 1 h at 65 °C followed by UV exposure for 1.5 min. The master was then treated again with Fluoro-TCS for 17 min to provide a semi-adhesive layer between the spin-coated PDMS layer and the subsequent capping PDMS layer (Figure 1B). Approximately 3 mm of PDMS precursor was poured onto the master immediately after the Fluoro-TCS treatment (Figure 1C). The coated master was then heated at 65 °C for at least 12 h. Excessive Fluoro-TCS treatment resulted in incomplete removal of the PDMS patterns from the silicone master, while decreased treatment resulted in poor release between the transferred pattern and the PDMS “cap.” The PDMS mold was cut and peeled from the master and washed with ethanol and water. The channel side of the PDMS mold was UV exposed (180 nm permissive) in air for 4 min before being transferred to a clean glass slide.21 Hand pressure was used to ensure intimate contact between the glass and the PDMS mold. The glass/PDMS pattern was then heated at 65 °C for 1.5 h (Figure 1D). After heating, the PDMS “cap” was peeled off, leaving the PDMS walls on the glass (Figure 1E). Figure 1G shows a composite image of the transferred PDMS pattern on glass. To make the transferred, open-channel pattern hydrophilic, we treated the surface with O2 plasma (300 W, 20 (21) Berdichevsky, Y.; Khandurina, J.; Guttman, A.; Lo, Y. H. UV/ozone modification of poly(dimethylsiloxane) microfluidic channels. Sens. Actuators, B 2004, 97 (2-3), 402-408.
cm3/min, 10 min) through reactive ion etching (RIE). Agarose was introduced into the channels by passive wicking from the larger filled reservoir at 55 °C. This process is driven by capillary forces (Figure 1F). B. Hydrogel Modification. Agarose is a thermoreversible hydrogel with low inherent bioadhesive properties.14,16 Its structure consists of alternating copolymers of (1-3)-linked β-Dgalactose and (1-4)-linked (3-6)-anhydro-R-L-galactose (Scheme 1(1)). The agarose preparation used in this work melts at approximately 85 °C and solidifies at 35 °C. Other agarose preparations solidify over a wide range of temperatures, 17-40 °C, depending on the molecular weight and degree of hydroxyethyl substitution on its side chains.14,22 Once solidified, hydrogels remain thermally stable until they are heated above the melting temperature. For this work, melted hydrogels were maintained at 50-60 °C during patterning. To make the agarose more amenable to ligand modification, the hydroxyl moieties were modified to form a carboxymethyl ether group (Scheme 1(3)) via Williams ether synthesis, following published protocols.18,23 The acid-modified hydrogels can then be further modified with amine-containing ligands via a peptide bond (Scheme 1(4)) using EDC/NHS chemistry. C. Elastic Modulus of Uniform Hydrogels. Both acidmodified and unmodified hydrogels were melted in a microwave and kept at 50-60 °C prior to patterning, and were then gelled at room temperature. While the addition of the carboxylic acid groups did not require changing any of the working protocols at any of the concentrations studied (0.5-2.0 wt/vol %), it did strongly impact the elastic moduli of the solidified hydrogels. To characterize the impact of concentration and acid modification on the resulting hydrogel stiffness, the elastic moduli of a range of unmodified and modified agarose gels were measured using atomic force microscopy (AFM). This technique probes the top 300 to 1000 nm of the hydrogel by nanoindentation of the force probe tip into the gel substrate.11,20 Force curves were analyzed using a method described previously by Domke and Radmacher.20 For the initial mechanical characterization of the gels, uniform gel samples were cast into 60-mm dishes and cooled, resulting in an approximate gel thickness of 1.5 mm. Cast samples were stored and measured under phosphate buffered saline (PBS). For the lowest concentration of agarose tested, 0.5 wt/vol %, unmodified hydrogels had a measured elastic modulus of 3.6 ( 0.5 kPa, while that of acid-modified hydrogels was 2.2 ( 0.3 kPa. The elastic moduli of both acid-modified and unmodified hydrogels increased linearly (R2 ) 0.99) with increasing agarose up to the maximum concentration tested, 2.0 wt/vol % (Figure 2). Unmodified hydrogels had, on average, an elastic modulus 2.9 times that of acid hydrogels at each of the concentrations tested. At 2.0 wt/vol %, unmodified hydrogels had an elastic (22) Normand, V.; Lootens, D. L.; Amici, E.; Plucknett, K. P.; Aymard, P. New insight into agarose gel mechanical properties. Biomacromolecules 2000, 1 (4), 730-738. (23) Peterson, E. A.; Sober, H. A. Chromatography of proteins. I. Cellulose ion-exchange adsorbents. J. Am. Chem. Soc. 1956, 78, 751-5.
1486 Langmuir, Vol. 23, No. 3, 2007
Dong et al.
