Ion Separation in Air Using a Three-Dimensional ... - ACS Publications

Article pubs.acs.org/ac

Ion Separation in Air Using a Three-Dimensional Printed Ion Mobility Spectrometer Adam Hollerbach,† Zane Baird,‡ and R. Graham Cooks*,† †

Chemistry Department, Purdue University, 560 Oval Drive, West Lafayette, Indiana 47907, United States S Supporting Information *

ABSTRACT: The performance of a small, plastic drift tube ion mobility spectrometer (DT-IMS) is described. The IMS was manufactured using threedimensional (3D) printing techniques and operates in the open air at ambient pressure, temperature, and humidity. The IMS housing and electrodes were printed from nonconductive polylactic acid (PLA, housing) and conductive polyethylene terephthalate glycol-modified polymer containing multiwalled carbon nanotubes (PETG-CNT, electrodes). Ring electrodes consisting of both an inner disk and an outer ring were used to prevent neutral transmission while maximizing ion transmission. As a stand-alone instrument, the 3D printed IMS is shown to achieve resolving powers of between 24 and 50 in positive ion mode using tetraalkylammonium bromide salts (TAA), benzylamines (mono-, di-, and tri-), and illicit drugs (MA, MDEA, and haloperidol). Resolving powers of between 29 and 42 were achieved in negative ion mode using sodium alkyl sulfates (C8, C12, C16, and C18). Reduced ion mobilities of TAA cations (C2−C8) were calculated at 14% relative humidity in air to be 1.36, 1.18, 1.03, 0.90, 0.80, 0.73, and 0.67, respectively. The effect of humidity on reduced ion mobilities of TAA cations is discussed. 3D printing is shown to be a quick and cost-effective way to produce small IMS instruments that can compete in performance with conventionally manufactured IMS instruments that also operate in the open air. An important difference between this IMS and other instruments is the absence of a counter gas flow.

T

a background gas.8,15,17−21 A typical DT-IMS instrument consists of a drift tube made of a series of metal rings. A voltage gradient is established by applying decreasing DC voltages to successive rings. Ions injected into the drift tube follow the voltage gradient toward the detector where they are distinguished by time spent in the drift tube. Ion separation occurs due to collisions with a background gas, which is often flowed in the direction opposite to ion motion to lengthen ion drift time and increase resolving power. The pressure and composition of the background gas plays a critical role in IMS. Most DT-IMS instruments operate using nitrogen or helium as the background gas at pressures around 1 Torr.1 However, there are a few examples of DT-IMS instruments which utilize atmospheric air as background gas. Baether et al. used a nonradiative electron ionization source and a counterflow of atmospheric air in their modified AP-DT-IMS to successfully analyze toluene-2,4-diisocyante (TDI) and dimethyl-methylphosphonate (DMMP).22,23 One likely reason for the small number of reported DT-IMS systems that operate using atmospheric air is that the composition of air can change considerably throughout a single day. As will be discussed later, moisture levels affect mobility data.

here is an increasing demand for portable high performance analytical instrumentation. Ion mobility spectrometry (IMS) is a field which has seen rapid development.1,2 High performance commercial IMS systems are manufactured by conventional machining methods and are used for a range of applications, including the analysis of illicit drugs3,4 and explosives,5 among other analytes.6 High performance is routinely achieved in both stand-alone IMS instruments and powerful ion mobility-mass spectrometry (IM-MS) hybrids.7,8 Notable examples are found in the work of Clemmer et al.9 on the separation of protein conformers by IM-MS and peptide characterization using multiple stage ion mobility systems. Ion mobility instruments have also been built to function over a range of pressures, from 1 Torr to atmospheric pressure (APIMS).10,11 It is advantageous to operate ion mobility instruments in the air because the instrumentation then becomes simpler. The challenge for current AP-IMS instrumentation is to achieve performance comparable to that of more conventional, vacuum-based instruments.1,12 Drift tube ion mobility spectrometers (DT-IMS) are ideal systems to fulfill the needs of high performance and portability because they are known to achieve high resolving powers and are electronically simple.13 Originally termed “plasma chromatography” by Karasek14,15 and later “ion chromatography” by Bowers,16 DT-IMS was one of the first ion mobility methods invented. In DT-IMS, gas phase ions are introduced into a drift cell containing a relatively weak electrostatic field and are separated based on their size, charge state, and interaction with © 2017 American Chemical Society

Received: February 7, 2017 Accepted: April 1, 2017 Published: April 6, 2017 5058

DOI: 10.1021/acs.analchem.7b00469 Anal. Chem. 2017, 89, 5058−5065

Article

Analytical Chemistry

Figure 1. Renderings of (a) 3D printed IMS, (b) focusing and drift electrodes, and (c) individual housing pieces.

