Skip to main content

Drugs and Precursor Sensing by Complementing Low Cost Multiple Techniques

Final Report Summary - CUSTOM (Drugs and Precursor Sensing by Complementing Low Cost Multiple Techniques)

Executive Summary:
The result of the project CUSTOM (Drugs and Precursors Sensing by Complementing Low Cost Multiple Techniques, FP7-SEC-2009.1.3-028) is the realization of a chemical detection system able to detect traces of drug precursors, such as ephedrine, safrole, acetic anhydride, Benzyl Methyl Ketone (BMK).
The system is modular and composed by two complementary sensing techniques, a micro-fluidic circuit and a pre-concentrator unit (PCU) to increase concentration and improve sensitivity of detection and reduce false alarm, a microprocessor including advanced data processing algorithms and database.
The system is equipped with a machine/user interface with a easy to use touch screen display.

All the modules are managed by a master control board interfacing all the modules and managing power supply.
The system implements two complimentary sensing techniques: 1) a fast response high sensitive spectroscopic technique based on photoacoustic transduction of IR absorption, able to drastically increase SNR respect to standard spectroscopy, and 2) highly selective fluorescence enhanced by the use of specially developed Organic macro-molecules able to bind to the analytes with specific body antibody interaction.
An operational demonstration has been performed to validate the system, obtaining a probability of detection of 98% and limit of detection below the ppm range.

Project Context and Objectives:
Drug Precursors are chemical substances having wide licit uses (e.g. in the manufacturing of pharmaceutical products, perfumes, cosmetics, fertilizers, oils, etc.). However, they can be extremely dangerous when diverted from the licit channels for the illicit manufacturing of drugs (such as heroin, cocaine, ecstasy, amphetamines, etc.). For example 200g of piperonal (a substance commonly used in perfumery, in cherry and vanilla flavorings, in organic synthesis, and in the manufacturing of mosquito repellents) are enough to produce 4,000 street doses of ecstasy.
A list of the drug precursors chemicals according to the article 12 of the 1988 United Nations Convention against the Illicit Traffic in Narcotic Drugs and Psychotropic Substances is reported in Table 1. The 23 listed substances are referred as essentials (Table I), if they are used to refine and process plant-based drugs chemicals and can be readily replaced by other chemicals with similar properties, or as precursors (Table II), if they are used as starting chemicals inputs for the production of synthetic drugs and are less likely to be substituted.

Table 1 -Tables I and II of the United Nations Convention, 1988
The need for a portable sensing platform able to perform chemical detection and identification over a large number of compounds is a fundamental requirement from the law enforcement agencies to control the manufacturing and the distribution of illegal narcotics and synthetic substances worldwide.
The project CUSTOM focuses on the development of a system for the detection of drug precursors for customs, airports and harbors check point scenarios, where inspection of trucks, cars, containers, as well as people and baggage is required.
A large number of different non-invasive hand held chemical sensors available on the market have been developed in the last decades. All of them are mainly based on a single detection approach, whether: Ion Mobility Spectrometry (IMS), Surface Acoustic Wave (SAW), Gas Chromatography or Raman Spectroscopy.
In order to achieve an optimum trade-off between opposite requirements in terms of high probability of detection and low false alarm rate, the sensor architecture has been conceived to integrate a multiple sensing techniques, which operate in a complimentary approach: the Biochemical Fluorescence (BF) and the Laser Photo Acoustic Spectroscopy. A further improvement of the sensor selectivity is achieved by implementing a chemical Pre-Concentrator (PC), based on the sorption/desorption cycles of a Syndiotactic Polystyrene (SP) polymer.

Project Results:
The general concept of the sensor architecture is schematically shown in Figure 1.
A micro-pump delivers the air samples to the Laser Photo Acoustic Spectrometer (LPAS) or to the Biochemical Fluorescence (BF) sensor depending on the selected operating mode.
The LPAS and the BF can be fed directly or through the PC unit.

