Simple, Expendable, 3D-Printed Microfluidic ... - ACS Publications

Feb 23, 2017 - Simple, Expendable, 3D-Printed Microfluidic Systems for Sample. Preparation of Petroleum. Érica M. Kataoka,. †. Rui C. Murer,. †. ...
2 downloads 0 Views 3MB Size
Subscriber access provided by UB + Fachbibliothek Chemie | (FU-Bibliothekssystem)

Article

Simple, Expendable, 3D-Printed Microfluidic Systems for Sample Preparation of Petroleum Erica Megumi Kataoka, Rui Cesar Murer, Jandyson Machado Santos, Rogerio Mesquita de Carvalho, Marcos Nogueira Eberlin, Fabio Augusto, Ronei Jesus Poppi, Angelo Luiz Gobbi, and Leandro Wang Hantao Anal. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.analchem.6b04413 • Publication Date (Web): 23 Feb 2017 Downloaded from http://pubs.acs.org on February 24, 2017

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

Analytical Chemistry is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 9

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Analytical Chemistry

Simple, Expendable, 3D-Printed Microfluidic Systems for Sample Preparation of Petroleum Érica M. Kataoka,a Rui C. Murer,a Jandyson M. Santos,b Rogério M. Carvalho,c Marcos N. Eberlin,b Fabio Augusto,b Ronei J. Poppi,b Angelo L. Gobbi,a and Leandro W. Hantaoa,b,* a

Laboratório Nacional de Nanotecnologia, Centro Nacional de Pesquisa em Energia e Materiais, Campinas, SP 13083-100, Brazil b Instituto de Química, Universidade Estadual de Campinas, Campinas, SP 13083-970, Brazil c Centro de Pesquisas e Desenvolvimento Américo Miguez de Mello, Petrobras, Rio de Janeiro, RJ, Brazil ABSTRACT: In this study, we introduce a simple protocol to manufacture disposable, 3D-printed microfluidic systems for sample preparation of petroleum. Such platform is produced with a consumer-grade 3D-printer, using fused deposition modeling. Successful incorporation of solid-phase extraction (SPE) to microchip was ensured by facile 3D element integration using proposed approach. This 3D-printed μSPE device was applied to challenging matrices in oil & gas industry, such as crude oil and oil-brine emulsions. Case studies investigated important limitations of non-silicon and non-glass microchips, namely, resistance to nonpolar solvents and conservation of sample integrity. Microfluidic features remained fully functional even after prolonged exposure to nonpolar solvents (20 min). Also, 3D-printed μSPE devices enabled fast emulsion breaking and solvent deasphalting of petroleum, yielding high recovery values (98%) without compromising maltene integrity. Such finding was ascertained by high-resolution molecular analyses using comprehensive two-dimensional gas chromatography and gas chromatography/mass spectrometry by monitoring important biomarker classes, like C10 demethylated terpanes, ααα-steranes, and monoaromatic steroids. 3D-printed chips enabled faster and reliable preparation of maltenes by exhibiting a 10-fold reduction in sample processing time, compared to the reference method. Furthermore, polar (oxygen-, nitrogen-, and sulfur-containing) analytes found in low-concentrations were analyzed by Fourier transform ion cyclotron resonance mass spectrometry. Analysis results demonstrated that accurate characterization may be accomplished for most classes of polar compounds, except for asphaltenes, which exhibited lower recoveries (82%) due to irreversible adsorption to sorbent phase. Therefore, 3D-printing is a compelling alternative to existing microfabrication solutions, as robust devices were easy to prepare and operate.

INTRODUCTION Analytical methods used for molecular characterization of petroleum, i.e. petroleomics,1 rely on high-resolution instrumental techniques, like comprehensive twodimensional gas chromatography (GC×GC)2 and Fourier transform ion cyclotron resonance mass spectrometry (FTICR MS).3,4 In this context, converting the sample to a suitable format that is compatible with the analytical instrumentation involves multistep procedures. Excessive sample handling may, however, change sample composition and decrease method throughput.5 Occurrence of nonprocessed crude oil as stable emulsions, for instance, hinders use of simple extraction protocols and direct analysis.6 Presence of fine solids, salts, and chemical additives further aggravate such limitations.7 Current best achievement in standard technologies relies on the use of tailored centrifuges for petroleum, water, and sediment separation.6 Effective concurrent routes employ solvent extraction and filtration. These protocols are, nevertheless, frequently incompatible with miniaturization and automation. Therefore, such methods may be inapplicable to process trace-amounts of crude oil. Indeed, selection of an appropriate sample preparation method is critical for

obtaining reliable downstream measurements and consequently warrants careful consideration. Microfluidics is a powerful platform to perform sophisticate chemical analyses by establishing well-controlled microenvironments for manipulating reduced quantities of fluids. Selection of materials for device manufacturing plays a dominant role in this field, once it determines specific microfabrication strategies and native properties of device.8 Traditional glass and silicon chips are ideal for applications involving organic solvents, but require laborious and expensive fabrication, as trained personnel and specialized facilities for device manufacturing (like clean rooms and machining centers) are needed.8,9 Furthermore, integration of functional 3D elements to microfluidic devices is challenging.9,10 Alternatively, elastomer-based platforms are rapid and inexpensive to manufacture, but its application to petroleum processing is eclipsed by solvent swelling8,11 and analyte sorption by substrate, e.g. poly(dimethylsiloxane) (PDMS).12 For example, Shirure et al. reported significant loss of hydrophobic analytes in PDMS-based chips, due to absorption.13 Perfluoropolymers seems to mitigate conventional elastomer-related downsides by exhibiting inertness and antifouling properties.14–16 These findings are,