form of 3D printing. The plastic IMS was used to separate and focus ions in the open air, a recent concept.44 3D printed structures have previously been used to separate and focus ions in the open air,45 and now, a functional analytical instrument has been constructed using 3D printed plastic structures. FDM printers offer the ability to produce structures with different geometries in-lab at low cost. Often, these structures can be used the same day that they were designed, a feature usually unavailable by conventional machining methods. The goal of this study was to compare the performance of a 3D printed ion mobility spectrometer operating in the open air to conventionally manufactured IMS systems. Resolving powers and reduced ion mobilities were calculated for tetraalkylammonium cations (C2−C8), benzylamines, illicit drugs, and sodium alkyl sulfates. A discussion on the effects of humidity on atmospheric pressure IMS systems is given.

A collective term, the reduced ion mobility (K0), is used to standardize measurements obtained from different ion mobility spectrometers to a gas number density of 2.69 × 1025 m−3, a temperature of 273 K, and a pressure of 760 Torr:24,25 K0 =

P 273 L2 1 760 T V td

(1)

where K0 is the reduced ion mobility typically reported in cm2 s−1 V−1. This value is used in the determination of an ion’s collisional cross section (CCS). Several research groups have tabulated CCS values for proteins,26−30 peptides,31−34 tryptic digests,35,36 oligonucleotides,37 carbohydrates and lipids,38 pharmaceutical compounds, quaternary amines, polycyclic hydrocarbons, and fullerenes39 using different gas compositions, ionization methods, and types of ion mobility spectrometers, providing guidance for other researchers in calibrating their own systems. If accurate K0 values cannot be determined directly, calibration is required. Hill and co-workers proposed calibrating nonideal and/or contaminated instruments by using standards which possess well-known mobilities and which closely resemble the desired analyte in size.40 Correction factors then need to be calculated for each analyte under every set of temperatures, pressures, humidity levels, and drift voltages using the relationship:41 K 0(unknown) K 0(standard)

=

■

EXPERIMENTAL SECTION 3D Printing Parameters. The plastic IMS was manufactured using a modified fused deposition modeling (FDM) 3D printer based on the MakerFarm Prusa i3 design (Makerfarm, UT, United States). The plastic housing and electrodes used to construct the IMS were designed in Autodesk Inventor. The drawings were exported as stereolithography (.stl) files and imported into the slicing software Simplify3D. The housing for the IMS was made from a mixture of polylactic acid (PLA) and polyhydroxyalkanoate (PHA) polymers (Colorfabb, Venlo, Netherlands). The electrodes were made from polyethylene terephthalate glycol-modified polymer doped with multiwalled carbon nanotubes (PETG-CNT) (3DXTech, MI, United States). The carbon nanotubes gave the PETG polymer sufficient conductivity to allow for high voltage applications. Table S1 shows optimized printing parameters for each plastic component. A description of FDM is also provided in the Supporting Information for the interested reader. Housing and Electrode Design. Specific dimensions of the housing and electrodes have been described previously.46 More details regarding design of the IMS housing and electrodes are also provided in the Supporting Information. Briefly, the IMS consisted of six different plastic parts: (1) the housing, (2) a focusing electrode, (3) an injection electrode, (4) 30 drift electrodes, (5) housing for a Faraday cup detector,

td(standard) td(unknown)

(2)

DT-IMS instruments typically suffer from low duty cycles and can suffer from low ion transmission efficiencies when long drift tubes are used. However, Bush et al. were able to improve ion transmission efficiencies in a commercial DT-IMS by applying a radiofrequency (RF) waveform to the drift cell with a 180° phase difference between adjacent electrodes in addition to an applied DC voltage gradient.42,43 The applied RF significantly reduced the number of ions lost through collisions with electrodes by radially confining the ions to the center of the DT-IMS. One novel aspect of the work presented in this study comes from the fact that a drift tube ion mobility spectrometer was manufactured by fused deposition modeling (FDM), which is a 5059


Article


Figure 2. Ion mobility spectra of (a) tetraalkylammonium cations (C2−C8), (b) drugs MA, MDEA, and haloperidol, and (c) monobenzylamine, dibenzylamine, and tribenzylamine, all acquired in positive ion mode, and (d) sodium alkyl sulfate anions (C8, C12, C16, and C18), acquired in negative ion mode. All spectra were acquired using a 0.30 ms injection time.