Figure 1 – Schematic view of the sensor architecture and hydraulic connections between the Syndiotactic Polystyrene Preconcentrator (SP-SPC), the Biochemical Fluorescence (BF) module and the Laser Photo Acoustic Spectrometer (LPAS)
The sensor has two main operational modes:
• A first alarm response, where the LPAS module is used for a fast screening of the gas samples;
• A confirmation results, where the BF module is used to confirm the presence of target analyte.
The two operational modes are complimentary and are used in sequence. As the LPAS is generally faster than the BF, it is used in the “high POD” mode to make a rapid check of the presence of the target analyte at first. Whenever the probability of the target molecule presence is above a fixed threshold, the operative mode is changed to the more selective and accurate low FA rate mode, which makes use of the BF.
Depending on the field of application, the PC could be used to increase the sensitivity of the sensor, or bypassed to reduce the air sapling time and the power consuming. An electronic control board is deputed to manage all the hardware modules of the sensor (PC, BF, LPAS), as well as the data processing used to elaborate the spectra acquired by the LPAS and perform chemical recognition, the communication functions between each modules of the sensor, the user interfaces and the power supply.
A picture of the final prototype is reported in Figure 2. The complete operational features are accessible by the touch screen panel, the vapour samples are collected by an inlet funnel.

Figure 2 – Final prototype. The bare hardware system is represented (left) and the system mounted inside its own the vessel (right). Complete operational features are accessible by the touch screen control panel and the inlet funnel
Hardware and software control - The Master Control Board (MCB) is the central hardware that controls, interconnects and coordinates the overall system.
The MCB is based on an ARM cortex-M3 microcontroller: a purposely developed firmware is burned into the chip ROM flash and a Unix like real-time operating system runs. The microprocessor coordinated action upon the devices and sensors allows for data acquisition, processing and results presentation under control of the User.
The final version of the board includes control for the LPAS and FLUO sensors, the Pre-Concentrator unit, the air sampler and hydraulic circuit. User actions are entered via a TFT display panel with touch capability (7”); chain of custody requirements are assured by permanent storage of relevant data including a time and geolocation (GPS device) (see Figure 3).

Figure 3: screenshot of the TFT display showing the panel for the mode option setting: First alarm (LPAS) end Fine Screening (FLUO). LPAS is run first providing Fast analysis. If the response is above a probability of presence threshold, a further analysis is performed on the sample by means of FLUO, which is intrinsically selective.
A serial diagnostic interface (USB) is available as well. Performances assessment on the hardware and software control concern the overall robustness and capability to properly manage critical situation, like malfunctioning and or partial failures of connected devices.
In normal operation the Users is only allowed to interact with the touch panel display; no special commands or settings are required; on the other hand extended local diagnostic allows a skilled User or servicing personnel to make a thorough tests of every elementary hardware functionality implemented by the firmware. The activation of diagnostic mode requires use of the USB serial communication.
Laser Photoacoustic spectrometer - Among various detection schemes, the LASER Photo Acoustic Spectroscopy (LPAS) offers some unique features such as low sample volume, high sensitivity and wide linear dynamic range. Signal is generated only from the absorption thus being a so called zero-background technique .
The prototype of the LPAS sensor Figure 4) is a 320 mm x 141 mm x 184 mm module, which includes a quantum cascade laser, a photoacoustic cell and a readout interferometer. It is enclosed into a vibrations-isolated baseplate integrating the electronics and the thermal control system.

Figure 4 – Layout of the LPAS module
The LPAS module performance is verified by means Gasera standard test. These include the noise level test of the interferometer, the PA-cell leakage test and the signal tests with a DFB laser measuring 100 ppm of acetylene.
The PA-cell leakage test is performed with an external pump, pressure meter and individually controlling the cell valves so that the final result is the cell leakage in mbar’s over defined time period. The air tightness of the PA-cell is a critical factor as the pressure inside the photoacoustic cell affects the spring constant of the cantilever pressure sensor and thus different measurement pressures affect the signal response. The pressure increase of the current LPAS module was 2 mbar in 10 seconds. A change of 10% in the pressure typically causes a response change of 1%, so this is acceptable considering the repeatability during 2 minutes of measurement time.
Concerning the Noise level test, the frequency response of the cantilever is recorded by collecting the PA-signal with cell windows blocked for at least 10 seconds and Fourier Transformed to frequency domain (Figure 5).

