On-site illicit-drug detection with an integrated near-infrared spectral sensor: A proof of concept

Ruben F. Kranenburg; Fang Ou; Petar Ševo; Maurangelo Petruzzella; Renee de Ridder; Anne van Klinken; Kaylee D. Hakkel; D.M.J. van Elst; P.J. van Veldhoven; Francesco Pagliano; Arian C. van Asten; Andrea Fiore

doi:10.1016/j.talanta.2022.123441

On-site illicit-drug detection with an integrated near-infrared spectral sensor: A proof of concept

Ruben F. Kranenburg (Corresponding author)
, Fang Ou
, Petar Ševo
, Maurangelo Petruzzella
, Renee de Ridder
, Anne van Klinken
, Kaylee D. Hakkel
, D.M.J. van Elst
, P.J. van Veldhoven
, Francesco Pagliano
, Arian C. van Asten
, Andrea Fiore

Research output: Contribution to journal › Article › Academic › peer-review

51 Citations (Scopus)

223 Downloads (Pure)

Abstract

Illicit-drug production, trafficking and seizures are on an all-time high. This consequently raises pressure on investigative authorities to provide rapid forensic results to assist law enforcement and legal processes in drug-related cases. Ideally, every police officer is equipped with a detector to reliably perform drug testing directly at the incident scene. Such a detector should preferably be small, portable, inexpensive and shock-resistant but should also provide sufficient selectivity to prevent erroneous identifications. This study explores the concept of on-site drugs-of-abuse detection using a 1.8 × 2.2 mm² multipixel near-infrared (NIR) spectral sensor that potentially can be integrated into a smartphone. This integrated sensor, based on an InGaAs-on-silicon technology, exploits an array of resonant-cavity enhanced photodetectors without any moving parts. A 100% correct classification of 11 common illicit drugs, pharmaceuticals and adulterants was achieved by chemometric modelling of the response of 15 wavelength-specific pixels. The performance on actual forensic casework was investigated on 246 cocaine-suspected powders and 39 MDMA-suspected ecstasy tablets yielding an over 90% correct classification in both cases. These findings show that presumptive drug testing by miniaturized spectral sensors is a promising development ultimately paving the way for a fully integrated drug-sensor in mobile communication devices used by law enforcement.

Original language	English
Article number	123441
Number of pages	8
Journal	Talanta
Volume	245
DOIs	https://doi.org/10.1016/j.talanta.2022.123441
Publication status	Published - 1 Aug 2022

Bibliographical note

Funding Information:
This research was partially funded under Penta Call 2 project Environmental Sensors for AIR Quality (ESAIRQ) n. 16113, NWO TTW project n. 16670 and NWO Zwaartekracht Research Center for Integrated Nanophotonics.

Funding Information:
For the measurements in each set, three preprocessing methods were applied to the triplicate-averaged absorbance values, and compared: mean-centering, sum normalization and standard normal variate (SNV). Six classifiers were compared for building the classification model: linear discriminant analysis (LDA), partial least square ? discriminant analysis (PLS-DA), support vector machine (SVM), principal component analysis (PCA)-LDA, random forest (RF) and PLS-RF. Fivefold groupwise cross-validation was used to optimize the parameters of the PLS, PCA and RF-based models, and groupwise randomized search cross-validation was used to optimize the SVM models that had a larger number of parameters to tune. In all cases of cross-validation, replicate measurements from each sample were kept together in one group. The combination of the preprocessing method and classifier that resulted in the best prediction performance for each experiment set is shown in this manuscript. The algorithms used in analysis and modeling were implemented in Python using packages from NumPy [50], Matplotlib [51], and Scikit-learn [52].This research was partially funded under Penta Call 2 project Environmental Sensors for AIR Quality (ESAIRQ) n. 16113, NWO TTW project n. 16670 and NWO Zwaartekracht Research Center for Integrated Nanophotonics.