to each drift electrode, establishing a 270 V/cm drift field strength. The voltage divider circuit was placed in its own plastic housing and could be easily inserted into the ends of the electrodes in the IMS. Ion injection was performed by applying floated high voltage pulses from a high voltage switch (GHTS 60A, Behlke, MA, United States) at a rate of 10 Hz to a 67% open steel mesh (E-Fab, Santa Clara, CA, United States) placed in contact with the injection electrode. A pulse generator (Model DG535, Stanford Research Systems, CA, United States) supplied the initial injection waveform to the high voltage switch. Negative ion mode experiments were performed analogously to those in positive ion mode. A grounded Faraday cup detector consisting of a circular copper plate and a soldered copper wire was used to detect injected ion packets. The Faraday cup wire was connected to a Keithley 428 current amplifier with an inverting output (Keithley Instruments, OH, United States). Data were recorded using a Tektronix TDS 2024C 4-channel digital oscilloscope (Tektronix, OR, United States). Humidity, temperature, and pressure measurements were conducted just outside of the IMS using a NIST traceable digital barometer (Cole-Palmer, IL, United States). The barometer was equilibrated for several hours prior to performing experiments. In each experiment presented in this

and (6) housing for a voltage divider circuit (Figure 1). Only 1−4 are displayed in Figure 1 for ease of viewing. Each drift electrode consisted of both a 30 mm diameter outer ring and a 15 mm diameter centered inner disk held in place by three 7.5 × 0.7 × 0.8 mm support columns. A focusing electrode was placed at the front of the IMS to confine the electrospray plume and possessed the same dimensions as the drift electrodes, except it had a 30 mm extended cylinder on its front. After ions were injected into the IMS, they traversed the region between the inner and outer electrode rings where they separated axially due to collisions with stagnant air molecules. The total drift length was 79.18 mm when all electrodes were placed into the housing. This value is the distance from the middle of the injection electrode (during injection, V = 2500) to the detector. A simulation showing ion trajectories in the focusing electrode has been described previously.46 Electronics. The electronics used to control the IMS have been previously described.46 More information, including a block diagram (Figure S1) and table of voltage parameters (Table S2), is provided in the Supporting Information. Briefly, a custom-built voltage divider circuit consisting of 34 10 kΩ precision resistors (Mouser Electronics, TX) and gold-plated headers (2.56 mm pitch) was used to supply high DC voltages 5060


Article


used to smooth the data. Full-width at half-maximum (fwhm) values were obtained by fitting detected peaks to Gaussian profiles via the polyf it function available in MATLAB.

study, the humidity changed by only a few tenths of a percent or less. However, as is discussed later, this measurement is a source of error when attempting to calculate reduced ion mobilities directly without calibration. Chemicals. Four different sets of chemicals were used in this study. The first set consisted of seven tetraalkylammonium bromide salts (TAA) with different hydrocarbon chain lengths: tetraethylammonium bromide (TAA, C2), tetrapropylammonium bromide (TAA, C3), tetrabutylammonium bromide (TAA, C4), tetrapentylammonium bromide (TAA, C5), tetrahexylammonium bromide (TAA, C6), tetraheptylammonium bromide (TAA, C7), and tetraoctylammonium bromide (TAA, C8). The second set of compounds consisted of three different illicit drugs: methamphetamine (MA), 3,4-methylenedioxy-N-ethylamphetamine (MDEA), and haloperidol. The third set consisted of three different benzylamines: monobenzylamine (MBA), dibenyzlamine (DBA), and tribenzylamine (TBA). The last set consisted of four different sodium alkyl sulfates (SAS) of different hydrocarbon chain lengths: sodium octyl sulfate (SAS, C8), sodium dodecyl sulfate (SAS, C12), sodium hexadecyl sulfate (SAS, C16), and sodium octadecyl sulfate (SAS, C18). All TAA salts, benzylamines, and SASs were purchased as powders (Sigma-Aldrich, MO, United States). MA, MDEA, and haloperidol were purchased as 1 mg/mL methanolic solutions (Cerilliant, TX, United States). Working solutions of TAA salts and sodium alkyl sulfates were prepared in acetonitrile (Sigma-Aldrich), while working solutions of benzylamines and illicit drugs were prepared in acetonitrile with 1% formic acid (Sigma-Aldrich). All working solutions of compound mixtures and individual compounds were prepared at 1 mM concentrations except for the solution containing a mixture of TAA salts, which was prepared at 100 μM. This more dilute solution was used to minimize space charge effects observed with more concentrated mixtures. It should be noted that the sodium hexadecyl sulfate standard was received with a contaminant of sodium octadecyl sulfate as confirmed by MS. However, it will be shown later that this contamination was actually beneficial for this study. Electrospray Ionization and Mass Spectrometer Parameters. Borosilicate glass capillaries (Sutter Instruments, CA) were pulled into nanospray emitters using a Sutter micropipette puller (Model P-97, Sutter Instruments, CA). All nanospray tip diameters were between 3−5 μm. A 0.51 mm platinum 10% iridium wire (California Fine Wire Company, CA, United States) was used to supply voltage to the solutions. Low-resolution mass analysis was performed on a linear ion trap mass spectrometer (LTQ, Thermo Fisher, CA, United States) in both positive and negative ion modes. The spray voltage, capillary voltage, tube lens voltage, and capillary temperature were ±1.5 kV, ±15 V, ±65 V, and 150 °C, respectively. Full scan mass spectra were acquired using three microscans and 150 ms injection times, while collision-induced dissociation (CID) spectra were acquired using three microscans and 500 ms injection times. All mass spectra were averaged in the Xcalibur software and exported to MATLAB, where spectra were replotted to provide high resolution images. Data Analysis. All IMS spectra are averages of 128 scans and were background subtracted using an average background spectrum. MATLAB was used to replot the spectra from the CSV files obtained from the oscilloscope. The x-axis in all spectra are plotted from 2 to 90 ms to eliminate interference from the injection waveform, which was significantly larger than the ion packet signal. A 3-point centered moving average was