ACS Paragon Plus Environment

Analytical Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

however, preliminary in nature, as the investigation of the sorptive properties of nonpolar analytes, such as crude oil components, by perfluoropolymers is incipient. In addition, proposed monomers are not readily available for purchase. Application of microfluidic systems to oil & gas industry (O&G) are limited to silicon and glass-based microfluidic devices. For example, controlled microenvironments were used for the investigation of transport phenomena in O&G.17 Moreover, microfluidic devices probed chemical composition,17,18 properties of water-oil interface,19 phase behavior, pressure, volume, and temperature of reservoir fluids.20,21 Prospection of microfluidic platforms for sample preparation of petrochemicals is, however, still inexistent. In this context, plain uptake of microfluidic devices for sample preparation in petroleomics is defied by two constraints. First is the limited prospection of materials that can withstand extended exposure to nonpolar solvents, like n-heptane and toluene. Next, arduous integration of functional 3D elements to microfluidic platforms restrains implementation of successful macroscale methods, including solid-phase extraction (SPE).6,22 In an effort to overcome current restrictions of fabrication approaches for microfluidic devices, we describe a simple protocol to manufacture expendable 3D-printed microfluidic devices for SPE (µSPE). We employed a miniaturized approach of a benchtop SPE method for petroleum demulsification, previously reported by our group.6 A hallmark of the method was the ability to circumvent the requirement of sophisticated instruments, enabling prompt incorporation to microfluidics. Herein, 3Dprinted devices for crude oil processing were produced with an inexpensive and readily accessible 3D printer and a commercially available substrate. Such microchips are easy to fabricate and assemble, even by untrained personnel in resource-limited laboratories. We further demonstrate the potential of 3D-printing as an alternative method for the fabrication of microfluidic systems. In this context, performance of µSPE devices was investigated using two demanding sample preparation applications for petroleomics, namely, demulsification of trace-amounts of synthetic oil-brine emulsion and solvent deasphalting. For the first application, samples were characterized by comprehensive two-dimensional gas chromatography with flame ionization detection (GC×GCFID) for fingerprinting of maltenes. The second application comprised one of the most important sample processing methods used in all GC-based petroleomics studies, namely, solvent deasphalting. Since prospection and characterization of biomarkers are fundamental to organic geochemical investigations, highly-sensitive selected ion monitoring (SIM) by gas chromatography/mass spectrometry (GC/MS) was employed for screening analysis of the most common classes of biological markers. Additionally, in untargeted analysis, conservation of analyte integrity during sample preparation is fundamental. Hence, unwanted analyte sorption by substrate or coextraction of interfering compounds from device must be avoided. Here, careful investigation of such parameter was performed using additional instruments with varying selectivity and sensitivity, including FT-ICR MS and nuclear magnetic resonance (NMR). These case studies showcased that 3D-printing may be a powerful technique for the fabrication of disposable micro analysis systems for O&G, once potential limitations of non-silicon and non-glass microchips have been overcome. Aspects as solvent swelling and conservation of sample integrity were effectively solved by using proposed route. Moreover, to the best of our knowledge, deployed approach

Page 2 of 9

pioneered the use of non-silicon and non-glass microfluidic systems to process petrochemical matrices.

MATERIALS AND METHODS 3D-Printing Computer-aided design (CAD) of microfluidic device was created with Inventor software (Autodesk – San Rafael, CA, USA). Slic3r (www.slic3r.org) was used to translate CAD file to computer numerical control (CNC) code. Printing instructions were uploaded to Graber i3 3D printer (GTMax – Americana, SP, Brazil) by Repetier-Host software (Hot-World GmbH & Co. KG – Willich, Germany). Micro devices were produced via 3D-printing using fused deposition modeling (FDM). Devices were printed in poly(lactic acid) (PLA) (UP3D – Louveira, SP, Brazil). Full printing settings can be found in Section S1 (Supporting Information).

Microfluidic Devices A fully fabricated µSPE device is shown in Figure 1A. Detailed description of procedures used to assemble the microchip can be found in Section S2 (Supporting Information). Leak tests were performed to measure the maximum pressure that the device could withstand. Outlet of µSPE device was blocked, while aqueous dye solution was pumped into the inlet at 1 mL min-1. The pressure required to drive the water was monitored throughout the experiment. Maximum operating pressure was determined by visual inspection of leaks. SPE experiments were performed using a syringe pump (New Era Pump Systems, Inc. – Farmingdale, NY, USA), as shown in Figure S3 (Supporting Information). Solvents were injected to the microfluidic system at a constant flow rate of 1 mL min-1. Celite 545 (Sigma-Aldrich – Missouri, USA), a flux-calcined diatomaceous earth, was used as stationary phase for emulsion breaking and solvent deasphalting. Samples were collected using glass test tubes (Figure S3B). Sample preparation methods are described in Sections S3 to S8 (Supporting Information).

