Detection and Quantitation of Trace Fentanyl in Heroin by Surface

Subscriber access provided by UNIV OF LOUISIANA

Article

Detection and Quantitation of Trace Fentanyl in Heroin by Surface Enhanced Raman Spectroscopy Abed Haddad, Mircea Alex Comanescu, Omar Green, Thomas A. Kubic, and John R. Lombardi Anal. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.analchem.8b02909 • Publication Date (Web): 24 Sep 2018 Downloaded from http://pubs.acs.org on September 24, 2018

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 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 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.

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 25 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

Fig 1. Molecular structures of (a) fentanyl and(b) heroin. 80x76mm (300 x 300 DPI)

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

Fig 2. Characterization of AgPaper using: Scanning electron microscopy (SEM) at (a) 11,000x and (b) 90,000x magnification; (b) optical microscopy at 20x magnification; and (c) Raman spectrum showing Ag-Cl stretching mode at 238 cm¬-1¬. 177x71mm (300 x 300 DPI)


Page 2 of 25

Page 3 of 25 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


Fig 3. Characterization of AgPaper using SERS mapping at 20x over 780 spots and 1.7 mm2. Table 1. lists the percent of total acquisitions represented in each range. 88x38mm (300 x 300 DPI)



Fig 4. Limit of detection (LOD) for fentanyl on AgPaper, where the peak at 1005 cm-1 is detected down to 100 ng/l µg; inset shows the same peak at 100ng/mL 181x89mm (300 x 300 DPI)


Page 4 of 25

Page 5 of 25 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


Fig 5. Limit of detection (LOD) for fentanyl mixed with heroin, where the peak at 1005 cm-1 is detected down to 1% (100ng fentanyl:10 µg total); inset shows fentanyl peaks at 1% fentanyl in heroin. 182x88mm (300 x 300 DPI)



Fig 6. Calibration curve using the peak height ratios of fentanyl to heroin (1005cm¬-1 to 628cm-1) from 05%, and from 0-50% in the inset. Open circles (○) represent the calibration points, triangles (▲) the mean of predictive values, and red lines mark the 95% confidence interval for the calibration model. 80x39mm (300 x 300 DPI)


Page 6 of 25

Page 7 of 25 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


Fig 7. Swab test using AgPaper on two different surfaces were able to detect (a) heroin and (b) fentanyl (5% and 1%) in heroin. 75x77mm (300 x 300 DPI)



TOC Graphic 249x112mm (300 x 300 DPI)


Page 8 of 25

Page 9 of 25 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


Detection and Quantitation of Trace Fentanyl in Heroin by Surface Enhanced Raman Spectroscopy Abed Haddada,b, Mircea A. Comanescuc, Omar Greend, Thomas A. Kubicc, and John R. Lombardib a) Ph. D Program in Chemistry, City University of New York Graduate School and University Center, 365 5th Avenue, New York, NY, 10016 USA b) Department of Chemistry, City University of New York, City College of New York, 160 Convent Avenue, New York, NY 10031 USA (*Corresponding address; e-mail:

[email protected]) c) Ph. D Program in Criminal Justice, Forensic Science Specialization, City University of New York, John Jay College of Criminal Justice, 524 West 59th Street, New York, NY, 10019 USA d) Ionica Sciences, Inc., McGovern Center for Venture Development in the Life Sciences, 413 Weill Hall, 526 N. Campus Drive, Ithaca, NY 14853

ABSTRACT The identification of fentanyl, a main culprit in opioid overdose deaths, has become critical. While Raman spectroscopy is an effective tool for detecting illicit drugs, the weak intensity of Raman scattering can make it difficult to distinguish trace materials. This shortcoming is addressed by surface‐enhanced Raman spectroscopy (SERS), which produces strong signal enhancements when target compounds are near metal nanoparticles. This work examines the use of a paper-based substrate impregnated with silver nanoparticles for the detection of trace quantities of fentanyl alone, and as an adulterant in heroin. In addition, intensity ratios of diagnostic peaks associated with each substance were fitted to a Langmuir isotherm calibration model and used for quantitative analysis of fentanyl in heroin mixtures. Linearity was observed at

< 6% fentanyl, a significant finding that is consistent with

concentrations found in drugs seized during law enforcement efforts. In addition, swabbing with these paper-based SERS substrates facilitated the recovery of fentanyl from surfaces, showing this to be applicable for crime scene investigations. However, assessment using the calibration model proved difficult for swabbed samples. Overall, this work demonstrates a potentially



Page 10 of 25

simple and sensitive technique for forensic analysis and quantitation of fentanyl in trace amounts. Since the 1990’s, the synthetic opioid fentanyl (Fig. 1.a) has been widely prescribed and used as treatment for cancer and chronic pain.1 A Schedule II drug under the Controlled Substance Act, fentanyl has been used increasingly as an adulterant in illicit substances, such as heroin (Fig 1.b). Fentanyl is powerful enough that a few milligrams may cause overdose, and it has been recently implicated in a sharp uptick in overdose deaths in the United States and abroad.2–4

Fig 1. Molecular structures of (a) fentanyl and (b) heroin.