■

RESULTS AND DISCUSSION The performance of the 3D printed IMS was tested by using standard solutions ionized by nanoelectrospray ionization (nESI). The position of the nanospray emitter was optimized in each experiment to provide the highest signal obtainable. The performance of the IMS in positive ion mode was first tested by electrospraying a mixture of TAA cations (C2−C8) (Figure 2a). Each TAA salt was present at 100 μM. Spectra were acquired using a 0.30 ms injection time. As can be seen, all seven compounds were baseline resolved. Interestingly, the peak intensities of the TAA cations were observed to decrease as the size of the cation increased. Upon closer inspection, the C2 peak was approximately four times larger than the C8 peak. Larger cations might be injected less efficiently into the drift region due to their inherently lower mobilities. IMS spectra taken using solutions of individual TAA cations also showed that heavier TAA cations produced lower signal intensities when the same concentrations and injection times were used (Figure S2). Note, too, that no changes in peak positions were observed in the course of a run, indicating that deposition of material on the insulating plastic surfaces does not occur. A solution of MA, MDEA, and haloperidol in acetonitrile with 1% formic acid was also analyzed in positive ion mode (Figure 2b). Each analyte was present at 1 mM. All three compounds were resolved, though baseline separation was not achieved. The first peak (drift time ∼18 ms) was present in solvent blanks and is probably caused by formic acid−water clusters. Haloperidol possessed the longest drift time and yielded a signal intensity lower than those of either MA or MDEA, which is likely because of the same reduced efficiency observed for the larger TAA cations. However, a significant amount of tailing was observed for both MA and MDEA, while haloperidol exhibited only a small amount of tailing. It is important to emphasize that the 3D printed IMS is not heated and does not currently have a desolvation region. This means tailing is likely to occur due to incomplete desolvation and/or adduction of ambient water in the drift cell.47 The performance of the IMS was further characterized in positive ion mode by analyzing a mixture of mono-, di-, and tribenzylamines in 1% formic acid (Figure 2c), each at 1 mM. Once more, all three compounds were resolved without baseline separation, and some tailing was also observed. Interestingly, the MBA peak was significantly suppressed compared to the DBA and TBA peaks. It is possible that space charging caused this phenomenon because the IMS possesses a rather small injection region, and 1 mM concentrations of analyte were used. Spectra of individual solutions of MBA, DBA, and TBA did not exhibit suppression when they were analyzed using the same concentration and injection time as the mixture (Figure S3). The Coulombic forces in the ion cloud may cause expulsion of the light MBA ions from the injection region but may not be strong enough to expel significant amounts of the heavier DBA or TBA ions. Interestingly, fewer ion mobility standards are reported in the literature for negative ion mode than for positive ion mode.48,49 Some researchers have proposed fatty acids as negative ion mode standards because they possess a wide range of masses.25,50 However, it was discovered in this study that sodium alkyl sulfates could be used as good negative ion mode 5061