GC×GC and GC/MS Analyses Crude oil-brine emulsions were prepared using samples and methods provided by Petrobras (Section S4, Supporting Information).6 Oil phase attained by emulsion breaking (Section S5, Supporting Information) was characterized by GC×GC-FID. This system consisted of a TRACE 1310 GC-FID gas chromatograph (Thermo Fisher Scientific – Walthman, MA, USA) equipped with a split/splitless injector and a AS 1300 105-position autosampler. GC×GC separations were performed using a high-speed Deans switch modulator based on SilFlow platform (SGE Analytical– Victoria, Australia).6 Modulation period was set to 2.0 s. Column set comprised a 30 m × 0.25 mm MEGA-5 HT (0.15 µm film thickness) primary column and a 2.0 m × 0.25 mm MEGAWAX-HT (0.15 µm film thickness) secondary column (MEGA s.n.c – Megano, MI, Italy). Samples were analyzed in quadruplicate. Maltenes attained by solvent deasphalting of crude oil (Section S6, Supporting Information) were characterized by GC/MS. Geochemical biomarkers were screened on a Shimadzu TQ8030 instrument (Shimadzu Corporation – Kyoto, Japan). Separations were executed using a 30 m × 0.25 mm Rtxi-5ms (0.25 µm film thickness) column (Restek – Bellefonte, PA, USA). Mass analyzer was operated in Q3 SIM mode. Samples were analyzed in quadruplicate. Full

ACS Paragon Plus Environment

Page 3 of 9

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Analytical Chemistry

GC×GC-FID and GC/MS conditions are presented in Sections S5 and S6 (Supporting Information).

NMR and FT-ICR MS Analyses 1H

NMR experiments were performed on Avance III 400 MHz instrument (Bruker BioSpin – Billerica, MA, USA). Samples were prepared by solvent removal of crude oil extracts and dissolution in 700 µL of deuterated chloroform (Tedia – Rio de Janeiro, RJ, Brazil). FT-ICR MS analysis was performed by dissolving the samples in 1:1 (v/v) toluene:methanol, obtaining a final solution of 0.5 mg mL-1. FT-ICR MS experiments were performed using a 7.2T LTQ FT Ultra mass spectrometer (Thermo Fisher Scientific) equipped with a direct infusion electrospray ionization (ESI) source, operating in positive [ESI(+)] or negative [ESI(–)] modes. For ESI(+), formic acid was added to facilitate protonation of the basic compounds. For ESI(–), aqueous ammonium hydroxide was added to facilitate deprotonation of acidic compounds. Data acquisition was performed along the m/z range of 100 – 1000 by LTQ FT Ultra 2.0 software. The mass resolving power was m/∆m50% = 400 000 at m/z 400. NMR and FT-ICR MS conditions are detailed in Sections S7 and S8 (Supporting Information).

cohesion between PLA polymeric chains, due to existence of strong intra- and intermolecular hydrogen-bonding. Accordingly, ABS and PET are more susceptible to swelling by nonpolar solvents, because of weaker intermolecular forces (dispersive forces) between its polymeric chains, compared to PLA. Therefore, PLA was chosen as the substrate, once this thermoplastic was able to withstand organic solvents and displayed the lowest post-printing shrinking. Also, PLA-derived devices may be considered earth-friendly, because this material is biodegradable and it is produced from renewable sources, like corn.25

RESULTS AND DISCUSSION 3D-Printed Microfluidic Devices The cost, time, and restrictions on creative flexibility associated with current fabrication methods, like photolithography, micromachining, and injection molding,23 present significant challenges to the development of microfluidic devices.24 3D-printing enables rapid prototyping of microchips, circumventing the need to produce a master mold prior to microfabrication process.24 In this report, we carefully evaluated the most common, consumer-grade 3D printer that uses FDM to produce microfluidic platforms (Section S1, Supporting Information). Printing parameters such as extruder Z-offset, nozzle deposition speed, and extrusion temperature (Tables S1 – S6, Supporting Information) were meticulously balanced to favor structural strength of microfluidic devices and enhance fusion of adjacent layers. Consequently, devices were capable of operating at pressures in excess of 5,000 KPa. Higher pressures were not evaluated, because average flow rates of 1.0 mL min-1 of n-heptane and toluene were easily attained under 800 KPa. Flow rates above 1.0 mL min-1 jeopardized process of emulsion breaking and solvent deasphalting, as these processes are most effective using conditions that favor mass transfer.6 Numerous thermoplastic substrates were evaluated to impart Spartan-like characteristics to microfluidic device, namely, the ability to remain fully functional after extended exposure to organic solvents, including n-heptane and toluene (up to 20 min). Nylon and poly(propylene) displayed the highest post-printing shrinking, causing 3D elements to warp and delaminate, in comparison to ABS, PET, and PLA. Compared to ABS and PET, polar PLA exhibited less solvent swelling by nonpolar and aromatic solvents, as n-heptane and toluene. Most likely, this feature is attributed to higher

Figure 1. Characterization and surface depth profiles of 3D-printed µSPE system for sample preparation of crude oil. Microfluidic device is shown in (A), showcasing base component (top) and totally assembled device (bottom). This device comprises a cylindrical sample compartment and a rectangular container for stationary phase. Dimensions of microchip are 25 mm × 50 mm of base and 10 mm of height. Photos by digital stereoscope of microchannels, highlighting half-printed microchannel (B) and cross section view of approximately circular channel (C). Dyes were used to enhance visualization. Profilometry of half-printed base component of µSPE output (D) and microfluidic channel (E). Simple fabrication using 3D-printing and PLA as thermoplastic substrate enabled production of microchips (Figure 1A) less expensive (ca. USD 0.50 per unit) than devices produced with glass and silicon – which are commonly used for fabrication of such small devices. Overall dimensions of the microfluidic device are 25 mm × 50 mm of base and 10 mm of height, comprising a cylindrical sample compartment, a rectangular stationary phase container, and two stainless steel connection ports (Figure S2, Supporting Information). Such 3D elements are interlocked by circular channels. The first millifluidic channel connected the sample compartment to sorbent recipient. This channel reduced pneumatic resistance to flow of highly viscous samples, like oil-brine emulsions. Next, microfluidic channel (Figure 1B) was positioned between stationary phase compartment and μSPE outlet.