Consequently, the identification of fentanyl in trace quantities and in mixtures has become urgent in the face of the current epidemic. Structural characterization of fentanyl has been reported via nuclear magnetic resonance (NMR) and infrared (IR) spectroscopy coupled with density functional theory (DFT) calculations,5 and surface enhanced Raman spectroscopy (SERS) coupled with DFT calculations.6 Additionally, an assessment of fentanyl in a patient’s blood or after autopsy has been studied from a toxicological standpoint.7 Forensic work traditionally employs hyphenated methods such as gas chromatography-mass spectrometry (GC-MS) or liquid chromatography-mass spectrometry (LCMS), given established knowledge of their reliability and reproducibility.8,9 While methods like GC-MS, LC-MS are more commonly used, they require large sample sizes and long investigation times, are expensive to own and operate, and can be inappropriate for in-situ analysis. On the other hand, fentanyl often makes up 0.1% to 10% of seized street samples,4 and its identification can be difficult in small percentages even though it would provide crucial forensic insight. Alternatively, Raman spectroscopy has been increasingly used for precise evaluation of pure and adulterated samples heroin and cocaine.10–13 Raman spectroscopy is classified as a “Category A” technique by the Scientific Working Group for the Analysis of Seized Drugs (SWGDRUG) based on discriminating power, and has gained traction as a simple and robust 2 ACS Paragon Plus Environment

Page 11 of 25 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


method in forensic laboratories.14 Surface enhanced Raman spectroscopy (SERS) is especially sensitive because trace detection of laced drugs can be accomplished by analyte bonding to metal nanoparticles (NP). Investigative methods for SERS detection of fentanyl have recently been proposed via analysis of saliva,15 Au- or Ag-doped sol-gels immobilized in glass capillaries,16 or SERS followed by desorption electrospray ionization mass spectrometry (DESI-MS) using commercially available Au-coated paper substrates.17 While SERS traditionally relies on colloidal metal NP for signal enhancement, novel substrates amenable for use in the field have become a central focus for new avenues of analyte detection. Among those, paper-based substrates impregnated with metal NP have gained favor for their ease-of-use and cost-effectiveness.18-19 More significantly, they are highly appropriate for in-situ forensic sampling by swabbing surfaces of interest.20 Several methods for obtaining paper-based substrates have been reported, such as: inkjet printing of NP;21,22 direct application23 of or dip-coating paper in colloidal suspensions of NP;24 or in-situ nucleation and growth of NP.25 Some illustrations of SERS detection using paper-based substrates include dyes,26 explosives,27 pesticides,28 and biological materials.29 In this work, we report the detection of fentanyl using a paper-based SERS substrate made of lab-grade filter paper immersed in a colloidal suspension of AgNP. We evaluate the suitability of this method for finding the percent content of fentanyl in heroin, simulating its use as an adulterant in trace quantities. The applicability of this substrate— referred to as AgPaper for the rest of this study—

for identifying the presence of fentanyl on surfaces was also

investigated. It should be noted that this analysis is a first step that only considers homogenous mixtures of heroin and fentanyl, which can be rare in street samples; most often, benign fillers such as lidocaine, benzocaine, and sodium carbonate are also appear in other analyses.30–32 MATERIALS AND METHODS Reagents and Materials Silver nitrate (AgNO3) (99.9999% trace metals basis), sulfuric acid (ACS reagent, 95.098.0%), glucose (≥99.5%), and sodium citrate (≥99%, FG) were used for AgNP synthesis. Whatman® filter paper (Grade 1) and sodium chloride (BioXtra, ≥99.5%) were used in substrate fabrication. Heroin (1.0 mg/mL in acetonitrile, certified reference material, Cerilliant) and fentanyl (1.0 mg/mL in methanol and 100µg/mL in methanol; both certified reference material, Cerilliant) were used as analyte sources for SERS detection. All materials whose source is not 3 ACS Paragon Plus Environment