Figure 5– Noise level measure of LPAS sensor
The acceptable limits, that indicate the proper functioning of the cantilever and the interferometer is to have the lowest noise level< 50 µV (typically around 300 Hz) and the first cantilever resonance peak of value> 200 µV. In the figure 2.12 (see below), the lowest noise level is about 35 µV and the cantilever resonance peak is 300 µV. These indicate that the interferometer module is tuned to the cantilever correctly, and the vibration insulation is working as expected.
As to the Signal Test, a modulated DFB laser is used to measure the acoustic signal with the PA-cell filled with acetylene in Nitrogen and to measure the acoustic background with the PA-cell filled with pure Nitrogen. From these measurements a signal-to-noise-ratio is calculated and the detection limit is obtained by dividing the sample gas concentration by the SNR (2*stdev). Results are shown in table below.
Table 2– SNR and detection limit assessment
Signal (C2H2 100ppm) Noise (N2) 1 s (2xrms) SNR Det. limit [ppb]
LPAS module (w/DFB laser) 353,1 0,077 4593 23 ppb
The laser used for the prototype is a room temperature monomodal tunable External Cavity Quantum Cascade Laser (EC-QCL) delivering 2.5 mW average output power (pulsed with a duty cycle of 3%) in the MID-IR (1100 cm-1) range with 50 cm-1 of wavenumber tunability and line width 0.5cm-1.
The EC-QCL spectra for different angles of the grating is shown in Figure 6a. The tunability of the ECQCL is determined by measuring the side mode suppression ratio (SMSR). For a SMSR higher than 15dB, the laser can be considered monomode (50 cm-1) Figure 6b. The linear behaviour of the wavelength versus the angle of the grating is reported in Figure 6b.
All the characterizations have been measured at room temperature. This is very important because liquid nitrogen is not necessary to cool the laser down. The Peltier stage, fans and the heat sink integrated in the LPAS system are adapted to the regulation of the laser

Figure 6- Spectra of the ECQCL for different angle of the grating (up). Side mode suppression ratio versus the wavenumber (down left). Evolution of the wavelength versus the angle of the grating (down right)
First tests using the LPAS prototype have been conducted on the AcAn and BMK (Figure 7). The estimation of the detection limit using univariate analysis is 500 ppb and 2ppm respectively for BMK and Acetic Anhydride.

Figure 7 – Examples of BMK (left), Acetic anidride (right) experimental LPAS spectra
In order to increase the detection capability against the target analyte, the acquired spectra coming from the LPAS module are processed and elaborated using the Partial Least Squares - Discriminant Analysis (PLS-DA) method. For a given set of wavenumbers, this multivariate method calculates one model for each drug precursor, giving its probability of presence. The advantages offered by the multivariate method consist in the fact that it does not only consider the useful (discriminant) information brought by each variable (wavenumber) separately, but also the information deriving from the interactions among all the analyzed variables.
A complex strategy of spectral response simulation has been developed in order to select the optimal wavelengths range to efficiently detect the drug precursor molecules in presence of interfering species and background air. For this purpose spectra of gases from literature databases have been collected, denoised by means of the Wavelet Transform and mixed together according to a concentration matrix. The optimal range has been defined by maximizing the classification efficiency function.
937 LPAS spectra, acquired in the measurement sessions performed between November 2013 and February 2014, were considered for the development and validation of multivariate classification models through the PLS-DA algorithm.
Due to a limitation on the performance of the actual QCL employed for spectra acquisition, the final PLS-DA classification models were mainly optimized for the detection of acetic anhydride.
On the whole, 394 LPAS spectra were included in the training set and 543 LPAS spectra were included in the external test set. Furthermore, in order to check the possible effect of the presence of a wider number of pollutants in a wide set of complex gas mixtures, the calculated model was also tested on another external test set of 1000 simulated spectra, including 9 air components, 4 target molecules and 20 pollutants. A 1 latent variable PLS-DA classification model was selected by means of random groups cross-validation, which led to a satisfactory performance as for the identification of acetic anhydride based on the LPAS spectra. The performance of this classification model, estimated in terms of Sensitivity (SENS), Specificity (SPEC) and Classification Efficiency is reported in Table 2.
Figure 8 reports the actual values of the probability of presence of acetic anhydride for the dataset.
The LPAS spectra measured on samples containing acetic anhydride are indicated as red diamonds, the LPAS spectra measured on samples without acetic anhydride are indicated as green squares, while the LPAS spectra not assigned to any class are indicated as grey circles. All the samples lying above the red dashed threshold line (corresponding to the probability of presence of acetic anhydride equal to 50%) are classified as samples with acetic anhydride.
SENS SPEC EFF
LPAS spectra training set (calibration results) 99.5% 100.0% 99.8%
LPAS spectra training set (cross-validation) 99.5% 99.9% 99.7%
LPAS spectra external test set 98.0% 94.1% 96.0%
Simulated spectra external test set 88.8% 94.1% 91.4%
Table 2 – PLS-DA classification results for the detection of acetic anhydride