ACS Paragon Plus Environment

Analytical Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 4 of 9

Figure 2. Comparison of methods for emulsion breaking of petroleum. GC×GC-FID chromatograms of samples were attained using centrifugation (reference method) (A) and 3D-printed µSPE device (B). Additionally, the surfaces of microfluidic elements were profiled. Digital stereoscope images determined that proposed approach enabled printing of microchannels approximately circular in cross section (Figure 1C), with an average diameter of 475 ± 20 μm (n = 10). Profilometry data, as shown in Figure 1E, also highlighted increased average surface roughness (Ra of 112 μm), compared to glass-based microchannels attained by wet etching (Ra ~ 0.48 μm).26 Conversely, emulsion breaking performance of µSPE device remained unaffected by such value of Ra, compared to macroscale approach.6 Element geometry and positioning were critical to avoid dead-volume in μSPE device. Channels were positioned in an off-axis manner, as shown in Figure S2 (Supporting Information) to favor crude oil mixing and dissolution by organic solvents. Absence of dead-volume zones was ascertained by inspection of sample compartment after total elution of crude oil with mobile phase. No trace of petroleum was detected. Furthermore, analysis of stationary phase confirmed complete elution of maltenes from µSPE device (Figure S5, Supporting Information).27 Such observation was supported by recovery experiments (98 ± 3%), previously reported by our group,6 and confirmed by highresolution molecular analysis, as detailed below.

Evaluation of Microfluidic Devices for Sample Preparation of Petroleum Petrochemical samples are among the most complex samples known to analytical chemists. To illustrate the bewildering complexity of such matrix, it is estimated that 75 isomers may be found for a carbon number (Cn) of 10, while 4347 distinct compounds are available for a Cn of 20. Thus, nearly one million compounds may integrate a middle distillate sample.28 So, the utility of our 3D-printed microplatform was challenged by processing common matrices found in O&G, namely, crude oil and oil-brine emulsion. Group-type analysis of petrochemical derivatives is often performed, instead of targeted analysis,28 due to the exceedingly complex nature of sample matrix. Using such untargeted approach, applications of analytical techniques to unveil oil composition, petroleum quality, and geochemical parameters have been reported.1,29,30 In screening analysis, conservation of sample integrity is a prime requisite of any

sample preparation platform. Indeed, unwanted analyte losses, due to sorption by device’s substrate, and coextraction of interfering compounds from substrate must be kept to a minimum. In this study, careful investigation of such parameter was performed, using instruments with varying selectivity and sensitivity.

Case Study 1: Microfluidic Device Applied to Geochemical Analysis – Emulsion Breaking SPE is a well-established alternative to conventional solvent extractions, as it enables sample preparation with reduced sample manipulation and solvent economy. More specifically, selective water retention may be attained using diatomaceous earth as sorbent material. Considering oilbrine emulsions, sample dewatering was readily accomplished, as multiphasic equilibration allowed aqueous phase to adsorb to the surface of microamorphous silica.6 Next, oil phase was easily desorbed using an eluotropic series. We have previously reported that quantitative retention of water, hydrophilic salts, and sediments was achieved using proposed SPE method.6 However, such approach using 3 mL glass syringes was inadequate to process trace-amounts of crude oil and it required laborious cleaning of SPE components to avoid sample contamination. Thus, we evaluated the feasibility of 3D-printing to produce reliable and expendable µSPE systems for emulsion breaking of petroleum. Comparative analysis was performed to evaluate the applicability of proposed platform. Samples of oil phase were obtained by phase separation using μSPE device and centrifugation – adopted as the reference procedure, in accordance to Petrobras internal protocol. GC×GC is currently one of the most effective techniques for separation of volatile and semi-volatile organic compounds.31 Considering that over 90 wt% of crude oil comprises hydrocarbons, GC×GC analysis is ideal for fingerprinting maltenes, since enhanced chromatographic resolution and well-ordered 2D structures are readily attained. For instance, in Figure 2, analytes are separated primarily by their vapor pressure in the first dimension. In the second dimension, the least retained and most intense compounds are the paraffins – normal and branched acyclic hydrocarbons. Next, naphthenic hydrocarbons (cyclic alkanes) and olefins (unsaturated hydrocarbons) are more retained in the 2D, but are still adjacent to the class of paraffins in the chromatogram. Together, the paraffins, naphthenes, and

ACS Paragon Plus Environment

Page 5 of 9

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Analytical Chemistry

olefins comprised the bulk of the unresolved complex mixture. The more retained solutes in the 2D space are the aromatic hydrocarbons. Comparison of the chromatographic profiles (Figure 2) and peak integration results (Table S8, Supporting Information) indicated that both sample preparation approaches are equivalent, as hydrocarbon compositions were not significantly different from each other (n = 4, p = 0.05). Such finding demonstrated that maltene integrity was conserved during sample preparation. More specifically, nonpolar solvents did not coextract detectable amounts of volatile additives or traces of PLA oligomers from microchip. Also, unwanted sorption of analytes by PLA substrate was insignificant. Hence, µSPE system enabled reliable emulsion breaking of trace-amounts of petroleum. Overall, the use of µSPE devices allowed a 10-fold reduction in sample preparation time, compared to centrifugation. Furthermore, proposed approach enabled simultaneous water removal, sample desalting, and oil fractionation,6 which would otherwise require sophisticated techniques, including ultrasonic radiation treatment,32,33 electrostatic-based 34,35 and microwave assisted systems.36 Also, our sample preparation method bypassed the use of chemical additives to enhance phase separation during emulsion breaking. These additives may result in deleterious and unforeseeable matrix effects, such as ion suppression in MS.37