otherwise listed were purchased from Millipore-Sigma. AgNP Synthesis and Analyte Application AgNP were obtained through isothermal microwave-assisted reduction. The microwave technique, which was chosen in over the more commonly used Lee-Meisel method,26,33 provides more uniformity and stability of the NP. Silver sulfate (Ag2SO4) was first precipitated upon dropwise addition of 10% H2SO4 to an aqueous solution of AgNO3 in an ice bath. The resulting product was then dissolved in water to yield a final concentration of 5 x 10-4 M. 1 mL of 1% (w/v) glucose, 0.5 mL of 1% (w/v) sodium citrate, and 12.5 mL of the Ag2SO4 solution were added to a Teflon container and heated for 1 min at 120 ◦C and under high-pressure using a labgrade microwave digestion system (Anton Parr Microwave 3000). This was followed by centrifugation to remove excess sodium citrate and to concentrate the supernatant 20 times over using ultra-pure water (Millipore Milli-Q, 18.0 MΩ). This ensures that excess sodium citrate is removed and precludes citrate bands from appearing in the spectra. Substrate Fabrication The following method was adapted from Huang et. al., where this substrate was optimized for robust signal enhancement.29 In brief, filter paper discs (0.5 cm in diameter) were produced with a single-hole punch. They were then soaked for 5 min in 20mM solution of NaCl to imbue them with halide ions that promote AgNP aggregation and result in SERS-active “hot spots.”24 Each disc was individually soaked in 100 µL of concentrated suspensions of AgNP for 1 hr. The resulting substrates were stored in Eppendorf tubes and refrigerated until use. Whereas fresh substrates were used for all presented spectra, a longevity study was also performed. Substrates worked well after being stored and tested over a period of 15 days (Fig. S1). Substrate Characterization Electron microscopy (SEM) images were taken a Zeiss Supra 55 VP system at 5kV after drying and coating the substrate with 10nm of graphite, and energy dispersive X-ray spectroscopy (EDS) was done using a TESCAN Vega 3 XMU equipped with an EDAX Si(Li) detector at 20kV. Nanoparticle distribution measurements were done in ImageJ. Reflectance spectroscopy in the 400-700 nm range was taken with an eXact X-Rite spectrophotometer equipped with a gas filled tungsten filament Raman and SERS Analysis Normal Raman and SERS spectra were collected using a Renishaw In-via Raman system 4 ACS Paragon Plus Environment

Page 12 of 25

Page 13 of 25 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


equipped with a 532-diode laser, 1200 lines/grating, and a Leica confocal microscope with a 20x objective. Each of the illustrated spectra, unless otherwise specified, represents an average of 10 baseline-corrected,34 30 second acquisitions collected from different areas in the center of the substrate. For normal Raman, 10 µL of stock 1mg/mL fentanyl solution was dropped onto a silicon wafer and dried prior to acquisition (Fig. S2). SERS analysis of fentanyl solutions at decreasing concentrations, and binary mixtures containing fentanyl and heroin at different compositions, were applied to newly made AgPaper. Substrates were dried for 30 minutes before spectral analysis. In a parallel set of experiments that demonstrate the capacity of AgPaper for extracting target compounds, surfaces to which 20 µL of a binary mixture had been applied and evaporated were swabbed with discs moistened with methanol.

Fig 2. Characterization of AgPaper using: Scanning electron microscopy (SEM) at (a) 11,000x and (b) 90,000x magnification; optical microscopy at 20x magnification (b) before and (d) after exposure to AgNP; and (e) Raman spectrum showing Ag-Cl stretching mode at 238 cm-1.

RESULTS AND DISCUSSION Substrate Characterization Closer examination via SEM imaging showed adsorption of AgNP into the matrix with a good distribution (Fig 2.a) of spherical nanoparticles 45-65 nm in size (Fig. 2.b), and further corroborated by color change before and after soaking (Figs. 2.c,d, respectively). Halide-induced agglomeration was observed, where 1-2 nm gaps between AgNP can produce hotspots of heightened electromagnetic activity that contribute to sizeable enhancement factors.35–37 A Raman spectrum of blank AgPaper also reveals a characteristic peak at 238 cm-1 that corresponds to the Ag–Cl stretching vibration.38 EDS phase mapping of Ag L lines verified the adsorption of 5 ACS Paragon Plus Environment


Page 14 of 25

AgNP after drying (Fig. S3). Reflectance spectroscopy confirmed the presence of AgNP with a surface plasmon resonance (SPR) at 460nm (Fig S4.a). Whereas microwave-produced AgNP in colloid are characterized by a narrow SPR band at 406 nm (Fig. S4.b), surface modes of multiple AgNP are coupled in aggregates and the resulting SPR will be red-shifted and broadened when compared to individual nanoparticles.39–41 Using ImageJ software, the calculated surface area of nanoparticles comprised ~23.4 % of the substrate surface area. This level of coverage (~25%) was also observed by Hasi, et al when using a similar approach. 24 Table 1. Percent of Enhancement Factor (EF) Distribution Range Percent 1 x 107 - 2 x 107

16.1

2 x 107 - 3 x 107

36.1

7

7

29.3

4 x 107 - 5 x 107

14.7

3 x 10 - 4 x 10 7

5 x 10 - 6 x 10

7

3.90

Fig 3. Characterization of AgPaper using SERS mapping at 20x magnification over 780 spots and 1.7 mm2. Table 1. lists the percent of total acquisitions represented in each range.

Fentanyl Normal Raman and SERS The enhancement factor (EF) across a 1.7 mm2 area in the center of the substrate was assessed by mapping of 780 SERS baseline-corrected spectra of 10 µL of 1mg/mL fentanyl, all obtained for 1 second at 2.5 mW (Fig. 3). For direct comparison, an average of 10 normal Raman spectra of fentanyl using the same parameters was obtained (Fig. S2). EF is the ratio of SERS signal intensity per molecule (ISERS/ NSERS) divided by the normal Raman signal intensity per molecule (INR/NNR). The most intense line in the spectrum was selected for this calculation, a CC-C trigonal stretch in benzene located at 1005 cm-1.6 NNR and NSERS are the number of molecules exposed to the laser during normal Raman and SERS measurements, respectively. EF can be rearranged as: =

6 ACS Paragon Plus Environment

(1)

Page 15 of 25 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


The number of molecules under laser exposure in NSERS assumes that a monolayer bonded to AgNP causes an enhanced spectrum and not a normal one. Using a beam (r) radius of 1 µm, the flat area of spot is πr2 = 3.14 µm2. Fentanyl attaches to AgNP is via a carbonyl bond, which leaves it flat with respect to the surface.6 The cross-sectional area of a fentanyl molecule is calculated using its length and width (1.8 nm and 0.42 nm, respectively) for a total area = 0.76 nm2 per molecule. Presuming full monolayer coverage of AgNP, the number of molecules in NSERS would ~23.4% of the total area; i.e.

