Forensic Chemistry 3 (2017) 7480

Contents lists available at ScienceDirect

Forensic Chemistry
journal homepage: www.elsevier.com/locate/forc

Case Reports

Analysis of illicit carfentanil: Emergence of the death dragon

John F. Casale , Jennifer R. Mallette, Elizabeth M. Guest
U.S. Drug Enforcement Administration, Special Testing and Research Laboratory, Dulles, VA 20166, USA

a r t i c l e i n f o a b s t r a c t

Article history: The United States is currently in the midst of an unprecedented illicit fentanyl crisis. In the last year alone
Received 8 December 2016 there have been more recent fentanyl-related overdose deaths than in the previous 60 years. The current
Available online 16 February 2017 crisis is multi-faceted and involves a global supply of fentanyl and related substances being smuggled
into the United States. Illicit carfentanil hydrochloride has recently entered the drug market causing mul-
Keywords: tiple overdoses and deaths across the U.S. To date, over 400 confirmed carfentanil cases have been iden-
Carfentanil tified. Carfentanil is approximately 100 times more potent that fentanyl with only 20 lg of material
Analysis
required to produce a lethal dose. Due to a lack of published spectra for carfentanil HCl, analytical profiles
Chromatography
Spectroscopy
are provided for three recent carfentanil submissions to our laboratory which include infrared spec-
Spectrometry troscopy, nuclear magnetic resonance spectroscopy, gas chromatography-mass spectrometry, isotope
Forensic chemistry ratio mass spectrometry, and quantitative determination via gas chromatographyflame ionization
detection. The three submissions were determined to contain 0.62%, 1.87%, and 0.31% carfentanil HCl,
respectively. Each exhibit also contained a fentanyl-related substance (fentanyl or 2-furanylfentanyl).
Acetylcarfentanil was characterized as an impurity in two exhibits. Isotopic analyses of two exhibits sug-
gest they are intimately related.
Published by Elsevier B.V.

1. Introduction
Over the past 40 years there have been three significant
fentanyl-related epidemics within the United States. The first epi-
demic began in the late 1970s and was dubbed China White due
to the nature, purity, and color of the fentanyl and fentanyl-related
compounds (FFRCs). The FFRCs were determined to be of Chinese
origin and caused hundreds of overdose deaths. Many of the FFRCs
were new compounds not previously seen by the forensic commu-
nity, and extensive research was conducted to characterize and
identify these new substances [13]. These FFRCs were a major
influence in the implementation of the Controlled Substance Ana-
logue Enforcement Act of 1986, which controlled drug analogs of
listed Schedule I and II drugs. A second wave of over 1000 fatal
overdoses from illicit fentanyl occurred during the 20052007
time frame. Illicit fentanyl was determined to be originating from
a clandestine laboratory in Mexico. Our laboratory was tasked with
providing chemical intelligence on seized fentanyl exhibits. A
fentanyl signature profiling research project was established to
determine the clandestine synthetic routes and link seized sam-
Corresponding author at: Special Testing and Research Laboratory, U.S. Drug
Enforcement Administration, 22624 Dulles Summit Court, Dulles, VA 20166-9509,
USA.
E-mail address: John.f.casale@usdoj.gov (J.F. Casale).
ples. Two profiling methods were developed, one of which was
published [4]. In that work, 40 synthetic markers were identified;
12 were Janssen route-specific, 7 were Siegfried route-specific,
and 21 were non route-specific. The markers were significant in
determining relationships between samples.
The current outbreak of FFRCs began in late 2013 and accounted
for over 5000 overdose deaths within the U.S. during 2014, as
reported by the Center for Disease Control (CDC) [5,6]. A compila-
tion of FFRC overdose deaths in 2015 is not yet available, but Ohio
has reported over 1100 for that state alone. The only 2016 data
available at this time is for New Hampshire with 200 opiate/opioid
deaths in the first 8 months. However, the National Forensic Labo-
ratory Information System (NFLIS) has tracked a dramatic increase
in fentanyl submissions to local, state, and federal laboratories over
the last 4 years (CY 2013 = 934, CY 2014 = 7864, CY 2015 = 13,059,
and CY 2016 through September = 16,263). If the relationship of
NFLIS sample submissions to CDC overdose deaths is consistent
between 2014 and 2016, there could well be over 10,000 overdose
deaths during CY 2016. Recently, the extremely potent FFRC, car-
fentanil (Fig. 1) (10,000 more potent that morphine, 100 more
potent that fentanyl) has been detected in seizures and overdose
death toxicology samples. Carfentanil was implicated in 125
deaths that occurred during the Moscow Theatre Siege of 2002
[7]. A lethal dose of carfentanil is estimated at only 20 micrograms
(20 one-millionths of a gram) and is barely visible to the naked eye.

http://dx.doi.org/10.1016/j.forc.2017.02.003
2468-1709/Published by Elsevier B.V.
 J.F. Casale et al. / Forensic Chemistry 3 (2017) 7480 75

Fig. 1. Structures of carfentanil (1), acetylcarfentanil (2), fentanyl (3), acetylfentanyl (4), and 2-furanylfentanyl (5).