Figure 8 - Predicted probability values for the presence of acetic anhydride based on the LPAS spectra of the external test set samples
A cantilever-type pressure sensor acting as a microphone is designed to improve the sensitivity of PAS . It is made out of single crystal SOI-silicon with a specially developed dry-etching process that leads to a highly stable component; this is why the sensor is practically totally immune to temperature and humidity variations. In addition, the sensor is not suffering from wearing.
The displacement of the cantilever is measured by a spatial type optical interferometer. It avoids the so called "breathing effect of the pressure sensor readout, which occurs in capacitive measurement principle, where the other electrode damps the movement of the sensor.
The EC-QCL is located on the same vibration damped baseplate as the photoacoustic cell, which is a 95 mm long thermally stabilized gold coated cylinder having 4 mm diameter.
Biochemical Fluorescence based Module - The biochemical fluorescent (BF) module consists on an immunochemical competitive assay 0 combined with an optical readout method. The proof of concept of the BF sensor focuses on the detection of the 1-Phenyl-2-propanone (BMK).
The BF module exploits an immunochemical approach to bind selectively the target analyte and the Polarization Fluorescence (PF) method to transduce the chemical concentration as a variation of fluorescence intensity .
In Particular a new chemically synthesized BMK derivative covalently attached to an immunological carrier was used for producing antibodies against the BMK molecules. A fluorescence polarization-based bioassay was developed by using the produced anti-BMK antibodies and the BMK derivative.
The assay exhibits interesting analytical performances with a limit of detection of less than 100 nM and an almost linear response up to 600 nM The target compound contained in the sample competes to bind a specific antibody with a derivative of the target compound.

One of the main task for the realization of the fluorescence sensor is the production of the reagents for the detection of the BMK, namely BMK antibodies and their derivatives. The FP immunoassay was used to measure the competition between the tracer of unlabeled BMK in solution and BMK– BSA-488 for binding with specific antibodies anti-BMK.
Different samples with a fixed concentration of antibody (460 nM) were incubated with increasing concentration of BMK in the range of 0.2–1.0 mM. Each sample was mixed off-line and allowed to incubate for 30 minutes before the fluorescence polarization measurements.
Figure 9 shows the decrease of polarized fluorescence emission as a consequence of increase of unlabeled BMK in solution. By analysing the data shown in the Figure 5, it appears that with this method it is possible to detect amounts of BMK less than 50 nM.
The Biochemical Fluorescent (BF) hardware (Figure 6) consists on an analysis module, the sample interface and a reagents handling module. The latter comprises the fluidic system and some printed circuit boards together with the local control board. The control board contains a tiny microcontroller to acquire some process signals like temperature and reservoirs pressure. Besides, it controls the activation of the valves and the peristaltic pump in order to inject the required compounds depending on the sample target to be detected. Moreover, it enables the light source to excite the optochip and it receives the emitted signal of the sample after suitable conversions.
Providing the request of the Main Control Board (MCB) to detect some specific target, the local controller fills up a 450l cuvette with the suitable quantity of required biomolecules and reagents. Afterwards, the local tiny microcontroller requests to the MCB the activation of the air pump to bubble the sample target passing through the preconcentrator if it were needed. After the mixture fills up the measuring cell, the local microcontroller switches on the light source exciting the optochip with a specific wavelength depending on the sample target.