. . .

= 9.67 x 105 molecules. As there is

undoubtedly less than complete surface coverage, this approximation offers a lower limit for calculating EF, To find out the total amount of molecules for NNR, a laser penetration depth (d) of 0.1 µm was used. The volume of fentanyl taken up by the laser is πr2d = 3.14 µm2 (0.1 µm) = 3.14 x 1010

cm3. Using the density (ρ) and molecular weight of fentanyl (1.087 g/cm3 and 336 g/mol,

respectively), NNR = 5.4x 1011 molecules. While the (NNR/NSERS) ratio remained uniform at 5.58 x 105, (INR/ISERS) ratio was varied over a range of ISERS constant and INR = 85 counts. This produced enhancement factors between 1.03 x 107 to 5.87 x 107 (Eqn. 1). While not uniform, a greater majority of the values (65.4%) are within one standard deviation from an average of 3.05 x 107 (Table 1.)

Fig 4. Limit of detection (LOD) for fentanyl on AgPaper, where the peak at 1005 cm-1 is detected down to 100 ng/l µg; inset shows the same peak at 100ng/mL



Since fentanyl is used as an adulterant in low concentrations, a limit of detection (LOD) of fentanyl via serial dilutions with methanol was examined to demonstrate the efficacy of AgPaper(Fig. 4). 10 µL of each dilution was applied to AgPaper, and fentanyl was faintly but readily detected at 100 ng/mL using the 1005 cm-1 peak. This peak has been used as a visual diagnostic for the presence of fentanyl alone or in mixtures.17,42,43 The following are percent relative standard deviation (RSD) values for each of the concentrations: 18.61% for 50 µg/mL; 24.37% for 10 µg/mL, 34.31 % for 1 µg/mL; and 40.48% for 100 ng/mL. The RSD in SERS has been repeatedly demonstrated to be variable and somewhat higher than traditional established quantitation methods, and dependent on the type of substrate, concentration, and solution being used.45

Fig 5. Limit of detection (LOD) for fentanyl mixed with heroin, where the peak at 1005 cm-1 is detected down to 1% (100ng fentanyl:10 µg total); inset shows fentanyl peaks at 1% fentanyl in heroin.

Binary Mixture Analysis The quantitation of binary mixtures can rely on ratios between isolated signature peaks for each of the target analytes (Eqn. 2).44 Depending on binding interactions with AgNP, the relationship between fentanyl and heroin concentrations could act as a standard against which unknown mixtures are evaluated. Peak heights were based on the maxima of chosen diagnostic peaks for each drug: 628 cm-1 for heroin – assigned to several ring modes16 – and 1005 cm-1 for fentanyl (Fig. 5). The ratios of peak height were calculated using the following equation: 8 ACS Paragon Plus Environment

Page 16 of 25

Page 17 of 25 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


=

( − ) ( + )

(2)

where I is the peak intensity after baseline correction and R is the resulting ratio. All calibration data ratio values for zero percent fentanyl produced negative values. Subsequently, each data set was initially forced through the origin by adding the smallest zero fentanyl ratio to all data points, effectively giving them positive values for calibration. This constant was retained for each model and added to the unknown binary mixture intensity ratios for accurate prediction and were fitted to a Langmuir isotherm with the equation:

!=

"#$ (1 + #$)

(3)

where q and b were calculated by taking the inverse intercept, and the intercept divided by the slope of the best fit line for the reciprocal of each data set. In this model, qb is the initial slope of the curve, while b is considered the affinity coefficient of fentanyl to the surface, q the maximum ratio threshold, and θ is the calculated peak ratio. A calibration model was constructed by plotting peak height ratios for standards with a total analyte mass of 10 µg, altering heroin volume from stock according to the desired mass/mass percent composition(Fig. 6). 10 µL of each solution was applied to AgPaper for

9 Fig 6. Calibration curve using the peak height ratios of fentanyl to heroin (1005cm-1 to 628cm-1) from 0-5%, ACS Paragon Plus Environment and from 0-50% in the inset. Open circles (○) represent the calibration points, triangles (▲) the mean of predictive values, and red lines mark the 95% confidence interval for the calibration model.