Case Study 2: Microfluidic Device Applied to Geochemical Analysis – Solvent Deasphalting The analyses of biomarkers in crude oil can provide important information related to source, diagenesis, catagenesis, biodegradation, and lithology.38 Biomarkers are generally minor components of crude oil, which are usually found in the nonpolar fraction, requiring highly selective and sensitive methods of analysis, such as GC/MS. In organic geochemical analysis, sample contamination is a critical issue, as unreliable data may lead to incorrect characterization of crude oil. In this context, contamination downsides may be amplified when processing trace-amounts of petroleum. So, the production of expendable sample preparation microplatforms is of great relevance. Here, we evaluated the potential of 3D-printing for the fabrication of such devices. One of the most important sample processing methods consists of solvent deasphalting prior to GC/MS analysis. For instance, maltenes were purified by diluting crude oil into nheptane, followed by asphaltene precipitation and isolation using centrifugation (Figure S6, Supporting Information). Analogously to saturate, aromatic, resin, and asphaltene (SARA) oil analysis, SPE allowed removal of asphaltenes using diatomaceous earth as stationary phase, once this sorbent is also widely used in industry as filtration media.6 Samples of oil phase were obtained using microfluidic method and centrifugation, which was adopted as the reference protocol. Comparative analysis (Figure 3) was attained by evaluating extracted ion chromatograms of important biomarker-related m/z channels, including C10 demethylated terpanes (m/z 177), tri-, tetra-, and pentacyclic terpanes (m/z 191), ααα-steranes (m/z 217), triaromatic steroids (m/z 231), C-ring monoaromatic steroids (m/z 253), and diasteranes, tetracyclic polyprenoid (m/z 259) (Figures S7 – S12, Supporting Information). Noteworthy, although C-ring monoaromatic steroids are found in trace concentrations, as shown in Figure S10, their GC/MS profiles remained virtually identical. Integration results (Table

S10, Supporting Information) ascertained that both methods are equivalent, as biomarker compositions were not significantly different from each other (n = 4, p = 0.05). Therefore, these results evidenced lack of contamination during sample preparation, and negligible biomarker sorption to PLA substrate. These findings confirmed that maltene integrity was preserved in all evaluated m/z channels.

Figure 3. Comparison of sample preparation techniques for solvent deasphalting. GC/MS chromatograms of samples were obtained by reference protocol (black) and 3D-printed µSPE device (red). Important biomarker-related channels were monitored, including m/z 177 (C10 demethylated terpanes) (A), m/z 191 (tri-, tetra-, and pentacyclic terpanes) (B), and m/z 217 (ααα-steranes) (C). Further sample characterization by GC/MS is shown in Supporting Information. Baseline of reference analysis was offset to facilitate chromatogram visualization.

Case Study 3: 3D-Printed Microfluidics Meets NMR Fingerprinting and High-Resolution Petroleomics Previous investigations using highly sensitive GC-based techniques mitigated hypothesis of alteration of hydrocarbon composition during sample preparation. However, analysis of traces of polar components of crude oil (~ 10 wt%) was also necessary to confirm broad applicability of 3D-printed μSPE for petroleomics and related applications in O&G. Therefore, additional techniques were used to investigate conservation of sample integrity, including 1H NMR and FT-ICR MS. NMR commonly probes bulk properties of crude oil. For example, saturate, aromatic, and polar fractions of crude oil were quantitated using NMR and chemometric techniques, like support vector regression and genetic algorithm.39 1H NMR spectra (Figure 4) were measured to determine information on aliphatic and aromatic hydrocarbons in petroleum. Oil composition was unaffected by sample preparation method through inspection of the most intense signals, including diaromatic CH (δ = 7.62 ppm), monoaromatic CH (δ = 7.22 ppm), olefinic CH2 (δ = 4.23 ppm), aliphatic CH2 (δ = 1.28 ppm), and γ-CH3 (δ = 0.92 ppm). 1H NMR full spectra are presented in Figure S13 (Supporting Information).

ACS Paragon Plus Environment

Analytical Chemistry

(A)

405.22249

100

(A)

299.20177 390.19903

443.29562

339.23304

468.36362 496.39490 536.42619

7.36

7.08

(B)

Relative abundance / %

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 6 of 9

566.47314 608.52006 622.53573 650.56700 678.59829

257.12105 243.10541

200

200 250

732.64536

300

350

400

450

500

550

600

650

700

750

788.70780

800

840.73925

850

405.22245

100

900

956.89553

1 000

950

(B)