Article


field. Similar effects were also observed for the mixtures of TAA cations, illicit drugs, and benzylamines (Figure S5). Interestingly, the same four effects were again observed when the injection time was kept constant at 0.5 ms and the spray voltage was varied from 4 to 5.5 kV in 500 V increments (Figure S6). These data suggest that space charge can contribute significantly to peak shapes. The resolving power of an ion mobility spectrometer is commonly calculated by dividing the mean drift time (td) of an analyte by the fwhm of the analyte peak:52

standards because they possess higher molecular weights and noticeably higher ionization efficiencies compared to those of fatty acids. As a further performance test, a solution containing three sodium alkyl sulfates with different carbon chain lengths (C8, C12, and C16) at 1 mM each was analyzed in negative ion mode (Figure 2d). Like the TAA cations, the sodium alkyl sulfates appeared to be rather unreactive with water, as indicated by the lack of peak tailing. Surprisingly, spectra collected from the three-component sodium alkyl sulfate solution showed an unexpected fourth resolved peak. This fourth peak eluted at a drift time slightly longer than that of the SHDS peak. To determine the origin of this peak, mass analysis was performed on individual sodium alkyl sulfate solutions (Figure S4). The results showed that the original SHDS powder also contained a relatively large amount of sodium octadecyl sulfate (SODS, C18) in addition to SHDS (Figure S4d). This result shows that the 3D printed IMS can easily separate ions of 28 Da mass difference. In general, spectra collected in negative ion mode were slightly noisier than spectra collected in positive ion mode. The presence of small background waves (Figure 2d, drift times ∼30−90 ms) in negative ion mode can be explained by electronic instabilities associated with the wiring required to operate in negative ion mode. The injection time of the IMS was adjusted to monitor its effect on the signal intensities of sodium alkyl sulfate anions (Figure 3). The different colored traces represent spectra

Rp =

td μ = fwhm 2.355σ

(3)

In this study, MATLAB was used to fit IMS spectra to Gaussian functions to obtain both td and fwhm values. Mean drift times were obtained by using the mean value of the fitted Gaussian (μ), and fwhm values were obtained by multiplying the standard deviation of the fitted Gaussian (σ) by 2.355 pursuant to the definition of fwhm for a Gaussian. The TAA cations yielded the highest resolving powers at 0.30 ms injection times of all the compounds analyzed, with values between 40−50 (Figure 4a). The TAA C7 cation yielded the highest individual resolving power of 49.7. An AP-DT-IMS system reported in the literature gave resolving powers between 42 and 54 for trichloroethylene, tetrachloroethylene, methyl tert-butyl ether (MTBE), and methyl isobutyl ketone (MIBK), which are common environmental contaminants.52 The resolving powers obtained using the 3D printed IMS compare quite well to these literature values. Other compounds analyzed in positive ion mode yielded resolving powers between 24 and 36 for 0.30 ms injection times (Figures 4b and c), while the sodium alkyl sulfates yielded slightly higher values between 31 and 43 in negative ion mode (Figure 4d). As can be seen, the use of longer injection times tended to decrease the resolving power of the IMS. While the use of short injection times improved the resolution of all peaks, the improvement was accompanied by a decrease in signal intensity, which is typical of all analytical instrumentation. Reduced ion mobilities were calculated according to eq 1 for solutions of individual TAA cations using a pressure, temperature, and relative humidity of 757 Torr, 297 K, and 14%, respectively. To ensure accurate drift times for each compound were obtained, every spray emitter used in this part of the study was placed at the distance which yielded a peak intensity of 25 mV instead of the highest intensity possible. The reduced ion mobilities of all the TAA cations were calculated to be significantly lower than the reduced ion mobilities reported in the literature for both nitrogen and helium drift gases (Table 1, columns 1−3).48 Interestingly, the smaller TAA cations (C2− C4) showed deviations from reported values greater than those of the larger TAA cations (C5−C8) (Table 1, column 4). The difference between the measured reduced ion mobility values and the literature values can likely be attributed in part to the presence of ambient water inside the IMS as well as incomplete desolvation of analyte ions. Even though humidity is not considered in the calculation of reduced ion mobilities (eq 1), high humidity levels are known to increase the drift times of analytes.53 Because reduced ion mobility and drift time are inversely proportional, a significant increase in an analyte’s drift time due to intermolecular interactions with water will cause a decrease in the calculated reduced ion mobility. It should be emphasized that the humidity, pressure, and temperature measurements used in the reduced mobility calculations were

Figure 3. Overlaid mobility spectra showing the effect of drift time on the peak shapes of a 1 mM mixture of SAS anions. The highest resolutions were obtained using 0.30 ms injection times (green trace), while increased signal intensities were obtained using longer injection times of 0.65 and 1.00 ms (blue and red traces, respectively).

recorded for the same solution at injection times of 0.30, 0.65, and 1.00 ms. When short injection times were used (green trace), baseline separation between each SAS anion was observed. Several effects were observed when longer injection times were used (blue and red traces): increased peak intensities, peak widths, background signal between peaks, and a slight increase in the mean drift time of each compound. Although these four effects are commonly observed with increases in the injection times of most IMS systems, the last three effects are likely enhanced by space charging associated with a small injection region51 and a lack of a radially confining 5062


Article


Figure 4. Plots of resolving power versus injection time for mixtures of (a) TAA cations, (b) MA, MDEA, and haloperidol, (c) benzylamines, and (d) SAS anions.