SERS analysis. The curve represents 1, 3, 5, 10, 30 and 50% mass of fentanyl in heroin (100ng:10 µg, 300ng:10 µg, 500:10 µg, 3 µg:10 µg, and 5 µg:10 µg fentanyl: total mass respectively). The calibration curve (Eqn. 3) exhibits linear behavior before beginning to plateau gradually after 5% fentanyl and severely after 20% (Fig. 6-inset). Such linearity was also seen at sub-microgram levels in a Langmuir calibration of SERS intensity versus concentration of heroin.21 The flattening effect is attributed to higher concentrations of fentanyl and saturation of the AgPaper that produces nonlinear behavior unfit for predictive evaluation.44 The presence of a high density of hot-spots in the AgPaper led to greater enhancement of fentanyl with increasing concentrations and similar peak ratios above 20%. Nevertheless, the critical concentrations of fentanyl found in street samples of heroin are represented in the linear portion of the curve.4 The curve exhibits an R2 of 0.9644 with a p-value of 2.2x10-16 which, in addition to the residuals plot of the regression data (Fig. S5) showing a small range consistent with random error, supports the calibration model as a good fit for the data set. Predictive Analysis Spectra of solutions containing varying concentrations between 0.1% and 20% mass fentanyl in heroin were collected to test the predictive ability of the model. A total analyte mass of 10 ug was maintained, altering heroin (1 mg/mL) and fentanyl (both 1mg/mL and 100µg/mL) volumes from stocks according to the desired mass/mass percent composition. The mean of each prediction, RSD, absolute and percent errors to the actual concentration of fentanyl, and the mean absolute percent error (MAPE), which indicates overall accuracy of the model, are all reported here (Table 2.). Small MAPE values convey better overall accuracy and good predictive behavior. At lower concentrations (0 – 6%) of fentanyl, a MAPE of 5.80% is observed, and higher (> 6%) concentrations show a MAPE of close to 20%. The model was a great fit to the data and predicted low-fentanyl content with success, evident by low percent biases to actual values and corroborated by the below 4.25% - 95% confidence interval RSD of the model relative to the reported means. If single-component solutions of medial interactivity with the substrates can have RSD of 5-16%,45 one would expect similar or slightly higher RSD values for concentration calculations in mixtures of components that have comparable binding affinities, especially towards the lower end of concentration curves, an effect described by the Horowitz function.46,47 10 ACS Paragon Plus Environment

Page 18 of 25

Page 19 of 25 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


Nevertheless, this approach could address the needs of forensic application since it quantified the trace amounts of fentanyl typically found in street samples. Other statistical methods have been explored to investigate the potential of expanding the quantitation to higher concentrations of fentanyl. Linear discriminant analysis (LDA) shows a good grouping of the standard replicate data (Fig. S6). However, analyses using either principal component regression (PCR) or partial least squares regression (PLSR) models (Figs. S7 and S8, respectively)— with 100 replications and an optimal number of components determined by the model building R— do not show an improvement on the peak-ratio calibration model when put in practice. Interestingly, PLSR determined two components for optimal analysis, in the same way our calibration model uses two components, i.e. two individual peaks from the target analytes.

Table 2. Prediction mean values of percent mass fentanyl in mixtures compared by peak height calibration curve Actual Percent Fentanyl 0.55 Error

0.70 Error

0.95 Error

1.5 Error

2.45 Error

3.65

Predictive Evaluation Mean RSD Absolute Percent 95% CI Mean RSD Absolute Percent 95% CI Mean RSD Absolute Percent 95% CI Mean RSD Absolute Percent 95% CI Mean RSD Absolute Percent 95% CI Mean RSD

Actual Percent Fentanyl

0.47 27.65 0.08 14.55 0.02 0.61 12.99 0.09 12.86 0.02 0.88 33.43 0.07 7.37 0.04 1.38 24.78 0.12 8.00 0.06 2.21 12.85 0.24 9.80 0.12 3.24 10.13

5.4 Error

6.1 Error

7.9 Error

12.3 Error

18.7 Error

MAPE (Low Conc.)


Predictive Evaluation Mean RSD Absolute Percent 95% CI Mean RSD Absolute Percent 95% CI Mean RSD Absolute Percent 95% CI Mean RSD Absolute Percent 95% CI Mean RSD Absolute Percent 95% CI

4.85 10.70 0.55 10.19 0.39 5.39 34.66 0.71 11.64 0.50 9.63 32.33 1.73 21.90 1.38 19.70 11.14 7.40 60.16 4.81 28.00 18.58 9.30 49.73 10.80 5.80


Error

Absolute Percent 95% CI

0.41 11.23 0.21

MAPE (Overall)

Swab Test SERS spectra of pure heroin at 10 µg were obtained by swabbing surfaces onto which solutions of the drug were introduced (Fig. 7.a). Two representative surfaces were selected: an absorbent, unfinished wooden desk and a nonabsorbent

epoxy counter.

