1 23
Your article is protected by copyright and all rights are held exclusively by SpringerVerlag. This e-offprint is for personal use only and shall not be self-archived in electronic repositories. If you wish to self-archive your work, please use the accepted author’s version for posting to your own website or your institution’s repository. You may further deposit the accepted author’s version on a funder’s repository at a funder’s request, provided it is not made publicly available until 12 months after publication.
1 23
Author's personal copy Microsyst Technol DOI 10.1007/s00542-012-1626-6
TECHNICAL PAPER
In-plane spectroscopy of microfluidic systems made in photosensitive glass Khalid Hasan Tantawi • William Gaillard • Jake Helton • Emanuel Waddell • Sergey Mirov Vladimir Fedorov • John D. Williams
•
Received: 5 April 2012 / Accepted: 6 July 2012 ! Springer-Verlag 2012
Abstract This work manifests the use of microstructures made in ApexTM photosensitive glass for in-plane investigation of microfluidic systems by demonstrating the ability to detect chemical fluids through etched glass sidewalls. Absorption spectra of liquid ethanol in the near infrared (NIR) region and Raman spectroscopy of dimethyl sulfoxide in microcuvettes made in ApexTM photosensitive glass demonstrate the high potential of photosensitive glass processing in microfluidic applications in which stacks of microfluidic systems are analyzed from the sidewalls. This eliminates the need for non-planar observation of microfluidic systems.
1 Introduction A major goal of the lab-on-a-chip industry is the production of a multiuse chemical reactor with on board real time spectroscopy. To date, this task has been partially achieved using silicon channels that are anodic bonded to glass lids and polydimethylsiloxane (PDMS) microchannels patterned on polymer or glass substrates (Pantoja et al. 2004); K. H. Tantawi ! W. Gaillard ! J. Helton ! J. D. Williams (&) Department of Electrical and Computer Engineering, University of Alabama in Huntsville, Huntsville, AL 35899, USA e-mail:
[email protected] E. Waddell Department of Chemistry and Material Science, University of Alabama in Huntsville, Huntsville, AL 35899, USA S. Mirov ! V. Fedorov Department of Physics, University of Alabama at Birmingham, Birmingham, AL 35294, USA
Dittrich and Manz 2006). While PDMS provides easier pumping schemes, the material is naturally hydrophobic (Chen and Lindner 2007), slightly porous (Berdichevsky et al. 2004), and soluble in several organic solvents. Thus, PDMS devices are limited to single use reactors for several organic chemistry processes. Furthermore, PDMS has limited optical transparency (Cai et al. 2008) windows when compared to quartz and commercial photosensitive glasses. While silicon and SiO2 are chemically resistant, optical spectroscopy in these channels is typically performed from above the channel using a benchtop spectrometer. Thus, in order to provide high quality optically transparent microfluidic reactors with in-plane spectroscopy, one should manufacture the entire channel in glass. However, quartz-based microfluidic chips are much more expensive, and it is more difficult to fabricate microdevices with quartz (Ou et al. 2009). To overcome limitations in semiconductors and PDMSbased technologies, many fabrication processes have been developed using glass to produce sidewalls with average surface roughness values in the nanometer range. Nevertheless, these glass technologies are limited to wall heights of about 160 lm such as glass etching (Queste et al. 2010; Bhatnagar et al. 2007; Lin et al. 2001) and laser drilling (Kim et al. 2005). An alternative technique, photodefinable glass processing, often resulted in very rough and opaque sidewalls (Dietrich et al. 1996; Flemming et al. 2009). However, a few research efforts have developed improved etch chemistries and post process anneals that result in optically transparent glass materials suitable for in-line spectroscopy of microfluidic channels (Cheng et al. 2003; Williams et al. 2010; Tantawi et al. 2011). These efforts demonstrate superiority in producing optically smooth and transparent sidewalls taller than 160 lm and aspect ratios greater than 30:1 (Williams et al. 2010). Furthermore,
123
Author's personal copy Microsyst Technol
Fig. 1 This technology allows for in-plane analysis of stacks or 3D microfluidic structures rather than the current non-planar observation
Fig. 3 Scanning electron microscope images illustrating the reduction in surface roughness of the glass sidewall from that shown on the left to that on the right after annealing lithographically-patterned glass
Fig. 2 A schematic graph of the experimental setup. The spectrometer light signal is transmitted through the microcuvette and liquid to the detector
photodefinable glass processing allows for the fabrication of embedded channels in the glass (Cheng et al. 2003). In addition, microstructures made in photostructurable glasses may be mass produced at low cost. The four-stage process shown in (Tantawi et al. 2011) results in glass microstructures with sidewalls having a root mean square (RMS) surface roughness (r) of approximately 33 nm. Hence, microstructures fabricated using this process, are suitable for optical measurements with wavelengths as short as 330 nm (Davies 1954; Guenther et al. 1984). This work demonstrates the use of the process outlined in (Tantawi et al. 2011) for performing an in-plane spectral analysis on stacked microfluidic systems through the glass sidewalls eliminating the need for a non-planar observation of the system as currently employed in microfluidic technologies (Pan et al. 2004; Birardaa et al. 2010; Prim et al. 2011) as elucidated in Fig. 1. Here we present measurements of the visible, near-infrared (NIR) absorption, and Raman spectra of fluids through the sidewalls of transparent glass microcuvettes.