Publisher Copyright:
© 2022 The Authors

Funding

This research was partially funded under Penta Call 2 project Environmental Sensors for AIR Quality (ESAIRQ) n. 16113, NWO TTW project n. 16670 and NWO Zwaartekracht Research Center for Integrated Nanophotonics. For the measurements in each set, three preprocessing methods were applied to the triplicate-averaged absorbance values, and compared: mean-centering, sum normalization and standard normal variate (SNV). Six classifiers were compared for building the classification model: linear discriminant analysis (LDA), partial least square ? discriminant analysis (PLS-DA), support vector machine (SVM), principal component analysis (PCA)-LDA, random forest (RF) and PLS-RF. Fivefold groupwise cross-validation was used to optimize the parameters of the PLS, PCA and RF-based models, and groupwise randomized search cross-validation was used to optimize the SVM models that had a larger number of parameters to tune. In all cases of cross-validation, replicate measurements from each sample were kept together in one group. The combination of the preprocessing method and classifier that resulted in the best prediction performance for each experiment set is shown in this manuscript. The algorithms used in analysis and modeling were implemented in Python using packages from NumPy [50], Matplotlib [51], and Scikit-learn [52].This research was partially funded under Penta Call 2 project Environmental Sensors for AIR Quality (ESAIRQ) n. 16113, NWO TTW project n. 16670 and NWO Zwaartekracht Research Center for Integrated Nanophotonics.

UN SDGs

This output contributes to the following UN Sustainable Development Goals (SDGs)

SDG 3 Good Health and Well-being

Keywords

Forensic on-scene analysis
Illicit-drug detection
Indicative testing
Integrated photonics
Portable devices
Spectral sensing
N-Methyl-3,4-methylenedioxyamphetamine
Cocaine
Substance Abuse Detection
Illicit Drugs
Smartphone

Access to Document

10.1016/j.talanta.2022.123441

pdfFinal published version, 3.22 MBLicence: CC BY

Cite this

Kranenburg, R. F., Ou, F., Ševo, P., Petruzzella, M., de Ridder, R., van Klinken, A., Hakkel, K. D., van Elst, D. M. J., van Veldhoven, P. J., Pagliano, F., van Asten, A. C., & Fiore, A. (2022). On-site illicit-drug detection with an integrated near-infrared spectral sensor: A proof of concept. Talanta, 245, Article 123441. https://doi.org/10.1016/j.talanta.2022.123441

@article{d9b634d95c3e4911be9c68641d6c3ce8,

title = "On-site illicit-drug detection with an integrated near-infrared spectral sensor: A proof of concept",

abstract = "Illicit-drug production, trafficking and seizures are on an all-time high. This consequently raises pressure on investigative authorities to provide rapid forensic results to assist law enforcement and legal processes in drug-related cases. Ideally, every police officer is equipped with a detector to reliably perform drug testing directly at the incident scene. Such a detector should preferably be small, portable, inexpensive and shock-resistant but should also provide sufficient selectivity to prevent erroneous identifications. This study explores the concept of on-site drugs-of-abuse detection using a 1.8 × 2.2 mm2 multipixel near-infrared (NIR) spectral sensor that potentially can be integrated into a smartphone. This integrated sensor, based on an InGaAs-on-silicon technology, exploits an array of resonant-cavity enhanced photodetectors without any moving parts. A 100\% correct classification of 11 common illicit drugs, pharmaceuticals and adulterants was achieved by chemometric modelling of the response of 15 wavelength-specific pixels. The performance on actual forensic casework was investigated on 246 cocaine-suspected powders and 39 MDMA-suspected ecstasy tablets yielding an over 90\% correct classification in both cases. These findings show that presumptive drug testing by miniaturized spectral sensors is a promising development ultimately paving the way for a fully integrated drug-sensor in mobile communication devices used by law enforcement.",