The intense peak at 628 cm-1 is consistently

enhanced

by

AgPaper, and those at 444, 500, 590, and 679 cm-1 show modest yet detectable signals that confirm the presence of heroin.12,17

While

the

roughness and absorptivity of

Fig 7. Swab test using AgPaper on two different surfaces were able to detect (a) heroin and (b) fentanyl (5% and 1%) in heroin.

the unfinished wood contributed to the weakness in the intensity of those features, the presence of heroin is unmistakable in comparison to the Raman spectrum of blank AgPaper. The same swab test was applied to 10% and 5% (1 µg:10 µg and 0.5 µg:10 µg) fentanyl in heroin mixtures (Fig. 7.b). Again, the diagnostic peak at 1005 cm-1 for fentanyl is present in addition to that of heroin’s at 628 cm-1. Quantitation of the swabbed samples proved difficult due to the great enhancement of the fentanyl peak (Fig. 5.b). Nonetheless, the ability to swab surfaces and identify the presence of illicit substances can be useful for rapid analysis in a crime scene. CONCLUSION

The use of a low-cost AgPaper substrate prepared by immersion in AgNP suspension allowed for rapid and simple SERS identification of fentanyl, alone or in mixtures with heroin. 12 ACS Paragon Plus Environment

Page 20 of 25

19.77

Page 21 of 25 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


Quantitation of fentanyl as an adulterant in sub-microgram concentrations using a peak-height calibration model presents an appealing choice for forensic analysis and avoids complex internal standards. Reduced analysis time, compared to currently applied methods, could help unwind the long forensic sample backlogs vital to court cases and law enforcement work. It can also serve as a compliment to traditional chromatographic approaches. With the growing availability of portable Raman instrumentation, identification of fentanyl and heroin by swabbing would facilitate the use of this technique in the field. A quantitative study will be performed in the future to assess ternary and quaternary mixtures with common adulterants (e.g. sorbitol, mannitol, lidocaine, sodium bicarbonate), which better reflects the nature of street drugs. ACKNOWLEDGEMENTS

We would like to thank the Museum of Modern Art, namely Chris McGlinchey and Ana Martins, for instrumental access and research facilities. We are indebted to the National Science Foundation (CHE-1402750) for partial funding of this project. This work was also partially supported by National Science Foundation grant number HRD-1547830 (IDEALS CREST). Further support came from the City University of New York PSC-CUNY Faculty Research Award Program, Grant No. 69079. Author Information Corresponding Author * E-mail: [email protected] ORCID Abed Haddad: 0000-0001-8710-039 Notes The opinions, findings, and conclusions or recommendations expressed in this publication are those of the authors and do not necessarily reflect those of the National Science Foundation. The authors declare no competing financial interest. SUPPORTING INFORMATION We include SERS spectra showing the potential of AgPaper for detecting 10 µg fentanyl over 15 days. Normal Raman and SERS spectra used for enhancement calculations are also presented. A figure with a scanning electron microscopy (SEM) image and energy x-ray dispersive spectroscopy (EDS) of the same site that shows the distribution of silver on AgPaper sample is illustrated. We also show a reflectance spectrum of AgPaper— with an inset of that of 13 ACS Paragon Plus Environment


the blank filter paper— and an absorbance spectrum of AgNP over the same wavelength range. We also include various statistical tests done in conjunction with data presented in the manuscript: a residuals plot of the calibration data for fentanyl and heroin mixtures that shows a small range; a two-dimensional linear discriminant plot (LDA) of fentanyl-heroin mixture standards as used for the calibration model to show clustering of SERS spectra; and quantitative analyses of fentanyl-heroin mixtures by Principal component regression (PCR) using five components mixtures and by partial least squares regression (PLSR) using two components. REFRENCES (1) (2) (3) (4)

(5)

(6)

(7)

(8) (9) (10)

(11)

(12) (13)

(14) (15)