2 Experiment and method 2.1 Experimental setup ApexTM glass wafers of 500 lm thickness were purchased from Life Bioscience, Inc., based in Albuquerque, New Mexico. In-plane spectral analysis by UV/Visible and NIR spectroscopy are performed on a fluidic sample through the
123
Fig. 4 A microcuvette fabricated in ApexTM photosensitive glass in a SEM image (left) and in an optical micrograph (right)
sidewalls of microcuvettes made in ApexTM photosensitive glass, Raman spectral analysis was also performed on separate fluidic samples. The experimental setup is shown in Fig. 2, in which only the spectrometer light traversing the liquid-filled microcuvette is allowed to reach the detector, by means of a 200 lm hole in a barrier. 2.2 Sample preparation A process to fabricate microstructures from ApexTM photosensitive glass with smooth sidewalls that are transparent in the near UV, visible and mid IR regions was presented in (Tantawi et al. 2011). Briefly, the process starts by irradiating photosensitive glass to UV light to receive a certain dose that depends on glass thickness, during this step a reaction occurs in which cerous (Ce3?) ions are photo oxidized to Ce3?? ion and electron pairs as suggested in (Brandily-Anne et al. 2010). The Ce3?? ion differs from the Ce4? ion in that the photo activated electron stays in the proximity of the cation (Hu¨lsenberg et al. 2008). The glass is then baked using a two-stage process, in which temperature is first raised to 500 "C causing silver ions to reduce to silver and agglomerate into clusters. The temperature is then raised to 550–560 "C to induce the glass matrix to form into the crystalline-phase lithium metasilicate (Dietrich et al. 1996) using the silver particles as nucleation sites. The lithium metasilicate is preferred for etching by hydrofluoric (HF) acid over the unexposed glass
Author's personal copy Microsyst Technol
equipped with a blocking wall that has a hole of 200 lm in diameter as was shown in Fig. 2. Raman spectral analysis was then performed on a specimen of less than 0.24 lL of pure dimethyl sulfoxide (DMSO) (CH3)2SO in another glass microcuvette. DMSO is an aprotic solvent that is commonly used in Raman spectral studies. Raman spectroscopy was performed using a HR 800 Confocal Raman Spectrometer by HORIBA Jobin–Yvon. The laser excitation unit used was a He–Ne laser at a wavelength of 633 nm by Coherent Innova at a power of 500 mW. Fig. 5 Absorption spectra of Ethanol in the UV–Vis and NIR regions through the walls of a microcuvette made in Apex photosensitive glass
matrix with a 20:1 selectivity when a concentration of 5 % (by volume) hydrofluoric acid is used. This results in a three-dimensional image of the mask on the glass structure. Depending on glass composition, and method of exposure, the surface roughness after this step is in the range of 0.7–5 lm (Dietrich et al. 1996; Tantawi et al. 2011). When ApexTM glass is subsequently annealed at 530–545 "C, the RMS surface roughness is reduced to a value of 33 nm (Tantawi et al. 2011). This allows the glass to be used for applications involving detection of optical signals at wavelengths as short as 330 nm. Surface roughness was also shown to be reduced for FoturanTM glass at an annealing temperature of 620 "C (Williams et al. 2010) or at 560–570 "C (Cheng et al. 2006). The scanning electron microscope images presented in Fig. 3 illustrate the reduction in sidewall surface roughness due to the post-etch anneal. The annealed sample presented on the right side of the figure is suitable for optical and nearinfrared spectroscopy. A microcuvette with internal dimensions of 700 lm 9 700 lm 9 500 lm and with a 300 lm wall thickness was fabricated, and bonded to a glass substrate of the same type. A SEM image of the ApexTM microcuvette is shown in Fig. 4. An optical micrograph of the cuvette is also shown to demonstrate the achieved sidewall clarity. These structures demonstrate transparent walls achievable for spectroscopic applications. 2.3 Spectral analysis To examine the spectroscopic application of this technology, a microcuvette was filled with less than 0.24 lL of liquid 95 % ethanol, and the UV/Vis–NIR absorption spectra were taken for the ethanol specimen through the annealed glass sidewalls as was demonstrated in Fig. 2. The transmission spectra were taken at 2 nm intervals using the Cary 5000 UV–Visible–NIR Spectrophotometer by Varian. A special sample holder was designed and was
