Article pubs.acs.org/ac
Low-Temperature Microchip Nonaqueous Capillary Electrophoresis of Aliphatic Primary Amines: Applications to Titan Chemistry Morgan L. Cable, Amanda M. Stockton, Maria F. Mora, and Peter A. Willis* NASA Jet Propulsion Laboratory, California Institute of Technology, Pasadena, California 91109, United States S Supporting Information *
ABSTRACT: We demonstrate microchip nonaqueous capillary electrophoresis (μNACE) analysis of primary aliphatic amines (C1−C18) in ethanol down to −20 °C as a first step in adapting microfluidic protocols for in situ analysis on Titan. To our knowledge, this is the first report of a nonaqueous separation at −20 °C on-chip. Limits of detection (LODs) ranged from 1.0 nM to 2.6 nM, and we identified several primary amines ranging in length from C2 to C16 in Titan aerosol analogue (tholin) samples; new amines were also detected in a tholin sample exposed to oxygen and liquid water. This preliminary work validates the sensitivity and efficacy of microfluidic chemical analysis of complex organics with relevance to Titan aerosols and surface deposits. neutral) that can be separated. In addition, μCE with laserinduced fluorescence (LIF) detection instrumentation has been miniaturized and fully automated, with Mars as the planetary target.12 However, ultrasensitive detection techniques such as μCE−LIF are not limited to Mars, and could form the basis of a nondestructive analysis tool for interrogating complex organic mixtures on Titan. In nonaqueous capillary electrophoresis (NACE), an organic solvent or combination of solvents is used, usually methanol and/or acetonitrile.13,14 The advantages of NACE include solvation of more nonpolar species as well as decreased thermal diffusion and typically decreased electro-osmotic flow (EOF), resulting in enhanced resolution.11,15 The use of organic solvents also enables analysis over a broader temperature range than aqueous separations, and it opens up ion-pairing and acid/ base strength adjustment capabilities that are not possible in aqueous solutions.15,16 Here, we optimize fluorescent labeling and separation chemistry in nonaqueous solvents for μCE analysis of organics relevant to Titan. We focus on the detection of aliphatic primary amines: CH3(CH2)nNH2, where n = 0−17. As amino groups are found in compounds throughout the Titan atmosphere, primary amines are likely to be present in the aerosols and are presumed to exist in surface deposits as well.3,17,18 Amines are also a major constituent of Titan analogue organic material (tholins), regardless of variations in tholin generation conditions such as temperature, pressure, gas mixture, or energy source (cold plasma discharge, UV irradiation).19,20 Here, we report an optimized microchip nonaqueous capillary electrophoresis
T
itan, the largest moon of Saturn, is unique in the solar system because it is the only moon with a thick atmosphere and the only body aside from Earth with standing liquid on its surface. Because of the low temperatures in the atmosphere (70−190 K) and on the surface (90 K), the liquid phase of Titan is comprised of hydrocarbons, mainly methane and ethane.1 Short-wavelength ultraviolet radiation and energetic particles from the Sun and Saturn’s magnetosphere activate N2 and CH4 in the upper atmosphere;2 these reactive species form light hydrocarbons, amines, and nitriles, which, in turn, generate heavy hydrocarbons and more-complex organics at intermediate altitudes.3 At lower altitudes, organic aerosols combine and fall to the surface as larger particles,4,5 ultimately forming a layer of organics that is believed to cover nearly the entire surface of Titan.6 Thus far, the only in situ technique employed to study these complex organic samples was the Aerosol Collector Pyrolyzer (ACP) and gas chromatography−mass spectrometry (GC-MS) device on the Huygens Probe,7,8 which is a European contribution to the joint NASA−ESA Cassini−Huygens mission.1 However, this technology involves the rapid heating of samples to temperatures as high as 600 °C, which may lead to unpredictable side reactions and secondary product formation, obfuscating efforts to decipher the original sample composition.9 A recent Titan study10 suggested that liquidbased sample handling could allow for separation of complex mixtures on Titan, while minimizing perturbation of the sample. Microchip capillary electrophoresis (μCE) is a liquid-based technique that offers high performance, reagent economy, speed, and automation capabilities.11 Separation efficiencies in capillary electrophoresis (CE) and μCE are typically superior to other techniques, such as high-performance liquid chromatography (HPLC), as is the range of molecules (both charged and © XXXX American Chemical Society