DEA Issues Nationwide Alert on Fentanyl as Threat to Health and Public Safety; Drug Enforcement Administration, 2015. Seth, P. Overdose Deaths Involving Opioids, Cocaine, and Psychostimulants — United States, 2015–2016. MMWR 2018, 67. UNODC. Global SMART Update Vol. 17, 2017. Marinetti, L. J.; Ehlers, B. J. A Series of Forensic Toxicology and Drug Seizure Cases Involving Illicit Fentanyl Alone and in Combination with Heroin, Cocaine or Heroin and Cocaine. J. Anal. Toxicol. 2014, 38, 592–598. Asadi, Z.; Esrafili, M. D.; Vessally, E.; Asnaashariisfahani, M.; Yahyaei, S.; Khani, A. A Structural Study of Fentanyl by DFT Calculations, NMR and IR Spectroscopy. J. Mol. Struct. 2017, 1128, 552–562. Leonard J.; Haddad A.; Green Omar; Birke R. L.; Kubic T.; Kocak A.; Lombardi J. R. SERS, Raman, and DFT Analyses of Fentanyl and Carfentanil: Toward Detection of Trace Samples. J. Raman Spectrosc. 2017, 48, 1323–1329. Cobaugh, D. J.; Gainor, C.; Gaston, C. L.; Kwong, T. C.; Magnani, B.; McPherson, M. L.; Painter, J. T.; Krenzelok, E. P. The Opioid Abuse and Misuse Epidemic: Implications for Pharmacists in Hospitals and Health Systems. Am. J. Health-Syst. Pharm. 2014, 71, 1539–1554. Peters, F. T. Recent Advances of Liquid Chromatography–(Tandem) Mass Spectrometry in Clinical and Forensic Toxicology. Clin. Biochem. 2011, 44, 54–65. Brettell, T. A. Forensic Science Applications of Gas Chromatography. Modern Practice of Gas Chromatography; Wiley-Blackwell, 2004; pp 883–967. Penido, C. A. F. de O.; Pacheco, M. T. T.; Novotny, E. H.; Lednev, I. K.; Silveira, L. Quantification of Cocaine in Ternary Mixtures Using Partial Least Squares Regression Applied to Raman and Fourier Transform Infrared Spectroscopy. J. Raman Spectrosc. 2017, 48, 1732–1743. Penido, C. A. F.; Pacheco, M. T. T.; Lednev, I. K.; Silveira, L. Raman Spectroscopy in Forensic Analysis: Identification of Cocaine and Other Illegal Drugs of Abuse. J. Raman Spectrosc. 2016, 47, 28–38. Zhao, J.; Xu, C.; Duan, Y.; Zhang, P.; Si, M.; Yang, Q. Raman investigating of trace contraband heroin by micro Raman. Guangpuxue Yu Guangpu Fenxi 2002, 22, 588–590. Dies, H.; Raveendran, J.; Escobedo, C.; Docoslis, A. Rapid Identification and Quantification of Illicit Drugs on Nanodendritic Surface-Enhanced Raman Scattering Substrates. Sens. Actuators, B 2018, 257, 382–388. SWGDRUG. Recommendations Ed. 7.1; SWGDRUG, 2016. Inscore, F.; Shende, C.; Sengupta, A.; Huang, H.; Farquharson, S. Detection of Drugs of Abuse in Saliva by Surface-Enhanced Raman Spectroscopy (SERS). Appl. Spectrosc. 2011, 65, 1004–1008.


Page 22 of 25

Page 23 of 25 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


(16)

(17)

(18) (19) (20) (21) (22) (23) (24)

(25)

(26)

(27) (28)

(29)

(30) (31)

(32) (33) (34) (35)

Farquharson, S.; Shende, C.; Sengupta, A.; Huang, H.; Inscore, F. Rapid Detection and Identification of Overdose Drugs in Saliva by Surface-Enhanced Raman Scattering Using Fused Gold Colloids. Pharmaceutics 2011, 3, 425–439. Fedick, P. W.; Bills, B. J.; Manicke, N. E.; Cooks, R. G. Forensic Sampling and Analysis from a Single Substrate: Surface-Enhanced Raman Spectroscopy Followed by Paper Spray Mass Spectrometry. Anal. Chem. 2017, 89, 10973–10979. Betz, J. F.; Yu, W. W.; Cheng, Y.; White, I. M.; Rubloff, G. W. Simple SERS Substrates: Powerful, Portable, and Full of Potential. Phys. Chem. Chem. Phys. 2014, 16, 2224–2239. Chamuah, N.; Hazarika, A.; Hatiboruah, D.; Nath, P. SERS on Paper: An Extremely Low Cost Technique to Measure Raman Signal. J. Phys. D: Appl. Phys. 2017, 50, 485601. Lee, C. H.; Tian, L.; Singamaneni, S. Paper-Based SERS Swab for Rapid Trace Detection on Real-World Surfaces. ACS Appl. Mater. Interfaces 2010, 2, 3429–3435. Yu, W. W.; White, I. M. Inkjet-Printed Paper-Based SERS Dipsticks and Swabs for Trace Chemical Detection. Analyst 2013, 138, 1020–1025. Boyd, S.; Bertino, M. F.; Ye, D.; White, L. S.; Seashols, S. J. Highly Sensitive Detection of Blood by Surface Enhanced Raman Scattering. J. Forensic Sci. 2013, 58, 753–756. He, S.; Chua, J.; Tan, E. K. M.; Kah, J. C. Y. Optimizing the SERS Enhancement of a Facile Gold Nanostar Immobilized Paper-Based SERS Substrate. RSC Adv. 2017, 7, 16264–16272. Hasi, W.-L.-J.; Lin, X.; Lou, X.-T.; Lin, S.; Yang, F.; Lin, D.-Y.; Lu, Z.-W. Chloride Ion-Assisted Self-Assembly of Silver Nanoparticles on Filter Paper as SERS Substrate. Appl. Phys. A 2015, 118, 799–807. Sallum, L. F.; Soares, F. L. F.; Ardila, J. A.; Carneiro, R. L. Determination of Acetylsalicylic Acid in Commercial Tablets by SERS Using Silver Nanoparticle-Coated Filter Paper. Spectrochim. Acta, Part A 2014, 133, 107–111. Wu, M.; Li, P.; Zhu, Q.; Wu, M.; Li, H.; Lu, F. Functional Paper-Based SERS Substrate for Rapid and Sensitive Detection of Sudan Dyes in Herbal Medicine. Spectrochim. Acta, Part A 2018, 196, 110–116. Fierro-Mercado, P. M.; Hernández-Rivera, S. P. Highly Sensitive Filter Paper Substrate for SERS Trace Explosives Detection. Int. J. Spectrosc. 2012, 2012, Article ID 716527. Lee, M.; Oh, K.; Choi, H.-K.; Lee, S. G.; Youn, H. J.; Lee, H. L.; Jeong, D. H. Subnanomolar Sensitivity of Filter Paper-Based SERS Sensor for Pesticide Detection by Hydrophobicity Change of Paper Surface. ACS Sens. 2018, 3, 151–159. Huang, Z.; Cao, G.; Sun, Y.; Du, S.; Li, Y.; Feng, S.; Lin, J.; Lei, J. Evaluation and Optimization of Paper-Based SERS Substrate for Potential Label-Free Raman Analysis of Seminal Plasma. Int. J. Spectrosc. 2017, 2017, Article ID 4807064. Behrman, A. D. Luck of the Draw: Common Adulterants Found in Illicit Drugs. J. Emerg. Nurs. 2008, 34, 80–82. Broséus, J.; Gentile, N.; Bonadio Pont, F.; Garcia Gongora, J. M.; Gasté, L.; Esseiva, P. Qualitative, Quantitative and Temporal Study of Cutting Agents for Cocaine and Heroin over 9 Years. Forensic Sci. Int. 2015, 257, 307–313. Cole, C.; Jones, L.; McVeigh, J.; Kicman, A.; Syed, Q.; Bellis, M. Adulterants in Illicit Drugs: A Review of Empirical Evidence. Drug Test. and Anal. 2011, 3, 89–96. Lee, P. C.; Meisel, D. Adsorption and Surface-Enhanced Raman of Dyes on Silver and Gold Sols. J. Phys. Chem. 1982, 86, 3391–3395. Zhang, Z.-M.; Chen, S.; Liang, Y.-Z. Baseline Correction Using Adaptive Iteratively Reweighted Penalized Least Squares. Analyst 2010, 135, 1138–1146. Caro, C.; López‐Cartes, C.; Zaderenko, P.; Mejías, J. A. Thiol-Immobilized Silver Nanoparticle Aggregate Films for Surface Enhanced Raman Scattering. J. Raman Spectrosc. 2008, 39, 1162– 1169.