keywords = "Forensic on-scene analysis, Illicit-drug detection, Indicative testing, Integrated photonics, Portable devices, Spectral sensing, N-Methyl-3,4-methylenedioxyamphetamine, Cocaine, Substance Abuse Detection, Illicit Drugs, Smartphone",

author = "Kranenburg, \{Ruben F.\} and Fang Ou and Petar {\v S}evo and Maurangelo Petruzzella and \{de Ridder\}, Renee and \{van Klinken\}, Anne and Hakkel, \{Kaylee D.\} and \{van Elst\}, D.M.J. and \{van Veldhoven\}, P.J. and Francesco Pagliano and \{van Asten\}, \{Arian C.\} and Andrea Fiore",

note = "Funding Information: This research was partially funded under Penta Call 2 project Environmental Sensors for AIR Quality (ESAIRQ) n. 16113, NWO TTW project n. 16670 and NWO Zwaartekracht Research Center for Integrated Nanophotonics. Funding Information: For the measurements in each set, three preprocessing methods were applied to the triplicate-averaged absorbance values, and compared: mean-centering, sum normalization and standard normal variate (SNV). Six classifiers were compared for building the classification model: linear discriminant analysis (LDA), partial least square ? discriminant analysis (PLS-DA), support vector machine (SVM), principal component analysis (PCA)-LDA, random forest (RF) and PLS-RF. Fivefold groupwise cross-validation was used to optimize the parameters of the PLS, PCA and RF-based models, and groupwise randomized search cross-validation was used to optimize the SVM models that had a larger number of parameters to tune. In all cases of cross-validation, replicate measurements from each sample were kept together in one group. The combination of the preprocessing method and classifier that resulted in the best prediction performance for each experiment set is shown in this manuscript. The algorithms used in analysis and modeling were implemented in Python using packages from NumPy [50], Matplotlib [51], and Scikit-learn [52].This research was partially funded under Penta Call 2 project Environmental Sensors for AIR Quality (ESAIRQ) n. 16113, NWO TTW project n. 16670 and NWO Zwaartekracht Research Center for Integrated Nanophotonics. Publisher Copyright: {\textcopyright} 2022 The Authors",

year = "2022",

month = aug,

day = "1",

doi = "10.1016/j.talanta.2022.123441",

language = "English",

volume = "245",

journal = "Talanta",

issn = "0039-9140",

publisher = "Elsevier B.V.",

}

TY - JOUR

T1 - On-site illicit-drug detection with an integrated near-infrared spectral sensor

T2 - A proof of concept

AU - Kranenburg, Ruben F.

AU - Ou, Fang

AU - Ševo, Petar

AU - Petruzzella, Maurangelo

AU - de Ridder, Renee

AU - van Klinken, Anne

AU - Hakkel, Kaylee D.

AU - van Elst, D.M.J.

AU - van Veldhoven, P.J.

AU - Pagliano, Francesco

AU - van Asten, Arian C.

AU - Fiore, Andrea

N1 - Funding Information: This research was partially funded under Penta Call 2 project Environmental Sensors for AIR Quality (ESAIRQ) n. 16113, NWO TTW project n. 16670 and NWO Zwaartekracht Research Center for Integrated Nanophotonics. Funding Information: For the measurements in each set, three preprocessing methods were applied to the triplicate-averaged absorbance values, and compared: mean-centering, sum normalization and standard normal variate (SNV). Six classifiers were compared for building the classification model: linear discriminant analysis (LDA), partial least square ? discriminant analysis (PLS-DA), support vector machine (SVM), principal component analysis (PCA)-LDA, random forest (RF) and PLS-RF. Fivefold groupwise cross-validation was used to optimize the parameters of the PLS, PCA and RF-based models, and groupwise randomized search cross-validation was used to optimize the SVM models that had a larger number of parameters to tune. In all cases of cross-validation, replicate measurements from each sample were kept together in one group. The combination of the preprocessing method and classifier that resulted in the best prediction performance for each experiment set is shown in this manuscript. The algorithms used in analysis and modeling were implemented in Python using packages from NumPy [50], Matplotlib [51], and Scikit-learn [52].This research was partially funded under Penta Call 2 project Environmental Sensors for AIR Quality (ESAIRQ) n. 16113, NWO TTW project n. 16670 and NWO Zwaartekracht Research Center for Integrated Nanophotonics. Publisher Copyright: © 2022 The Authors