3 Results and discussion Absorption spectra of ethanol (CH3CH2OH) in the UV/Vis– NIR regions were taken in the wavelength range of about 300–2,900 nm. Ethanol is an organic compound that is commonly used in pharmaceutical drugs directed at the central nervous system, as a fuel additive for combustion engines, and as a chemical solvent. Ethanol has also been used for calibration of NIR instruments in the 1,100– 1,600 nm range of wavelengths (Burns and Ciurczak 2001). Furthermore the same NIR range was used in determining the water content in ethanol in (Cho et al. 2005). The transmission spectra of ethanol in the glass microcuvette are shown in Fig. 5 along with the spectra of the glass microstructure. There has been reported at least 17 absorption maxima of ethanol (Plyler 1952), within the 325–4,500 nm transmission range of ApexTM photosensitive glass, most of them however are classified as very weak except for three strong maxima at 3.340, 3.359 and 3.42 lm. The ability to detect the weaker maxima in the NIR below 3 lm that positively identify the chemical absorption lines of ethanol, validates the purpose of this device. The OH absorption band at 2,700 nm (Plyler 1952) is found in both ethanol and the photosensitive glass. Transmission of radiation through glass in the mid-IR is attenuated by the water content in the glass. The absorption bands characteristic of the alcohol such as the bands at about 2,583 and 2,300 nm which were classified as ‘weak bands’ (Plyler 1952) and the band at 2,480 nm classified as ‘very weak’ were all detected. The band near 1,700 nm and the band at 1,180 nm which result from the CH-stretching in ethanol (Cho et al. 2005) are clearly observed. Other bands such as the free OH band at 1,580 nm and the overtone of the OH stretch at 1,450 nm (Cho et al. 2005) are also detected. Absorption in the UV region is limited by the scattering losses due to the surface roughness of the structure. With an RMS surface roughness of about 33 nm, the microstructure is not suitable for detection of signals shorter than about 330 nm. A schematic for Raman
123
Author's personal copy Microsyst Technol
Fig. 6 Illustration of the setup for Raman spectroscopy
Fig. 7 Raman spectra of DMSO in a microstructure made from ApexTM photosensitive glass and of DMSO on a standard quartz substrate
spectroscopy is shown in Fig. 6. A microcuvette was capped with a 500 lm thick ApexTM glass substrate that has undergone the exact fabrication process used to create the cuvette. Raman spectroscopy was then performed to measure the optical emission of a dimethyl sulfoxide (DMSO) liquid sample inside the cuvette. Raman scattering peaks associated with DMSO in the microcuvette and spectra of DMSO on a standard Raman spectroscopy quartz plate at 301 K are shown in Fig. 7. In particular, the strong bands at 677 and 702 cm-1 associated with the C–S–C stretching (Alves and Antunes 2007) were observed. The antisymmetric and symmetric bands at the high wavenumbers 3,004 and 2,910 cm-1, which are associated with the C–H stretching (Martens et al. 2002), were weakly observed in the photodefinable glass. Although the quartz substrate provided a superior Raman signature, fabricating microfluidic structures with optically smooth sidewalls in quartz is much more difficult than in photosensitive glass, requiring laser drilling and micromachining technologies (Ceriotti et al. 2003). Furthermore, microstructures made in photosensitive glass can be easily mass produced. The Raman spectra show that