(36)

(37) (38) (39) (40)

(41)

(42) (43) (44)

(45)

(46) (47)

Sun, L.; Zhao, D.; Ding, M.; Xu, Z.; Zhang, Z.; Li, B.; Shen, D. Controllable Synthesis of Silver Nanoparticle Aggregates for Surface-Enhanced Raman Scattering Studies. J. Phys. Chem. C 2011, 115 , 16295–16304. Maher, R. C. SERS Hot Spots. In Raman Spectroscopy for Nanomaterials Characterization; Kumar, C. S. S. R., Ed.; Springer Berlin Heidelberg: Berlin, Heidelberg, 2012; pp 215–260. Kim, M.; Itoh, K. PH Dependence of Surface-Enhanced Raman Scattering Spectra from 2,2’Bipyridine Adsorbed on Silver Colloids. J. Phys. Chem. 1987, 91, 126–131. Bilankohi, S. M. Optical Scattering and Absorption Characteristics of Silver and Silica/ Silver Core/Shell Nanoparticles. Orient. J. Chem. 2015, 31, 2259–2263. Smitha, S. L.; Nissamudeen, K. M.; Philip, D.; Gopchandran, K. G. Studies on Surface Plasmon Resonance and Photoluminescence of Silver Nanoparticles. Spectrochim. Acta, Part A 2008, 71, 186–190. Karpov, S. V.; Gerasimov, V. S.; Grachev, A. S.; Isaev, I. L.; Podavalova, O. P.; Slabko, V. V. Experimental Manifestations of the Correlation between the Local Structure of Silver Nanoparticle Aggregates and Their Absorption Spectra. Colloid J. 2007, 69, 170–179. B&W Tek, Inc. https://www.azom.com/article.aspx?ArticleID=13473 (accessed May 20, 2018). Diagnostic anSERS Inc. http://www.diagnosticansers.com/application-examples/ (accessed May 20, 2018). Comanescu, M. A.; Muehlethaler, C.; Lombardi, J. R.; Leona, M.; Kubic, T. A. Competitive Binding Investigations and Quantitation in Surface-Enhanced Raman Spectra of Binary Dye Mixtures. Appl. Spectrosc. 2018, 72, 60–68. Muehlethaler, C.; Leona, M.; Lombardi, J. R. Towards a Validation of Surface-Enhanced Raman Scattering (SERS) for Use in Forensic Science: Repeatability and Reproducibility Experiments. Forensic Sci. Int. 2016, 268, 1–13. Horwitz, W.; Kamps, L. R.; Boyer, K. W. Quality Assurance in the Analysis of Foods and Trace Constituents. J. Asso.c Off. Anal. Chem. 1980, 63, 1344–1354. Mesley, R. J.; Pocklington, W. D.; Walker, R. F. Analytical Quality Assurance. A Review. Analyst 1991, 116, 975–990.


Page 24 of 25

Page 25 of 25 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


FOR TOC ONLY


Detection and Quantitation of Trace Fentanyl in Heroin by Surface

Recommend Documents