Table 1. Calculated and Calibrated Reduced Ion Mobilities Obtained from Individual Solutions of TAA Saltsa TAA

measured K0 values

literature K0 values48

K0(lit)/ K0(meas)

measured drift times (ms)

K0 using C2

K0 using C3

K0 using C4

K0 using C5

K0 using C6

K0 using C7

K0 using C8

C2 C3 C4 C5 C6 C7 C8

1.36 1.18 1.03 0.90 0.80 0.73 0.67

1.88 1.56 1.33 1.15 1.02 0.92 0.84

1.38 1.32 1.29 1.28 1.27 1.27 1.26

16.83 19.42 22.32 25.48 28.58 31.58 34.40

1.88 1.63 1.42 1.24 1.11 1.00 0.92

1.80 1.56 1.36 1.19 1.06 0.96 0.88

1.76 1.53 1.33 1.17 1.04 0.94 0.86

1.74 1.51 1.31 1.15 1.03 0.93 0.85

1.73 1.50 1.31 1.14 1.02 0.92 0.85

1.73 1.50 1.30 1.14 1.02 0.92 0.84

1.72 1.49 1.29 1.13 1.01 0.91 0.84

a

Bold values are out of the acceptable error range based on ±0.02 ms.31

has a significant effect on the performance of the 3D printed IMS. An attempt to calibrate the IMS was made by specifying different individual TAA cations as the calibrant and substituting the values for all the other TAA cations into eq 2 (Table 1, columns 6−12). Reduced ion mobilities are considered reasonable if the calculated values are within about ±0.02 cm2 s−1 V−1 of reported literature values.31 For the calculations presented in Table 1, values that fell outside of the acceptable error range are shown in bold text. The data shown in Table 1 indicate that both the C2 and C3 TAA cations were unacceptable calibrants for the larger TAA cations. However,

recorded just outside of the IMS instead of inside and thus could be prone to errors. To explore the relationship between humidity and TAA cation size, experiments were performed in which the Faraday cup detector was modified to allow for incorporation of a nitrogen bath gas. Several of the same TAA solutions previously mentioned were electrosprayed into the IMS. It was observed that drift times decreased for all TAA cations. The change in drift time was most significant for TAA C2, which changed by about 1.5 ms, while larger TAAs decreased only slightly. These preliminary experiments are mentioned because they support the suggestion that moisture 5063



■

acceptable calibrated reduced ion mobilities were obtained when the C4, C5, and C6 TAA cations were used as calibrants. This same effect has been reported for the C4 and C5 TAA cations.40 The moderately sized TAA cations are also likely good calibrants because their reduced ion mobilities fall in the middle of the values for the other analytes. Interestingly, all reduced ion mobilities obtained using C3 as a calibrant were only slightly outside the acceptable error range. The C4 reduced ion mobility was different only by ±0.03 cm2 s−1 V−1, and the C5−C8 reduced ion mobilities were different by ±0.04 cm2 s−1 V−1. A possible reason for the large discrepancy in reduced ion mobilities calculated using C2 and C3 as calibrants is that ion− molecule interactions taking place inside the IMS between analytes and water delay the transmission of smaller TAA cations through the IMS more than they do for larger TAA cations. Further study will likely elucidate this phenomenon.

CONCLUSIONS A plastic drift tube ion mobility spectrometer was constructed using a 3D printer and operated in the open air. The IMS was shown to perform comparably to conventionally manufactured ion mobility spectrometers operating under atmospheric conditions. It is possible that the resolution of the IMS might be increased and the effects of humidity decreased by employing a nitrogen counter flow gas. A larger injection region would also likely decrease the contributions from space charging. One advantage of 3D printing is that new designs can be rapidly constructed and tested. ASSOCIATED CONTENT

S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.analchem.7b00469. Description of models and designs; parameter tables; block diagram; and IMS, mass, and overlaid mobility spectra (PDF)