Received: October 16, 2012 Accepted: December 7, 2012
A
dx.doi.org/10.1021/ac3030202 | Anal. Chem. XXXX, XXX, XXX−XXX
Analytical Chemistry
Article
and stored at −20 °C. All reagents were used without further purification. All aqueous solutions were prepared using deionized water (resistivity 18.2 MΩ) and filtered prior to use with disposable 0.22-μm sterile mixed cellulose esters membrane syringe filters (Fisherbrand, Fisher Scientific, Inc.); all solutions prepared in organic solvents were filtered using disposable 0.2 μm nylon membrane syringe filters (Fisherbrand). Tholin samples were provided by the laboratory of Dr. Mark Smith and Dr. Chao He at the University of Arizona as part of the collaborative effort of “Titan as a Prebiotic Chemical System”, a NASA Astrobiology Institute (NAI) endeavor. Two tholin samples were analyzed in this study and are classified based on formation conditions, age, and exposure to oxygen. One tholin sample, which we will refer to as Type I tholin, was generated using a cold plasma (AC) discharge of a 95% N2, 5% CH4 atmosphere for 72 h at −108 °C (165 K). The reaction vessel was allowed to warm to room temperature under vacuum for 24 h, after which the solid material was isolated in a dry, oxygen-free glovebox and stored under N2 atmosphere.22 The sample was shipped under N2 atmosphere and stored at −80 °C prior to use (5 days). A second sample, which we define as Type II tholin, was generated in the same laboratory in 2006 under similar discharge conditions23 at 30 °C (303 K), stored at room temperature, and was exposed to air prior to use. μNACE−LIF Method. Primary amines are readily labeled under basic conditions using PB (λex = 405 nm, λem = 455 nm) to generate a stable fluorescent amide (see Figure S1 in the Supporting Information).24 Unless otherwise specified, amines were labeled at high concentration (20 μM) in ethanol with a 10-fold excess of PB and 25 mM triethylamine to maintain basic conditions. After reaction overnight (16 h), samples were diluted 1:200 into a separation buffer for analysis. The microfluidic channel was preconditioned prior to each analysis with 0.1 M HCl(aq), followed by deionized water (resistivity 18.2 MΩ) and ethanol. The channel was then filled with separation buffer, and an aliquot of sample was pipetted into the sample well (separation buffer was pipetted into the other three wells). Injection potentials for μNACE were 2 kV from the sample (negative) to the waste well (positive), with bias applied to the other two wells (−0.4 and −0.8 kV) to pinch the injection plug for the 30 s injection. Separation was performed at a potential of 5 kV along the separation channel, with the positive potential at the end opposite the cross. All analyses were performed in triplicate at 20 °C with the exception of the temperature dependence study. Data fitting was performed with PeakFit (Seasolve Software, Inc., v4.12).
(μNACE) separation protocol for primary amines in ethanol with LIF detection. We determine limits of detection for this analysis method and demonstrate separations at low temperature (−20 °C). We also use this technique to identify primary amines in two tholin samples. Finally, we comment on the implications of this work in terms of in situ analysis on Titan.
■
EXPERIMENTAL METHODS Instrumentation and Microdevices. The microchip capillary electrophoresis (μCE) system with laser-induced fluorescence (LIF) detection was constructed specifically for this study on a small optical table; the entire system measures 52 cm × 25 cm × 50 cm, excluding the chiller used for temperature control. Briefly, a 405-nm, 16.0 mW laser beam (Power Technology, Inc., Little Rock, AR) is directed through an excitation filter (FF01-405/10-25, Semrock, Rochester, NY), to a dichroic filter (Di01-R405-25, Semrock) and through a 20× microscope objective (Plan Fluor 20X/0.50 DIC M/N2 ∞/0.17 WD 2.1, Nikon, Japan), focusing the beam ∼2 mm above the objective. The microfluidic chip is placed on a custom temperature-controlled stage (connected to a Melcor Liquid Chiller, Model MRC 300DH2-DV, Trenton, NJ) with the center of the channel at the focal point. LIF is directed through an emission filter (BLP01-405R-25, Semrock) and focused (using a Model LA1027-A-F35 lens) onto the entry window of a photomultiplier tube (PMT) (Model H9656-20; Model C7169 power supply, Hamamatsu, Japan). The PMT signal is processed using a data acquisition card (DAQCard 6036E, National Instruments, Austin, TX) and a laptop computer. A high-voltage sequencer (Model HVS448 6000D, LabSmith, Livermore, CA) is used to control the electrode potentials for μCE. Both the high-voltage sequencer and the PMT data acquisition are controlled by a custom program written in LabVIEW (National Instruments, 2011). Although integrated multilayer microfluidic devices designed and fabricated in our laboratory are capable of autonomous fluidic manipulation and separations,12,21 for the work reported here, commercial two-layer glass devices were used (Micralyne, Inc., Edmonton, Canada). The BF4-SC chips (16 mm × 95 mm × 2.2 mm) are comprised of low-fluorescence Schott Borofloat glass and contain two channels (8.0 mm and 85.0 mm) that intersect perpendicularly to form a 4-port simple cross with three arms of equal length (4.0 mm). The channel cross-section is approximately semicircular, with dimensions of 50 μm width by 20 μm depth. Nanoports (Upchurch, Idex Health & Science, Oak Harbor, WA) were adhered to the glass surface above each well, using port adhesive (N-100-01, Idex Health & Science), and sealed with epoxy to provide reservoirs for the sample/buffer. Materials. Methylamine (C1) hydrochloride, ethylamine (C2) hydrochloride, propylamine (C3), amylamine (C5), hexylamine (C6), nonylamine (C9), hexadecylamine (C16), and octadecylamine (C18) were purchased from Sigma− Aldrich; 1-dodecylamine (C12) was purchased from Alfa Aesar. All amines were 97% purity or greater, with the exception of hexadecylamine (90%). Ethyl alcohol (99.5%, ACS reagent) was purchased from Acros Organics. Ammonium acetate (99.99%) and triethylamine (99.5%) were purchased from Sigma−Aldrich. Acetic acid (glacial, 17.4 N, 99.9%), dimethylsulfoxide (DMSO, ACS reagent grade), and N,Ndimethylformamide (DMF, ACS reagent grade) were purchased from Fisher Scientific. Pacific Blue succinimidyl ester (PB) was purchased from Invitrogen (Life Technologies Corp.)
■
RESULTS AND DISCUSSION We implemented μNACE for this work for three reasons: (1) long aliphatic chains tend to be more soluble in organic solvents, (2) tholins have been reported to undergo hydrolysis in aqueous solution or when exposed to water vapor,22,25−28 and (3) the low viscosities and freezing points of some organic solvents may enable ultralow temperature chemical analyses on icy worlds such as Titan. We selected ethanol due to its low freezing point (159 K), high solubility of both short- and longchain aliphatic amines, and because it is a well-known solvent for NACE.15,29 We also note that, because μNACE is nondestructive, separations in a solvent of low viscosity such as ethanol facilitate future coupling of this technique to a downstream analysis tool such as mass spectrometry (MS), which would enable a second dimension of analysis. B
dx.doi.org/10.1021/ac3030202 | Anal. Chem. XXXX, XXX, XXX−XXX
Analytical Chemistry
Article
Optimization of μNACE. As aliphatic amines are not intrinsically fluorescent, we utilized a fluorescent dye to label this functional group and enable detection with high sensitivity. PB imparts a 200-fold improvement in sensitivity over other amine-reactive fluorescent probes (e.g., fluorescamine) and has been implemented for amine and amino acid detection on the Mars Organic Analyzer, which is a portable μCE−LIF instrument,30 as well as other laboratory systems.12 Because the PB labeling reaction must be performed under basic conditions,21 we utilized triethylamine, which is an organic base with a high solubility in ethanol. A standard containing 100 nM of C3, C6 and C9 amines was used to optimize labeling and separation conditions. For labeling optimization, the amine standard was prepared with concentrations of triethylamine ranging from 5 mM to 200 mM, and μNACE was performed with a separation buffer containing 100 mM ammonium acetate and 1.0 M acetic acid in ethanol. High concentrations (≥50 mM) of triethylamine resulted in the appearance of an ethylamine (C2) contamination peak (see Figure S2 in the Supporting Information), whereas low concentrations (≤5 mM) did not allow the labeling reaction to go to completion within 24 h. Therefore, an optimal concentration of 25 mM triethylamine was used for subsequent experiments. A brief kinetics study was also performed to determine the efficiency of the labeling reaction over time and the stability of the labeled product. The amine standard (which also included C16 for this study) was labeled, and aliquots of 1 μL were removed periodically and diluted to 100 nM for μNACE analysis using a separation buffer containing 100 mM ammonium acetate and 1.0 M acetic acid in ethanol. Results indicated that labeling of primary amines with PB in ethanol was rapid (