123
Fig. 8 Common defects found in photosensitive glass structures: a Rough granules (shown on the left) after the etch stage resulting from back light reflection during exposure. b A defect in a sample baked at 40 "C higher than recommended temperature. c Silver precipitation on the glass surface
photosensitive glass demonstrates a high potential for microfluidic applications in which Raman spectral studies are required to be performed on fluids running in microfluidic devices. Unlike other technologies, defects associated with photosensitive glass technology are mainly related to the method of UV exposure and baking temperatures. Some of the common defects observed in improperly processed photodefinable glasses are shown in Fig. 8. A defect resulting from the back reflection of UV light rays during exposure is shown in Fig. 8a. This defect may be significantly reduced by proper adjustment of aligner tilts and exposure time, or eliminated using a laser source. A defect resulting from baking at temperatures higher than the optimal is shown in Fig. 8b. Whereas silver precipitation on the glass surface is an inherent defect (Williams et al. 2010) resulting from inhomogeneity in the glass matrix itself (shown in Fig. 8c). Excess silver redeposition is eliminated through the addition of a 10 % nitric acid rinse prior to the anneal step. That said, optimization of the exposure, etch, and post etch anneal results in nearly ideal structures with greater than 90 % repeatability.
4 Conclusion and future perspectives In this work we demonstrate the use of photostructurable glasses for in-plane analysis of microfluidic systems. This is shown by detecting different chemical substances in microstructures made in Apex photosensitive glass through the processed sidewalls. Ethanol with a concentration of 95 % was used as a sample specimen for UV–Vis–NIR transmission analysis, almost all weak and strong bands in the region 0.9–2.7 lm that are characteristic of the alcohol were identified. Raman spectroscopy of DMSO in a glass microcuvette was also presented. This technology eliminates the need for non-planar observation required for microfluidics analysis, allowing for stacks of microfluidic systems to be analyzed from the sides. Challenges associated with photosensitive glass technology are due to surface defects resulting from the back reflection of UV rays when a conventional contact aligner is used for
Author's personal copy Microsyst Technol
UV exposure, baking at temperatures higher than the optimal, and silver precipitation on the surface. Future work involves optical fiber IR transmission analysis of chemical and biological samples at different concentrations in microfluidic systems made in the two commercially produced photosensitive glasses, Apex and Foturan. A simulation software package will also be developed to use the RMS surface roughness and curvature in predicting the optimal dimensions and geometry of the microstructure to accompany these systems. Acknowledgments This research is sponsored by the Alabama EPSCoR Graduate Research Scholars Program, and supported by the Office of the Vice President for Research in the University of Alabama in Huntsville, Huntsville, AL, 35899