PY - 2022/8/1

Y1 - 2022/8/1

N2 - Illicit-drug production, trafficking and seizures are on an all-time high. This consequently raises pressure on investigative authorities to provide rapid forensic results to assist law enforcement and legal processes in drug-related cases. Ideally, every police officer is equipped with a detector to reliably perform drug testing directly at the incident scene. Such a detector should preferably be small, portable, inexpensive and shock-resistant but should also provide sufficient selectivity to prevent erroneous identifications. This study explores the concept of on-site drugs-of-abuse detection using a 1.8 × 2.2 mm2 multipixel near-infrared (NIR) spectral sensor that potentially can be integrated into a smartphone. This integrated sensor, based on an InGaAs-on-silicon technology, exploits an array of resonant-cavity enhanced photodetectors without any moving parts. A 100% correct classification of 11 common illicit drugs, pharmaceuticals and adulterants was achieved by chemometric modelling of the response of 15 wavelength-specific pixels. The performance on actual forensic casework was investigated on 246 cocaine-suspected powders and 39 MDMA-suspected ecstasy tablets yielding an over 90% correct classification in both cases. These findings show that presumptive drug testing by miniaturized spectral sensors is a promising development ultimately paving the way for a fully integrated drug-sensor in mobile communication devices used by law enforcement.

AB - Illicit-drug production, trafficking and seizures are on an all-time high. This consequently raises pressure on investigative authorities to provide rapid forensic results to assist law enforcement and legal processes in drug-related cases. Ideally, every police officer is equipped with a detector to reliably perform drug testing directly at the incident scene. Such a detector should preferably be small, portable, inexpensive and shock-resistant but should also provide sufficient selectivity to prevent erroneous identifications. This study explores the concept of on-site drugs-of-abuse detection using a 1.8 × 2.2 mm2 multipixel near-infrared (NIR) spectral sensor that potentially can be integrated into a smartphone. This integrated sensor, based on an InGaAs-on-silicon technology, exploits an array of resonant-cavity enhanced photodetectors without any moving parts. A 100% correct classification of 11 common illicit drugs, pharmaceuticals and adulterants was achieved by chemometric modelling of the response of 15 wavelength-specific pixels. The performance on actual forensic casework was investigated on 246 cocaine-suspected powders and 39 MDMA-suspected ecstasy tablets yielding an over 90% correct classification in both cases. These findings show that presumptive drug testing by miniaturized spectral sensors is a promising development ultimately paving the way for a fully integrated drug-sensor in mobile communication devices used by law enforcement.

KW - Forensic on-scene analysis

KW - Illicit-drug detection

KW - Indicative testing

KW - Integrated photonics

KW - Portable devices

KW - Spectral sensing

KW - N-Methyl-3,4-methylenedioxyamphetamine

KW - Cocaine

KW - Substance Abuse Detection

KW - Illicit Drugs

KW - Smartphone

UR - https://www.scopus.com/pages/publications/85127700063

U2 - 10.1016/j.talanta.2022.123441

DO - 10.1016/j.talanta.2022.123441

M3 - Article

C2 - 35405444

AN - SCOPUS:85127700063

SN - 0039-9140

VL - 245

JO - Talanta

JF - Talanta

M1 - 123441

ER -

On-site illicit-drug detection with an integrated near-infrared spectral sensor: A proof of concept

Abstract

Bibliographical note

Funding

UN SDGs

Keywords

Access to Document

Other files and links

Fingerprint

Cite this