440.33228 482.37922

325.18435

896.80167

496.39489 390.19898 524.42617 294.12888 255.23301

538.44187

339.23300

566.47315 578.47313 608.52010 636.55152 650.56718 692.54134 722.58821 778.65108

7.80

7.43

δ 1H/ ppm

Figure 4. Petroleum characterization by 1H NMR, illustrating important regions of spectra, like monoaromatic CH (A) and diaromatic CH (B) signals. Such bands are related to aromatic hydrocarbons, which are likely to irreversibly adsorb to substrates during sample processing. NMR fingerprinting was obtained using reference method (top) and 3D-printed µSPE device (bottom). Signal at 7.26 ppm is related to deuterated solvent (CD3Cl). FT-ICR MS offers the highest available mass resolution, mass resolving power, and mass accuracy for crude oil characterization,1 allowing direct separation, and identification of thousands of components in such complex organic mixture. Moreover, ESI(±) FT-ICR MS has been successfully used in petroleomics to characterize crude oils and its derivatives.7,40 Because many heteroatom-containing compounds of petroleum are highly polar, ESI is specific and especially efficient to generate their gas-phase ions.1 Thus, oxygen-, nitrogen-, and sulfur-containing compounds are accurately and comprehensively probed using ESI(±) FT-ICR MS. Mass spectra of oil samples attained by ESI(–) FT-ICR MS are shown in Figure 5. Comparative analysis revealed that samples obtained using both sample preparation methods exhibited the same broadband mass spectral patterns.4 In the same mass range of 290 – 390 Da for negative ESI, µSPE processing reduced interference of peaks at m/z 299.20177 and 390.19903 on the ionization of petroleum components, as shown in Figure S14 (Supporting Information). Such fact is confirmed by the peak values of ion at m/z 496.39489, which increased from 44% to 82% relative abundance, a 2-fold improvement. Hence, proposed microfluidic method allowed removal of important interfering compounds that suppressed petroleum ionization by ESI(–). Analogously, mass spectra of oil samples attained by ESI(+) were also appraised. Mass spectra of samples prepared by reference and µSPE method are shown in Figures S15 and S16 (Supporting Information), respectively. Comparative analysis indicated that trace-amounts of heavier oligomers of PLA were coextracted by toluene (Figure S15B). Such ions were detected within the m/z range of 703 to 991, spaced by 72 m/z units. As a consequence, average ion intensities of crude oil peaks were suppressed, resulting in a 4-fold reduction of relative intensities.

0 200

200250

300

350

400

450

500

550

600

m/z

650

700

750

800

832.69766

850

915.59969 932.82321 900

1 0001000

950

Figure 5. ESI(–) FT-ICR MS of crude oil extracts obtained using reference method (A) and using 3D-printed µSPE-based method (B). Mass resolving power of 400,000 at m/z 400. After resolving and converting the thousands of peaks in the FT-ICR mass spectrum, assignment of elemental composition from accurate mass measurements (m/∆m50% of 400,000) was performed. Successful identification was achieved with mass errors below 1 ppm. Heteroatom class distribution is shown in Figure 6. Comparative analysis demonstrated that differences in the polar fraction composition of petroleum probed by negative-ion ESI, were insignificant (n = 2, p = 0.05), considering both sample preparation methods. Similarly, differences in the composition of petroleum acquired by positive-ion ESI, using both sample processing approaches were not statistically significant (n = 2, p = 0.05). Figures S15 and S16 (Supporting Information) suggested, however, that the use of 3D-printed microfluidic devices made with PLA may have a negative effect on the molecular information attained only by ESI(+) FT-ICR MS analysis. The N-class distribution is highlighted in Figure 7, since molecular formulas with one nitrogen were predominant in both positive and negative ESI modes, as shown in Figure 6. The distribution of compounds based on elemental composition of a complex sample is better represented by the iso-abundance plot. Figure 7 shows the iso-abundance graph of N-class members for samples obtained using both sample preparation methods. In this plot, double bond equivalent values (DBE) are plotted against Cn with sizecoded relative abundance.4 In general, DBE values may represent the aromaticity and Cn the boiling point of the analyte. Peaks detected with Cn and DBE values below 60 and 25, respectively, indicated that the integrity of such polar fraction was preserved during sample preparation. However, composition differences are perceived when inspecting peaks with Cn and DBE values above 60 and 25, respectively. Despite this, such fact was not statistically significant, because the intra-class variation exceeded the inter-class variation. To confirm this finding, we investigated the adsorption of asphaltenes to the components of the µSPE device, including Teflon, PLA, and Celite (Figure S4, Supporting Information). This preliminary data demonstrated that this class of compounds irreversibly adsorbed to Celite, but not to Teflon and PLA (n = 3, p = 0.05). This observation is in agreement with previous reports, wherein highly aromatic and heavy molecules, e.g.

ACS Paragon Plus Environment

Page 7 of 9

(A)

Relative abundance / %

asphaltenes, strongly adsorb to metal oxides, including silica.6,41 45 40 35 30 25 20 15 10 5 0

(B)

Relartive abundance / %

N

(C)

NO

NO2 2

12 10 8 6 4 2 0

O

Relative abundance / %

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Analytical Chemistry

O22

O33

S

100 90 80 70 60 50 40 30 20 10 0

N

NS

resulting in a 4-fold reduction on peak intensities. Although such drawback did not exhibit significant impacts on the elemental composition of oil samples determined by ESI(±) FT-ICR MS, careful consideration is advised when considering the use of PLA-based microfluidic devices and positive-ion ESI mode. Successful proof-of-concept studies illustrated the potential of 3D-printed microsystems for sample preparation of petroleum. Tailored μSPE devices enabled reliable processing of trace-amounts of petroleum, yielding a 10-fold reduction in sample preparation time, compared to reference method. Besides increasing sample throughput, microfluidic processing using such devices bypassed the use of sophisticated instruments for sample processing. Developed 3D-printing method is a compelling alternative to existing fabrication methods of microfluidic devices. While this article does not report the hyphenation of µSPE to other instrumental techniques, besides GC/MS, GC×GC-FID and FT-ICR MS, it indicates that powerful, multiplatform investigations of petroleum may be accomplished. For instance, on-line coupling of µSPE to portable infrared spectrometers is an interesting solution for fast and on-site assessment of the bulk properties of crude oil.