Figure 9 – Titration of FP immune-assay with increasing concentration of unlabeled BMK

Figure 10 – Layout of the BF module. It consists of an optical source and a photomultiplier based detector, selectable UV filters and collimation lenses for the fluorescence analysis, a handling module to manage liquids and the electronic boards which control and monitor its status
At the same time the detector, a Photo Multiplier Tube, will capture the specific wavelength light emitted by the optochip blocking other wavelengths by means of a band pass filter. After measuring, the fluidic system is cleaned to get ready for a new measurement.
In the case of the BMK detection, BMK labeled antibody is excited with a 100Hz pulsed light source of 535nm and the emission from the sample is blocked with a band pass filter of 671nm.
The acquired raw data and some calculated mean values of the received light as voltage values are sent by the local microcontroller to the MCB through a Universal Asynchronous Receiver/Transmitter (UART) controller. The mean value is introduced into a calibrated curve in order to determine the concentration of the target.
Extensive laboratory tests have been performed to assess the FLUO sensor performances, to optimize the procedure of gas exchange between the air sampler and the measuring cell; and to find the optimal physical conditions for cell loading and for cleaning procedure. Some problems still are unsolved concerning the reagent loading procedure, then measurements can be performed only in semiautomatic manner.
Table 3 - competition polarization assays used for FLUO assessment

Sample Description
Blank antibody against BMK (460 nM) in absence of the unlabeled BMK with a solution of ethanol/PBS 30:70 (150uL); used as reference;
BMK 600 nM antibody against BMK (460 nM) in presence of solution of BMK diluted in a solution of ethanol/PBS 30:70 (150uL) with a final concentration of 600 nM
BMK fluxed in cuvette antibody against BMK (460 nM) in presence of solution of ethanol/PBS 30:70 (150 uL) in which was bubbled BMK directly in cuvette
BMK fluxed in eppendorf tube antibody against BMK (460 nM) in presence of solution of ethanol/PBS 30:70 (150 uL) in which was bubbled BMK directly in eppendorf tube
The FLUO assessment dealt with laboratory measurements performed by fluxing BMK directly in a solution of ethanol/PBS 30:70 in the cuvette and in the eppendorf tube (plastic vial); successively manual addition of the Antibody and BMK–BSA-488 were done. The final volume of the assay was 450 l and in the bubbled samples the gas mixture (BMK/Nitrogen) has been fluxed in 150l solution at 8ml/min for 1 minutes (it is the same condition designed also for the hydraulic circuit of FLUO module).
In particular, four samples of the competition polarization assays were used as detailed in Table 3.
The incubation for the samples was done at room temperature for 1 hour. After pre-incubation of these solutions in presence of antibody against BMK, all samples were incubated with BMK–BSA-488 (70 nM) for 1 hour at room temperature. After this last incubation period the polarization fluorescence measurements were carried out with Fluo module with a vertical excitation polarized filter and with a horizontal emission polarized filter.
Results are reported below in the Figure 12-Figure 14.

Figure 11- Fluorescence polarization emission of Blank sample (0 nM BMK) and in presence of 600 nM of BMK, acquired with MiniFluo Figure 12 - Fluorescence polarization emission of Blank sample (0 nM BMK) and sample bubbled with BMK in cuvette, acquired with MiniFluo. Figure 13 - Fluorescence polarization emission of Blank sample (0nM BMK) and sample bubbled with BMK in eppendorf tube, acquired with MiniFluo.
The results showed above demonstrated that the polarization developed fluorescence is suitable method for the detection of drug precursor. The methodology to detect Ephedrine and BMK based on inmuno competitive assays has been setup and validated in laboratory.
The reagents to perform Ephedrine and BMK detection tests have been designed and manufactured; it is worth notice that these reagents are not available in the market, and they have been produced specifically.
Sensitivity of 40 ppb is demonstrated with BMK in solution in semiautomatic operating mode and the specificity has been also assessed. Concentration of 1 ppm of BMK have been detected in vapor state.
Air Sampler and Pre Cconcentrator Unit - The Air Sampler Unit (ASU) and the Pre Concentrator Unit (PCU) provide for feeding the sensors with a sample of air to analyze, eventually concentrating the amount of analyte in order to increase the system sensitivity while not impairing the discrimination capability. The ASU and PCU are electrically and mechanically integrated in an appropriate slot on the system mechanical frame.
The air sampler is equipped by a membrane air pump and two transducers for measuring the pressure inside the hydraulic circuit and for measuring the air flux.