■

REFERENCES

(1) Cumeras, R.; Figueras, E.; Davis, C. E.; Baumbach, J. I.; Gràcia, I. Analyst 2015, 140, 1376−1390. (2) Yamaguchi, S.; Asada, R.; Kishi, S.; Sekioka, R.; Kitagawa, N.; Tokita, K.; Yamamoto, S.; Seto, Y. Forensic Toxicol. 2010, 28, 84−95. (3) Forbes, T. P.; Najarro, M. Analyst 2016, 141, 4438−4446. (4) Armenta, S.; Garrigues, S.; de la Guardia, M.; Brassier, J.; Alcalà, M.; Blanco, M.; Perez-Alfonso, C.; Galipienso, N. Drug Test. Anal. 2015, 7, 280−289. (5) Harris, G. A.; Kwasnik, M.; Fernández, F. M. Anal. Chem. 2011, 83, 1908−1915. (6) Armenta, S.; Alcala, M.; Blanco, M. Anal. Chim. Acta 2011, 703, 114−123. (7) Lanucara, F.; Holman, S. W.; Gray, C. J.; Eyers, C. E. Nat. Chem. 2014, 6, 281−294. (8) Kanu, A. B.; Dwivedi, P.; Tam, M.; Matz, L.; Hill, H. H. J. Mass Spectrom. 2008, 43, 1−22. (9) Clemmer, D. E.; Hudgins, R. R.; Jarrold, M. F. J. Am. Chem. Soc. 1995, 117, 10141−10142. (10) Eiceman, G. A.; Karpas, Z.; Hill, Jr., H. H. Ion mobility spectrometry, 3rd ed.; CRC Press: Boca Raton, 2013. (11) Davis, E. J.; Dwivedi, P.; Tam, M.; Siems, W. F.; Hill, H. H. Anal. Chem. 2009, 81, 3270−3275. (12) Cumeras, R.; Figueras, E.; Davis, C. E.; Baumbach, J. I.; Gracia, I. Analyst 2015, 140, 1391−1410. (13) Michelmann, K.; Silveira, J. A.; Ridgeway, M. E.; Park, M. A. J. Am. Soc. Mass Spectrom. 2015, 26, 14−24. (14) Karasek, F. W. Anal. Chem. 1974, 46, 710A−716A. (15) Cohen, M. J.; Karasek, F. W. J. Chromatogr. Sci. 1970, 8, 330− 337. (16) Bowers, M. T.; Kemper, P. R.; von Helden, G.; van Koppen, P. A. M. Science 1993, 260, 1446−1451. (17) Sacristan, E.; Solis, A. A. IEEE Trans. Instrum. Meas. 1998, 47, 769−775. (18) Purves, R. W.; Guevremont, R. Anal. Chem. 1999, 71, 2346− 2357. (19) Kailemia, M. J.; Park, M.; Kaplan, D. A.; Venot, A.; Boons, G.-J.; Li, L.; Linhardt, R. J.; Amster, I. J. J. Am. Soc. Mass Spectrom. 2014, 25, 258−268. (20) Schneider, B. B.; Nazarov, E. G.; Londry, F.; Vouros, P.; Covey, T. R. Mass Spectrom. Rev. 2016, 35, 687−737. (21) Shvartsburg, A. A.; Smith, R. D. Anal. Chem. 2008, 80, 9689− 9699. (22) Baether, W.; Zimmermann, S.; Gunzer, F. IEEE Sens. J. 2012, 12, 1748−1754. (23) Gunzer, F.; Zimmermann, S.; Baether, W. Anal. Chem. 2010, 82, 3756−3763. (24) Revercomb, H. E.; Mason, E. A. Anal. Chem. 1975, 47, 970− 983. (25) Zhang, F.; Guo, S.; Zhang, M.; Zhang, Z.; Guo, Y. J. Mass Spectrom. 2015, 50, 906−913. (26) Bush, M. F.; Hall, Z.; Giles, K.; Hoyes, J.; Robinson, C. V.; Ruotolo, B. T. Anal. Chem. 2010, 82, 9557−9565. (27) Salbo, R.; Bush, M. F.; Naver, H.; Campuzano, I.; Robinson, C. V.; Pettersson, I.; Jørgensen, T. J. D.; Haselmann, K. F. Rapid Commun. Mass Spectrom. 2012, 26, 1181−1193. (28) Valentine, S. J.; Counterman, A. E.; Clemmer, D. E. J. Am. Soc. Mass Spectrom. 1997, 8, 954−961. (29) Shelimov, K. B.; Clemmer, D. E.; Hudgins, R. R.; Jarrold, M. F. J. Am. Chem. Soc. 1997, 119, 2240−2248. (30) Valentine, S. J.; Anderson, J. G.; Ellington, A. D.; Clemmer, D. E. J. Phys. Chem. B 1997, 101, 3891−3900. (31) Bush, M. F.; Campuzano, I. D. G.; Robinson, C. V. Anal. Chem. 2012, 84, 7124−7130. (32) Wyttenbach, T.; von Helden, G.; Bowers, M. T. J. Am. Chem. Soc. 1996, 118, 8355−8364. (33) Counterman, A. E.; Valentine, S. J.; Srebalus, C. A.; Henderson, S. C.; Hoaglund, C. S.; Clemmer, D. E. J. Am. Soc. Mass Spectrom. 1998, 9, 743−759.