During CY 2016, this laboratory has been notified of over 400 con-
firmed carfentanil cases within the U.S., with the overwhelming
majority (84%) occurring in Ohio (Table 1). Herein, we report ana-
lytical data for illicit carfentanil from three case exhibits which
include infrared spectroscopy (IR), nuclear magnetic resonance
spectroscopy (NMR), gas chromatographymass spectrometry
(GC/MS), isotope ratio mass spectrometry (IRMS), and quantitative
determination via gas chromatographyflame ionization detection
(GCFID).
2. Cases
Exemplars of two cases comprised of three total exhibits were
submitted to this laboratory for in-depth analysis. Specific details
for each exhibit cannot be disclosed due to current and active
enforcement investigations; however, the following non-sensitive
information is provided:
Case #1: Two separate exhibits (1A and 1B) of white powder
seized in the central U.S. weighing approximately 2 and 20 g,

Table 1
Confirmed carfentanil cases examined by local, state, and federal forensic laboratories during 2016.a

Reported drug with carfentanil OH MI FL KY IL RI GA IN Total

Carfentanil only 48 1 8 7 0 0 3 0 67
Fentanyl 3 0 0 0 0 0 0 2 5
Furanylfentanyl 4 0 0 0 0 4 0 0 8
Heroin 20 0 2 5 0 0 0 0 27
Heroin/fentanyl 8 0 0 3 1 0 0 0 12
Heroin/furanylfentanyl 1 0 0 0 0 0 0 0 1
Heroin/fentanyl/furanyfentanyl 1 3 0 0 0 0 0 0 4
Other controlled substancesb 8 1 1 0 0 0 0 0 10
No detailed reporting 250 0 23 0 0 0 0 0 273
Total 343 5 34 15 1 4 3 2 407
a
Data compiled by DEA Special Testing and Research Laboratory as of October 26, 2016.
b
Includes mixtures containing cocaine, FUB-AMB, U47700, methamphetamine, and cannabinoids.
76 J.F. Casale et al. / Forensic Chemistry 3 (2017) 7480

respectively. Carfentanil and lactose were identified by the submit-
ting laboratory; however, quantitative data was not acquired.
Case #2: One exhibit of brown powder seized in the north cen-
tral U.S. weighing approximately one kilogram. Heroin, carfentanil,
fentanyl, and lactose were identified by the submitting laboratory;
however, quantitative data was not acquired.
3. Methods
All solvents were distilled-in-glass products of Burdick and
Jackson Laboratories (Muskegon, MI). All other chemicals were of
reagent-grade quality and products of Sigma-Aldrich Chemical
(St. Louis, MO). Carfentanil citrate (Wildnil) was a product of
Wildlife Pharmaceuticals (Fort Collins, CO). Carfentanil hydrochlo-
ride was obtained through purification and authentication of mate-
rial from one of the listed case exhibits.
Carfentanil quantitation was performed using an Agilent (Palo
Alto, CA) Model 7890A gas chromatograph with flame ionization
detection. The GC system was fitted with a 30 m  0.25 mm ID
fused-silica capillary column coated with DB-1 (0.25 mm) (Agilent,
Santa Clara, CA) in constant flow mode at 43.5 cm/s of hydrogen
carrier gas. The injection port and flame ionization detectors were
maintained at 280 C. The oven temperature was programmed as
follows: Initial temperature, 250 C; initial hold, 12.0 min; program