References Alves W, Antunes O (2007) Lewis acid–base adducts: a quantitative Raman analysis of formamide and dimethylsulfoxide mixtures. Spectrochim Acta Part A 67:847–851 Berdichevsky Y, Khandurina J, Guttman A, Lo YH (2004) UV/ozone modification of poly(dimethylsiloxane) microfluidic channels. Sens Actuators B 97:402–440 Bhatnagar P, Panda S, Edleman N, Allen S, Wise R, Mahorowala A (2007) Controlling line-edge roughness and reactive ion etch lag in sub-150 nm features in borophosphosilicate glass. J Appl Phys 101:076102 Birardaa G, Grencib G, Businarob L, Marmiroli B, Pacore S, Vaccari L (2010) Fabrication of a microfluidic platform for investigating dynamic biochemical processes in living samples by FTIR microspectroscopy. Microelectron Eng 87:806–809 Brandily-Anne ML, Lumeau J, Glebova L, Glebov L (2010) Specific absorption spectra of cerium in multicomponent silicate glasses. J Noncryst Solids 356:2337–2343 Burns DA, Ciurczak EW (2001) Handbook of near-infrared analysis. Marcel Dekker, Inc., New York Cai D, Neyer A, Kuckuk R, Heise H (2008) Optical absorption in transparent PDMS materials applied for multimode waveguides fabrication. Opt Mater 30:1157–1161 Ceriotti L, Weible K, de Rooij NF, Verpoorte E (2003) Rectangular channels for lab-on-a-chip applications. Microelectron Eng 67:865–871 Chen IJ, Lindner E (2007) The stability of radio-frequency plasma treated polydimethylsiloxane surfaces. Langmuir 23:3118–3122 Cheng Y, Sugioka K, Masuda M, Toyoda K, Kawachi M, Shihoyama K, Midorikawa K (2003) 3D microstructuring inside Foturan glass by femtosecond laser. Riken Rev 50:101–106 Cheng Y, Tsai HL, Sugioka K, Midorikawa K (2006) Fabrication of 3D microoptical lenses in photosensitive glass using femtosecond laser micromachining. Appl Phys A 85:11–14
Cho S, Chung H, Woo YA, Kim HJ (2005) Determination of water content in ethanol by miniaturized near-infrared (NIR) system. Bull Korean Chem Soc 26:115 Davies H (1954) The reflection of electromagnetic waves from a rough surface. Proc IEE Part IV Inst Monogr 101:209–214 Dietrich TR, Ehrfeld W, Lacher M, Kra¨mer M, Speit B (1996) Microelectron Eng 30:497–504 Dittrich PS, Manz A (2006) Lab-on-a-chip: microfluidics in drug discovery. Nat Rev Drug Discov 5:210–218 Flemming JH, Mack R, Buckley C (2009) Photosensitive glass for ultra high aspect ratio features. In: Nanotech conference and expo, Houston, TX Guenther K, Wierer P, Bennett J (1984) Surface roughness measurements of low-scatter mirrors and roughness standards. Appl Opt 23:3820–3836 Hu¨lsenberg D, Harnisch A, Bismarck A (2008) Microstructuring of glasses. Springer, Berlin Kim T, Campbell K, Groisman A, Kleinfeld D, Schaffer C (2005) Femtosecond laser-drilled capillary integrated into a microfluidic device. Appl Phys Lett 86:201106 Lin CH, Lee GB, Lin YH, Chang GL (2001) A fast prototyping process for fabrication of microfluidic systems on soda-lime glass. J Micromech Microeng 11:726–732 Martens WN, Frost RL, Kristof J, Kloprogge JT (2002) Raman spectroscopy of dimethyl sulphoxide and deuterated dimethyl sulphoxide at 298 and 77 K. J Raman Spectrosc 33:84–91 Ou J, Glawdel T, Ren C, Pawliszyn J (2009) Fabrication of a hybrid PDMS/SU-8/quartz microfluidic chip for enhancing UV absorption whole- channel imaging detection sensitivity and application for isoelectric focusing of proteins. Lab Chip 9:1926–1932 Pan T, Kelly RT, Asplund MC, Woolley AT (2004) Fabrication of calcium fluoride capillary electrophoresis microdevices for onchip infrared detection. J Chromatogr A 1027:231–235 Pantoja R, Nagarah JM, Starace DM, Melosh NA, Blunck R, Bezanilla F, Heath JR (2004) Silicon chip-based patch-clamp electrodes integrated with PDMS microfluidics. Biosens Bioelectron 20:509–517 Plyler EK (1952) Infrared spectra of methanol, ethanol, and n-propanol. J Res Natl Bureau Stand 48:281–286 Prim D, Crelier S, Segura JM (2011) Coupling of a microfluidic mixer to a Fourier transform infrared spectrometer for protein conformation studies. Chimia 65:815–816 Queste S, Salut R, Clatot S, Rauc JY (2010) Manufacture of microfluidic glass chips by deep plasma etching, femtosecond laser ablation, and anodic bonding. Microsystem Technol 16:1485–1493 Tantawi KH, Oates J, Kamali-Sarvestani R, Bergquist N, Williams JD (2011) Processing of photosensitive APEXTM glass structures with smooth and transparent sidewalls. J Micromech Microeng 21:017001 Williams JD, Schmidt C, Serkland D (2010) Processing advances in transparent Foturan glass MEMS. Appl Phys A 99:777–782
123