Figure 3. Mixtures of acid-modified and unmodified gels. The total gel concentration is kept at 2.0 wt/vol %, while the amount of acid-modified gel in the mixture is increased from 0% to 100%. Error bars show the standard deviation. Figure 2. Elastic modulus as a function of gel composition. (A) Straight lines represent various concentrations of modified and unmodified agarose gels. Squares represent unmodified agarose gels, and circles represent acid-modified gels. Closed symbols show the data obtained by AFM and open symbols data obtained by bulk measurements using a cup and spindle geometry. Dashed lines show the linear least-squares fit of the data. (B) Dotted lines show the bulk elastic modulus of unmodified gels as measured by a cup and spindle geometry is shown for frequencies of 0.1 to 10 Hz. Error bars show the standard deviation.
modulus of 45.2 ( 5.5 kPa compared to 16.2 ( 1.6 kPa for acid-modified hydrogels. This measured range, 2-45 kPa, is similar to literature values for AFM measurements of PAAM gels of 3-40 kPa.11 While the elastic moduli of agarose hydrogels depend strongly on the source and preparation method, the values reported here are consistent with trends reported elsewhere.16,24 The elastic modulus values obtained from AFM measurements were compared to bulk measurements of agarose using a cup and spindle rheometer. Values for elastic moduli were taken as an average value across the frequency range 0.1-10 Hz. The values obtained by both methods were consistent with each other (Figure 2). While the elastic moduli of acid-modified gels could have been increased by controlling the degree of acid addition during the modification procedure, the difficulties of controlling the kinetics of this reaction led us to examine mixtures of modified and unmodified gels. With the total gel concentration kept at 2.0 wt/vol %, various amounts of acid-modified material were titrated into the unmodified gel (Figure 3). The compositions mixed linearly from 0% to 100% modified gel (linear least-squares fit R2 ) 0.99). A wide range of elastic moduli and total level of acid-modified gel can be achieved by changing both the total gel percentage and acid to unmodified gel ratio (see Supporting Information). In addition, after 1 and 2 weeks in serum-containing medium at 37 °C and 5% CO2, there was no significant change in the measured elastic moduli (see Supporting Information). When mixtures of acid-modified and unmodified gels were further modified by the covalent attachment of fibronectin using EDCNHS chemistry, the measured elastic modulus for each mixture did not change (Figure 3). D. Patterned Hydrogels. To visualize the hydrogels in the channels, lissamine rhodamine B ethylenediamine (Rhodamine) was conjugated to the gel using EDC/NHS chemistry.25 Control experiments conjugating Rhodamine to unmodified agarose resulted in no detectable fluorescence in the gels after rinsing (not shown). The fluorescently-labeled agarose hydrogel was introduced into the channels by capillary force at 55 °C. Figure
Figure 4. Rhodamine-labeled 1.5 wt/vol % agarose after capillary filling of microchannels. The pattern shown has channels 100-µm wide with 300-µm spacing and features channel heights of 60 µm. Scale bar represents 2000 µm.
4 is a composite fluorescence image of the complete filled pattern area using 1.5 wt/vol % labeled gel. While all concentrations of gel filled the channels, only concentrations of 1.5 wt/vol % and lower consistently filled the entire length of the channels (1-cm path length). For the 2% gels, approximately 75% of the patterns filled to completion, while the remainder filled 25-75% of the pattern. This was attributed both to the higher viscosity of the 2.0 wt/vol % gels and to their faster gelation in the channels during filling. To analyze the differences in the liquid gel characteristics, the viscosity of both the unmodified and acidmodified gels was also measured at 55 °C using a cup and spindle rheometer (Figure 5A). The average values for modified and unmodified gels at 1.0, 1.5, and 2.0 wt/vol % agarose are, respectively, 0.8 ( 0.1, 1.7 ( 0.7, and 4.8 ( 1.2 cP. The lower (24) Stolz, M.; Raiteri, R.; Daniels, A. U.; VanLandingham, M. R.; Baschong, W.; Aebi, U. Dynamic elastic modulus of porcine articular cartilage determined at two different levels of tissue organization by indentation-type atomic force microscopy. Biophys. J. 2004, 86 (5), 3269-3283. (25) Grabarek, Z.; Gergely, J. Zero-Length Crosslinking Procedure With The Use Of Active Esters. Anal. Biochem. 1990, 185 (1), 131-135.
Variably Elastic Hydrogel
Langmuir, Vol. 23, No. 3, 2007 1487
Figure 6. Irregular structures are frozen into the channel upon gelation. A small subset of channels exhibit W-shaped menisci after gelation. Figure 5. The meniscus across the channel decreases in depth with increasing agarose suspension viscosity up to 1.5 wt/vol %. (A) The viscosity of both modified and unmodified gels at 55 °C is averaged for each concentration shown. The error bars indicate standard deviation. *Due to limitations of the viscometer used, the value at 0.5 wt/vol % is given for water at 55 °C. (B) Cross sections of channels filled with 0.5, 1.0, 1.5, and 2.0 wt/vol % are shown. The scale bar represents 100 µm on the z-axis.
viscosity limit of this cup and spindle rheometer did not allow measurement of the 0.5 wt/vol % suspensions. The value plotted in Figure 5A at 0.5 wt/vol % is for water at 55 °C, 0.5 cP. In the data of Figure 4, it was noted that the edges of the channels appeared brighter than the center of the channel and that this effect was more pronounced as the concentration of the gel decreased. Since capillary forces drive the channel filling, menisci at the leading edge of the filling channel as well as at the top of the channel are expected (and found, see below).26,27 We imaged the resulting menisci in the channels using confocal fluorescence microscopy. Three-dimensional reconstructions of the confocal images were assembled using Amira software and are shown in Figure 5B. The depth of the meniscus decreases with increasing liquidstate viscosity up to the 1.5 wt/vol % sample. At 2.0 wt/vol %, the depth of the meniscus once again increases. This results in deeper “valleys” for the lowest and highest viscosity filling solutions and a lower total volume of gel in the microchannel. The final shapes of the menisci for the different concentrations depend on several experimental parameters including the viscosity of the gel precursor, the surface tension, and the length of time required for gelation.28,29 The largest impact of the viscosity is the length of time required to reach the final equilibrium shape of the meniscus.28 The shapes represented here may be nonequilibrium states resulting from the gelation of the solutions during meniscus formation. Supporting this conjecture is an interesting phenomena observed with a small subset of channels (