Fig. 2. Partial GC/MS reconstructed total ion chromatograms for CH2Cl2 soluble extracts of (a) Case #1, Exhibit 1A, (b) Case #1, Exhibit 1B, and (c) Case #2, Exhibit 1. Peak
Identification: 1 = diphenhydramine, 2 = acetylcodeine, 3 = O6-monoacetylmorphine, 4 = heroin, 5 = fentanyl, 6 = acetylcarfentanil, 7 = carfentanil, 8 = 2-furanylfentanyl, and
9 = noscapine.
 J.F. Casale et al. / Forensic Chemistry 3 (2017) 7480 77

rate, 25.0 C/min; final temperature, 280 C; final hold, 3.8 min.
Nitrogen was used as the auxiliary make-up gas for the detector.
Samples were injected (2 lL injection) in the split mode (21:1)
by an Agilent 7683 Series Auto Injector. Linearity was evaluated
with carfentanil citrate. A stock solution was prepared by accu-
rately weighing 1.05 mg of carfentanil citrate into a 15-mL cen-
trifuge tube and extracting from aqueous base with ether
(2  5 mL). The ether extracts were combined, evaporated to dry-
ness, and reconstituted to volume in a 10.0 mL volumetric flask
with CHCl3 containing 20 lg/mL of tetracosane. A set of serial dilu-
tions were made in order to provide 4 more solutions. One mL of
each linearity solution was evaporated to dryness and reconsti-
tuted in 100 mL of CHCl3 and placed into a 200 lL insert for injec-
tion. Linearity was evaluated and the method was shown to have
a linear response across the concentration range (5.25105.0 lg/
mL) where m = 0.019, b = 0.016, and r = 0.99994. Illicit samples
were accurately weighed (ca. 20 mg) into a 15-mL centrifuge tube,
dissolved in 1 mL of water, rendered basic with 2 drops of concen-
trated NH4OH, extracted with 5 mL of ether containing 1.00 mL of
CHCl3 (containing 20 lg/mL of tetracosane). The organic phase was
passed through a cotton-plugged pipet, evaporated to dryness and
reconstituted in 100 mL of CHCl3 for injection.

Fig. 3. Electron Ionization Mass Spectra of (a) carfentanil 1 and (b) acetylcarfentanil 2. Molecular ion or M-2 not observed for 2.
78 J.F. Casale et al. / Forensic Chemistry 3 (2017) 7480

Gas chromatography/mass spectrometry (GC/MS) was per-
formed using an Agilent Model 5975C quadrupole mass-selective
detector (MSD) interfaced with an Agilent 7890A gas
chromatograph. The MSD was operated in the electron ionization
mode with an ionization potential of 70 eV, a scan range of
34700 mass units, and at 1.34 scans/s. The GC system was fitted
with a 30 m  0.25 mm ID fused-silica capillary column coated
with DB-1 (0.25 mm) (Agilent, Santa Clara, CA) in constant flow
mode at 36.5 cm/s of helium. The GC oven was temperature pro-
grammed as follows: Initial temperature, 100 C; initial hold,
0.0 min; temperature program rate, 6 C/min to 300 C; final hold,
5.6 min. The injector was operated in the split mode (22:1) and at a
temperature of 280 C. The auxiliary transfer line to the MSD was
operated at 280 C. An approximately 5 mg/mL equivalent of
CH2Cl2 extract of carfentanil was injected at a volume of 2 mL.
Infrared spectra were obtained on a Thermo-Nicolet Nexus 670
FT-IR equipped with a SensIR Dura-Scope single bounce attenuated
total reflectance (ATR) accessory. Instrument parameters were:
Resolution = 4 cm 1; gain = 1; optical velocity = 0.4747; aper-
ture = 150; and scans/sample = 32.
NMR spectra were obtained using an Agilent 600MR-DD2
600 MHz NMR with a 5 mm OneNMR pulse field gradient probe

Fig. 4. Infrared Spectrum (FTIR-ATR) of (a) carfentanil citrate and (b) carfentanil hydrochloride.
 J.F. Casale et al. / Forensic Chemistry 3 (2017) 7480 79

Fig. 5. Proton spectra of (a) carfentanil citrate in 1.0 mL of CDCl3 containing 0.03% v/v tetramethylsilane (TMS, 0 ppm reference) and 2 drops of CD3OD and (b) carfentanil HCl
in CDCl3 containing 0.03% v/v tetramethylsilane. Insets for analyte peak groups are beneath each full spectrum. The letter i above a peak group indicates an impurity.