■

■

Article

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. ORCID

R. Graham Cooks: 0000-0002-9581-9603 Present Address

‡ Z.B.: Indiana Biosciences Research Institute, 1345 W. 16th Street #300, Indianapolis, IN 46202, United States.

Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS This material is based upon work supported by the United States Department of Energy, Office of Science, Office of Basic Energy Sciences, Separations and Analysis Program, under Award DE-FG02-06ER15807. John M. Hollerbach is thanked for helpful discussions involving data analysis and for a critical review of the manuscript. Valentina Pirro and Christina Ferreira are thanked for their help in obtaining chemical standards. Stephen T. Ayrton and David Logsdon are thanked for helpful discussions involving data analysis. 5064


Article

Analytical Chemistry (34) Henderson, S. C.; Li, J.; Counterman, A. E.; Clemmer, D. E. J. Phys. Chem. B 1999, 103, 8780−8785. (35) Hoaglund, C. S.; Valentine, S. J.; Sporleder, C. R.; Reilly, J. P.; Clemmer, D. E. Anal. Chem. 1998, 70, 2236−2242. (36) Henderson, S. C.; Valentine, S. J.; Counterman, A. E.; Clemmer, D. E. Anal. Chem. 1999, 71, 291−301. (37) Hoaglund, C. S.; Liu, Y.; Ellington, A. D.; Pagel, M.; Clemmer, D. E. J. Am. Chem. Soc. 1997, 119, 9051−9052. (38) Fenn, L. S.; Kliman, M.; Mahsut, A.; Zhao, S. R.; McLean, J. A. Anal. Bioanal. Chem. 2009, 394, 235−244. (39) Campuzano, I.; Bush, M. F.; Robinson, C. V.; Beaumont, C.; Richardson, K.; Kim, H.; Kim, H. I. Anal. Chem. 2012, 84, 1026−1033. (40) Fernandez-Maestre, R.; Harden, C. S.; Ewing, R. G.; Crawford, C. L.; Hill, H. H. Analyst 2010, 135, 1433−1442. (41) Eiceman, G. A.; Nazarov, E. G.; Stone, J. A. Anal. Chim. Acta 2003, 493, 185−194. (42) Allen, S. J.; Bush, M. F. J. Am. Soc. Mass Spectrom. 2016, 27, 2054−2063. (43) Allen, S. J.; Giles, K.; Gilbert, T.; Bush, M. F. Analyst 2016, 141, 884−891. (44) Badu-Tawiah, A. K. Ion generation, ion collection and ionic reactions outside the mass spectrometer. Ph.D. Thesis, Purdue University, 2012. (45) Baird, Z.; Wei, P.; Cooks, R. G. Analyst 2015, 140, 696−700. (46) Baird, Z. Methods and instrumentation for the manipulation and characterization of electrosprayed ions under ambient conditions. Ph.D. thesis, Purdue University, 2015. (47) Amo, M.; Alvaro, A.; McColloch, R.; de la Mora, J. F. In 63rd ASMS Conference of Mass Spectrometry and Allied Topics; American Society for Mass Spectrometry: St. Louis, MO, 2015. (48) Kaur-Atwal, G.; O’Connor, G.; Aksenov, A. A.; Bocos-Bintintan, V.; Paul Thomas, C. L.; Creaser, C. S. Int. J. Ion Mobility Spectrom. 2009, 12, 1−14. (49) Forsythe, J. G.; Petrov, A. S.; Walker, C. A.; Allen, S. J.; Pellissier, J. S.; Bush, M. F.; Hud, N. V.; Fernandez, F. M. Analyst 2015, 140, 6853−6861. (50) Tufariello, J. A.; Grows, K.; Davis, E. J.; Harden, C. S.; Siems, W. W.; Hill, H. H. Int. J. Ion Mobility Spectrom. 2015, 18, 95−104. (51) Tolmachev, A. V.; Clowers, B. H.; Belov, M. E.; Smith, R. D. Anal. Chem. 2009, 81, 4778−4787. (52) Kanu, A. B.; Gribb, M. M.; Hill, H. H. Anal. Chem. 2008, 80, 6610−6619. (53) Mayer, T.; Borsdorf, H. Anal. Chem. 2014, 86, 5069−5076.

5065


Ion Separation in Air Using a Three-Dimensional ... - ACS Publications

Recommend Documents