The cleaning efficiency of the air sampler has been tested by evaluating the signal decrease as function of time for two different sample constituents, namely the water vapor and the ammonia. Gas mixture were prepared as follow:
• Water vapor (from air laboratory at RH 60% T=22C);
• Ammonia NH3 (30% in H2O solution) (Ftot/Fsample =300/250).
Cleaning procedure was tested by acquiring subsequent LPAS spectra with an interleave of 30s, and the results are shown in Figure 14 – Air sampler cleaning efficiency evaluated on LPAS sensor. The acquired experimental data is the decrease of the spectral peak amplitude as function of time.

Figure 14 – Air sampler cleaning efficiency evaluated on LPAS sensor
The cleaning efficiency can be related to the signal amplitude reduction; a decrease of more than one order of magnitude (-10dB) is reached in around 30s flushing time. Extrapolating the experimental data an average reduction of at least -30dB is reached in 120s.
Previous result enabled the MCB to perform a smart cleaning process, by considering an average peak reduction of 0.27db/s and relating the flushing time to the amplitude of the last measured LPAS peak. By this approach the cleaning time is not less than 30s and its duration will extend till when the signal amplitude is equal or falls below the noise level of the LPAS sensor.
The Pre-Concentrator (PC) is based on a nanoporous polymer material, the syndiotactic polystyrene (sPS), which is in form of aerogels beads with features of 500-1000 microns. It is used for its high sorption capacity and fast sorption/desorption kinetics with Volatile Organic Compounds (VOCs) .

Figure 15 – Preconcentrator layout
The preconcentrator consists on a metallic serpentine path of 500 mm length and of 2 cm3 of volume (Figure 15) filled with about 300mg of sPs. It has 90% porosity and so very low density: of about 0.1 g/cm3.
If thermal cycles are applied upon the PC, sPS will adsorb and concentrate VOC at low temperature (-10°C), during air sampling, and it releases them at higher temperature (e.g. +50°C).
Preliminary experiments conducted on the PC unit by using acetic anhydride demonstrated an improvement of a factor 13 of the concentration (from 5ppm to 65ppm) of the analyte in the sampled air.
In order to verify the PCU operation, sorption and desorption kinetics of acetic anhydride (AcAn) have been tested. In figure 2.3 the amount of AcAn absorbed by the PCU vs sorption time is reported, and the AcAn concentration in the inlet air into the PCU has also been indicated for each curve.

a) b) c)
Figure 16 - Amount of AcAn absorbed by the polymer vs the sorption time a). For each curve the AcAn concentration of inlet air has been indicated. Sorption kinetic curves of AcAn for each concentration tested (b), deorption kinetic curves of AcAn for each concentration tested (c).
In Figure 16 the sorption and desorption kinetic curves are reported in normalized form, i.e. Mt/M∞ versus time (the subscripts refer to mass of AcAn evaluated at time t and at sorption equilibrium ∞).
Sorption and desorption kinetics depend on temperature and air flow, but do not depend on the AcAn concentration in the inlet air, as shown by the kinetic curves. Sorption is temperature and air flow dependent, in particular, decreasing temperature and increasing the inlet air flow rate the amount and the kinetic of absorbed AcAn are enhanced. Desorption is instead faster by increasing the temperature and decreasing the outlet air flow rate.
For each tested concentration of AcAn in the inlet air, the total sorption of AcAn happens in less than 16 minutes, and the complete desorption in approximately 5 minutes. However, the most part of AcAn uptake (>80%) happens in less than 6 minutes, as well as the most part (>80%) of AcAn desorbs in about 1 minute.
Shorter sorption/desorption cycles allow to absorb/desorb significant amounts of AcAn, compared to the air volume contained into the PCU. As an example, for the AcAn concentration in the inlet air of 5ppm, after a sorption/desorption cycle, which takes 1 sorption minute and 1 desorption minute, the amount of AcAn released in the air is around 50 µg.