(Palo Alto, CA). The sample temperature was maintained at 25 C. A
standard Agilent proton pulse sequence was used (90 degree pulse,
45 s delay, 6 s acquisition time, 8 scans). Carfentanil citrate was
dissolved in 1 mL deuterated chloroform (CDCl3) containing
0.03% v/v tetramethylsilane (TMS, 0 ppm reference) with two
drops of deuterated methanol (CD3OD) to enhance solubility
(Cambridge Isotopes, Tewksbury, MA). 1,4-BTMSB-d4 (Wako Pure
Chemical Industries, Saitama, Japan) was used as the quantitative
internal standard. Carfentanil hydrochloride was dissolved in
deuterated chloroform (CDCl3) containing 0.03% v/v tetramethylsi-
lane (TMS, 0 ppm reference) (Cambridge Isotopes).
Carbon and nitrogen isotope ratio analyses were determined
using an elemental analyzer (EA) (Costech Analytical Technologies
Inc., Valencia, CA) interfaced with a Delta Plus XP isotope ratio
mass spectrometer (Thermo Fisher Scientific Inc., Bremen, Ger-
many). Typically, 0.81.1 mg of carfentanil was accurately weighed
into a tin capsule (Costech Analytical Technologies Inc.) and
introduced into the EA using a Costech Zero-Blank autosampler
80 J.F. Casale et al. / Forensic Chemistry 3 (2017) 7480
(Costech Analytical Technologies Inc.). The EA reactor tubes were
comprised of two quartz glass tubes filled with chromium (III)
oxide/silvered cobaltous oxide and reduced copper, held at
1040 C and 640 C for combustion and reduction, respectively. A
trap filled with magnesium perchlorate was used to remove water
from generated combustion gases, and a post-reactor GC column
was maintained at 65 C for separation of evolved N2 and CO2.
Helium (99.999% purity, ARC3 Gasses, Richmond, VA) was used
as the carrier gas, and the system head pressure was adjusted to
achieve a measured flow of 90 mL/min. Data was acquired and pro-
cessed using ISODAT 3.0 software (Thermo Fisher Scientific Inc.,
Bremen, Germany).
EA normalization and quality control materials consisted of
internally calibrated atropine (TCI, St. Louis, Missouri) and cocaine
base (DEA Laboratory, Dulles, Virginia). Sample sequences were
bracketed by an internally calibrated atropine secondary standard,
typically at intervals of one standard every seven samples. The
atropine and cocaine secondary standards were calibrated to pri-
mary isotopic standard materials IAEA-N-1, IAEA-N-2, USGS-25,
LSVEC, and NBS-19 and corresponding scale normalized isotope
values are expressed as d13CVPDB-LSVEC for carbon and d15NAIR for
nitrogen. Overall system reproducibility of reference materials
and replicate analyses of cocaine was consistently 0.1 and
0.2 or better for all EA-IRMS d13C and d15N measurements,
respectively.
4. Results and discussion
All three exhibits were examined by GC/MS to determine their
analytical trace profiles (Fig. 2). Each contained carfentanil (peak
#7) and at least one other fentanyl-related compound. In addition
to carfentanil, Case #1-1A was determined to contain traces of
diphenhydramine (peak #1), fentanyl (peak #5), and acetylcarfen-
tanil (peak #6) which produced a base peak at m/z 289 (Fig. 3b),
with no apparent molecular ion or M-2 ion observed due to the
low concentration of the compound. Acetylcarfentanil gave the
same relative retention time to carfentanil, similar to that of
acetylfentanyl to fentanyl. The spectrum of acetylcarfentanil vs.
carfentanil (Fig. 3a vs. b) was analogous to that of acetylfentanyl
vs. fentanyl, producing the same ions and relative abundances:
m/z 42, m/z 77, m/z 91, m/z 105, m/z 130, m/z 154, and m/z 187.
A mass difference of 14 Daltons was observed between acetylcar-
fentanil and carfentanil for ions found at m/z 229 vs. m/z 243, m/z
261 vs. m/z 275, m/z 289 vs. m/z 303, and m/z 321 vs. m/z 335, also
analogous to acetylfentanyl vs. fentanyl. Acetylcarfentanil presum-
ably arises from the clandestine synthesis of carfentanil, similar to
that of acetylfentanyl in illicit fentanyl exhibits. The observed frag-
mentation pattern was consistent with the MS/MS fragment struc-
tures identified by Feasel et al. [8]. Case #1-1B was determined to
contain traces of fentanyl (peak #5), acetylcarfentanil (peak #6),
and 2-furanylfentanyl (peak #8). Case #2 was determined to con-
tain diphenhydramine, heroin (Peak #4), fentanyl, and heroin-
related impurities. Direct FTIR-ATR spectra of the CH2Cl2 solubles