O33

Classes

Figure 6. Comparative analysis using heteroatom class distribution obtained from ESI(±) FT-ICR MS analyses of a Brazilian presalt crude oil. Class distribution for ESI(–) is presented in (A) and (B). ESI(+) data is shown in (C). Samples were prepared using reference approach (blue) and µSPE-based sample preparation (orange).

CONCLUSIONS Here, we described the use of a readily accessible, consumer-grade 3D printer to fabricate simple, expendable, microfluidic platforms for petroleum processing. Challenging case studies were used to investigate the main limitations of 3D-printed microchips to process petroleum samples, namely, resistance to nonpolar solvents and conservation of sample integrity. 3D-printed microfluidic elements on PLA substrate remained fully functional during sample preparation, even after prolonged exposure (at least 20 min) to solvents, including n-heptane, toluene, and methanol. Negligible sorption of analytes to microfluidic components was determined. However, minor adsorption ( 98%). Moreover, 1H NMR further confirmed that the bulk features of crude oil were unaffected by sample preparation method. ESI(±) FT-ICR MS was used to evaluate the composition of polar analytes including oxygen-, nitrogen-, and sulfur-containing compounds. The most accurate characterization of crude oil was obtained by combining µSPE-based method and ESI(–), as a 2-fold increase on signal intensity of petroleum-related peaks was observed. Conversely, minor coextraction of heavy oligomers of PLA was only detected by ESI(+) FT-ICR MS,

Figure 7. Iso-abundance plots of DBE versus carbon number for molecular formulas of the N-class of Brazilian presalt crude oil. ESI(–) data is shown in (A) and (B). ESI(+) data is illustrated in (C) and (D). Iso-abundance plots obtained using reference method are exhibited in (A) and (C). Data of samples prepared using µSPE device are illustrated in (B) and (D).

ASSOCIATED CONTENT Supporting Information The Supporting Information is available free of charge on the ACS Publications website. Full printing settings are shown in Section S1 (Additive Manufacturing Using Fused Deposition Modeling). Microfuidic device assembly is described in Section S2 (Assembly of Microchip for Solid Phase Extraction). Additional methods and characterizations related to proof-of-concept experiments are also detailed (S3 – Asphaltenes Adsorption; S4 – Emulsification; S5 – Oil-Brine Emulsion Breaking Using µSPE Device; S6 – Solvent Deasphalting Using Microfluidic System; S7 – Nuclear Magnetic Resonance Fingerprinting; S8 – Petroleum Analysis Using High Resolution Mass Spectrometry; S9 – References).

ACS Paragon Plus Environment

Analytical Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

AUTHOR INFORMATION Corresponding Author: *Leandro Wang Hantao Centro de Pesquisa em Energia e Materiais 10 000 Giuseppe Máximo Scolfaro, Campinas, SP 13083-100 E-mail: [email protected]. Phone: +55 019 35123566.

Notes All authors have given approval to the final version of the manuscript and declare no competing financial interest.

ACKNOWLEDGMENT Funding for this research was provided by the São Paulo Research Foundation (FAPESP - AP.R 2015/05059-9 and DR 2013/191614) and Petrobras (2014/00228-4). Stefano Galli (MEGA snc) is thanked for providing GC columns. The authors also acknowledge Gabriel Gaal and Prof. Antônio Riul Júnior (IFGW) for helpful discussions.

REFERENCES (1) Marshall, A. G.; Rodgers, R. P. Proc. Natl. Acad. Sci. 2008, 105, 18090–18095. (2) Byer, J. D.; Siek, K.; Jobst, K. Anal. Chem. 2016, 88, 6101–6104.

(3) Pudenzi, M. A.; Eberlin, M. N. Energy & Fuels 2016, 30, 7125– 7133. (4) Hsu, C. S.; Hendrickson, C. L.; Rodgers, R. P.; McKenna, A. M.; Marshall, A. G. J. Mass Spectrom. 2011, 46, 337–343. (5) Pawliszyn, J. Anal. Chem. 2003, 75, 2543–2558. (6) Higa, K. M.; Guilhen, A.; Vieira, L. C. S.; Carvalho, R. M.; Poppi, R. J.; Baptistão, M.; Gobbi, A. L.; Lima, R. S.; Hantao, L. W. Energy & Fuels 2016, 30, 4667–4675. (7) Santos, J. M.; Galaverna, R. de S.; Pudenzi, M. A.; Schmidt, E. M.; Sanders, N. L.; Kurulugama, R. T.; Mordehai, A.; Stafford, G. C.; Wisniewski, A.; Eberlin, M. N. Anal. Methods 2015, 7, 4450– 4463. (8) Ren, K.; Zhou, J.; Wu, H. Acc. Chem. Res. 2013, 46, 2396– 2406. (9) Kovarik, M. L.; Ornoff, D. M.; Melvin, A. T.; Dobes, N. C.; Wang, Y.; Dickinson, A. J.; Gach, P. C.; Shah, P. K.; Allbritton, N. L. Anal. Chem. 2013, 85, 451–472. (10) Patabadige, D. E. W.; Jia, S.; Sibbitts, J.; Sadeghi, J.; Sellens, K.; Culbertson, C. T. Anal. Chem. 2016, 88, 320–338. (11) Bhattacharjee, N.; Urrios, A.; Kang, S.; Folch, A. Lab Chip 2016, 16, 1720–1742. (12) Seethapathy, S.; Górecki, T. Anal. Chim. Acta 2012, 750, 48– 62. (13) Shirure, V. S.; George, S. C. Lab Chip 2017, DOI: 10.1039/C6LC01401A. (14) Rolland, J. P.; Van Dam, R. M.; Schorzman, D. A.; Quake, S. R.; DeSimone, J. M. J. Am. Chem. Soc. 2004, 126, 2322–2323. (15) Vitale, A.; Quaglio, M.; Marasso, S. L.; Chiodoni, A.; Cocuzza, M.; Bongiovanni, R. Langmuir 2013, 29, 15711–15718.