from both exhibits in Case #1 gave absorbances in the range of
23003000 cm 1, indicative of an amine hydrochloride ion-pair.
Their spectra were ultimately compared to a standard of carfen-
tanil hydrochloride (HCl) which was prepared from ethereal HCl
treatment of an ethereal basic extract of exhibit 1A (confirmed
via NMR). The salt form of Case #2 could not be spectroscopically
determined; however, the carfentanil was CH2Cl2 soluble, indicat-
ing it was not the citrate. FTIR-ATR spectra for carfentanil citrate
and carfentanil HCl are illustrated in Fig. 4. The proton spectra
for carfentanil citrate and carfentanil HCl are illustrated in Fig. 5
for reference purposes. Carfentanil citrate is easily differentiated
from the HCl via the two citric acid doublet resonances at ca.
2.7 ppm.
Each exhibit was quantitatively determined for carfentanil via
GC-FID. Cases 1-1A, 1-1B and 2 were determined to contain
0.62%, 1.87%, and 0.31% carfentanil as the HCl, respectively. The
methods limit of detection was approximately 2 lg/mL. We note
that the same quantitative values were obtained for each exhibit,
whether at a 1000 or 100 lL final dilution. The concentrations of
acetylcarfentanil in 1-1A and 1-1B were estimated to be approxi-
mately 0.001% and 0.004%, respectively, from the total ion current
via GC/MS.
Carfentanil was isolated and purified from both exhibits from
Case #1. Each was subjected to EA-IRMS analysis for nitrogen
and carbon isotope determination. Both exhibits gave virtually
identical measurements: d13C = 1.1 (0.1) and d15N = 27.6
(0.1) . Excessive heroin in Case #2 prevented the isolation of
carfentanil. Although not definitive, both exhibits in Case #1
appear to contain the same carfentanil or are intimately related,
although the GC/MS profiles contain different adulterants.
It is probably not a sheer coincidence that Exhibit 1A is exactly
one-third the carfentanil concentration of 1B, and that Case #2 is
exactly one-half the concentration of 1A. Although we were unable
to obtain sufficient data on Case #2, we contend that Cases #1 and
#2 may be related.
5. Addendum
Our laboratory has recently initiated a Fentanyl Profiling Pro-
gram which utilizes multiple signature methodologies patterned
after the DEAs Signature Programs. The objective is to enhance
current capabilities towards tactical and strategic intelligence for
the criminal justice system from a forensic chemistry perspective.
The current project will be the subject of a future communication.
Acknowledgements
The authors are indebted to Patrick Hays and David Morello of
this laboratory for their assistance in acquiring NMR and IRMS
data.
References
[1] D. Cooper, M. Jacob, A. Allen, Identification of fentanyl derivatives, J. Forensic
Sci. 31 (1986) 511528.
[2] T.C. Kram, D.A. Cooper, A.C. Allen, Behind the identification of China White,
Anal. Chem. 53 (1981) 1379A1386A.
[3] J.M. Moore, A.C. Allen, D.A. Cooper, S.M. Carr, Determination of fentanyl and
related compounds by capillary gas chromatography with electron capture
detection, Anal. Chem. 58 (1986) 16561660.
[4] I.S. Lurie, A.L. Berrier, J.F. Casale, R. Iio, J.S. Bozenko, Profiling of illicit fentanyl
using UHPLC-MS/MS, Forensic Sci. Int. 220 (2012) 191196.
[5] R. Frank. Rates of drug overdose deaths continue to rise, more action needed to
reverse troubling trends. U.S. Department of Health and Human Services. 2015
www.hhs.gov/blog/2015/12/10/rates-of-drug-overdose-deaths-continue-to-
rise.html> (Last accessed on 10-26-2016).
[6] R.M. Gladden, P. Martinez, P. Seth. Fentanyl law enforcement submissions and
increases in synthetic opioid-involved overdose deaths 27 states, 20132014.
Center for Disease Control and Prevention Morbidity and Mortality Weekly
Report. 65 (2016) 837885.
[7] J.R. Riches, R.W. Read, R.M. Black, N.J. Cooper, C.M. Timperley, Analysis of
clothing and urine from Moscow theatre siege casualties reveals carfentanil and
remifentanil use, J. Anal. Tox. 36 (2012) 647656.
[8] M.G. Feasel, A.W. Wohlfarth, J.M. Nilles, S. Pang, R.L. Hristovich, M.A. Huestis,
Metabolism of carfentanil, an ultra-potent opioid, in human liver microsomes
and human hepatocytes by high-resolution mass spectrometry, AAPS J. (2016),
http://dx.doi.org/10.1208/s12248-016-9963-5, Published online 05 August
2016.