Page 8 of 9

(16) Vitale, A.; Bongiovanni, R.; Ameduri, B. Chem. Rev. 2015, 115, 8835–8866. (17) Schneider, M. H.; Sieben, V. J.; Kharrat, A. M.; Mostowfi, F. Anal. Chem. 2013, 85, 5153–5160. (18) Bowden, S. A.; Wilson, R.; Parnell, J.; Cooper, J. M. Lab Chip 2009, 9, 828–832. (19) Mostowfi, F.; Czarnecki, J.; Masliyah, J.; Bhattacharjee, S. J. Colloid Interface Sci. 2008, 317, 593–603. (20) Fadaei, H.; Scarff, B.; Sinton, D. Energy & Fuels 2011, 25, 4829–4835. (21) Mostowfi, F.; Molla, S.; Tabeling, P. Lab Chip 2012, 12, 4381–4387. (22) Rowland, S. M.; Robbins, W. K.; Corilo, Y. E.; Marshall, A. G.; Rodgers, R. P. Energy & Fuels 2014, 28, 5043–5048. (23) Ho, C. M. B.; Ng, S. H.; Li, K. H. H.; Yoon, Y.-J. Lab Chip 2015, 15, 3627–3637. (24) Anciaux, S. K.; Geiger, M.; Bowser, M. T. Anal. Chem. 2016, 88, 7675–7682. (25) Stoleru, E.; Dumitriu, R. P.; Munteanu, B. S.; Zaharescu, T.; Tanase, E. E.; Mitelut, A.; Ailiesei, G-L.; Vasile, C. App. Surf. Sci. 2016, 367, 407-417. (26) Coltro, W. K. T.; Piccin, E.; Fracassi da Silva, J. A.; Lucio do Lago, C.; Carrilho, E. Lab Chip 2007, 7, 931–934. (27) Jarvis, J. M.; Robbins, W. K.; Corilo, Y. E.; Rodgers, R. P. Energy & Fuels 2015, 29, 7058–7064. (28) Adahchour, M.; Beens, J.; Vreuls, R. J. J.; Brinkman, U. A. T. TrAC Trends Anal. Chem. 2006, 25, 726–741. (29) Seeley, J. V.; Seeley, S. K. Anal. Chem. 2013, 85, 557–578. (30) Casilli, A.; Silva, R. C.; Laakia, J.; Oliveira, C. J. F.; Ferreira, A. A.; Loureiro, M. R. B.; Azevedo, D. A.; Aquino Neto, F. R. Org. Geochem. 2014, 68, 61–70. (31) Mostafa, A.; Górecki, T. Anal. Chem. 2016, 88, 5414–5423. (32) Harris, N. R.; Hill, M.; Beeby, S.; Shen, Y.; White, N. M.; Hawkes, J. J.; Coakley, W. T. Sens. Actuators B 2003, 95, 425−434. (33) Hamidi, H.; Mohammadian, E.; Asadullah, M.; Azdarpour, A.; Rafati, R. Ultrason. Sonochem. 2015, 26, 428−436. (34) Chokkalingam, V.; Ma, Y.; Thiele, J.; Schalk, W.; Tel, J.; Huck, W. T. S. Lab Chip 2014, 14, 2398−2402. (35) Sams, G. W.; Zaouk, M. Energy & Fuels 2000, 14, 31−37. (36) Kar, T.; Hascakir, B. Energy & Fuels 2015, 29, 3684−3690. (37) Santos, J. M.; Galaverna, R. S.; Pudenzi, M. A.; Schmidt, E. M.; Sanders, N. L.; Kurulugama, R. T.; Mordehai, A.; Stafford, G.; Wisniewski, A., Jr.; Eberlin, M. N. Anal. Methods 2015, 7, 4450−4463. (38) Peters, K. E.; Walters, C. C.; Moldowan, J. M. Biomarkers and Isotopes in the Environment and Human History, The Biomarker Guide, 2nd ed.; Cambridge University Press: Cambridge, 2005. (39) Filgueiras, P. R.; Portela, N. A.; Silva, S. R. C.; Castro, E. V. R.; Oliveira, L. M. S. L.; Dias, J. C. M.; Neto, A. C.; Romão, W.; Poppi, R. J. Energy & Fuels 2016, 30, 1972–1978. (40) Pereira, R. C. L.; Simas, R. C.; Corilo, Y. E.; Vaz, B. G.; Klitzke, C. F.; Schmidt, E. M.; Pudenzi, M. A.; Silva, R. M. C. F.; Moraes, E. T.; Bastos, W. L.; Eberlin, M. N.; Nascimento, H. D. L. Energy & Fuels 2013, 27, 7208–7216. (41) Chacón-Patiño, M. L.; Blanco-Tirado, C.; Orrego-Ruiz, J. A.; Gómez-Escudero, A.; Combariza, M. Y. Energy & Fuels 2015, 29, 1323-1331.

ACS Paragon Plus Environment

Page 9 of 9

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Analytical Chemistry

Table of Contents

9 ACS Paragon Plus Environment