The use of sPS based pre-concentrator hence improves the sensitivity of the whole sensors as well as its detection limit and selectivity.
All the components described above have been assembled and integrated all together to form the final Demonstrator, which is a result of a multidisciplinary fruitful Team Working between Partners of complementary Skills.
The final stage of the project was devoted to the test and the assessment of the demonstrator.
After several tests conducted in a laboratory environment successfully demonstrated the ability of the CUSTOM sensor to be used as a fast and reliable measuring unit for the detection of illicit substances.
Real scenario was simulated by introducing in a carton box a sealed vial containing acetic anhydride; the CUSTOM sensor was easily deployed, and showed the capability to positively detect trace of the illicit substance and to give an alarm to the User.

Potential Impact:
The natural domain of utilization of the final prototype realized in CUSTOM is the detection of drug precursors that is performed on-field by customs, law enforcement authorities and local units engaged in the fight against illicit drugs.
The main applications of the system is the on-field check of suspected substances in the following cases:
• control of incoming and outgoing freight/cargo at the airport, seaport and border;
• passenger control on import, export and transit at the airport and seaport;
• thorough check of vehicles;
• physical search for suspected persons;
• controls effected at courier bond, mail posts, and importers premises;
• control on the acquisition of chemicals used in illicit laboratories.

The main impact of the system will be in the
• Security: for the monitoring and control of the drug precursors manufacture and transportation in contests like Customs, Airports, car and truck checks, people, baggage and cargos inspections etc, where fast and accurate responses are needed by the law enforcement operators for the seizure of the drugs.
• Forensic: output data including sample spectra, GPS location, operator ID, environmental conditions can exported and stored with chain of custody integrity by means of digital signature for legal/forensic needs.

An extension of the target analytes to the toxic industrial material and the chemical warfare agents makes the innovation suitable for:
• military application: military brigades for battlefield monitoring and decontamination,
• civil application for environmental monitoring and industrial process monitoring.
The Laser Photoacoustic (LPA) module provides a fast detection feature to the sensor. It has also the potentiality to easily enlarge the selection of target molecules as well as a capability to flexible learning of ever changing new measurement scenarios.
The combination of the LPA with the cantilever enhanced photoacoustic cell and the Quantum Cascade Laser has been notified as a very potential concept for great variety of commercial applications in the field of gas analysis. Several inquiries from e.g. industry, security and safety, and also research sectors have already been asked, even thought the system is not yet fully developed for commercial purposes.
The result of the project is a portable device with good flexibility to adapt different specialized applications.

Regarding the FLUO module, the objective is an automatic drug precursors sensing technology with a very low false alarm rate. This technology complements LPA technology in order to get a totally trusty drug precursors sensor measurement. FLUO sensor expected final results can be split in the biochemical compounds and methodology able to automatically transduce the presence of several selected drug precursors in air on one hand, and the low volume and cost effective hardware handling the necessary chemical agents and optical and electrical signals on the other hand.
The use of a preconcentrator will further improve the performances of the final sensor in terms of selectivity and sensibility and reducing the limit of detection (LOD) up to the ppb range.
The possibility to develop different types of binding molecules for the FLUO module and to upgrade the algorithm and the database of the LPAS module makes the sensing platform of CUSTOM able to extend at any time the species of the target analytes and so to extend the range of the applications to warfare, chemical, biochemical and toxic Industrial agents.

In order to fit with the needs of the market related to the application fields described above, few aspects of the final product should be optimized and a reshape of the two main technologies used, i.e. the LPAS module and the biochemical fluorescence, should be considered.
As regard to the LPAS module, the QCL performanceS should be improved in order to get a higher sensitivity and reach a lower limit of detection (50ppb were claimed in the proposal, current limit of detection is 3/5ppm).
The limits of the current QCL prototype have been reported in D5.6. They are mainly related to the low average output power (2.7mW in pulsed mode instead of the 200mW in CW as claimed) and the narrow wavelength range (50cm-1 instead of 200cm-1). Since they are not ascribed to final technological limitations they could be overcome by process trimming in further fabrication runs.
As to the fluorescence module, since it was designed as flexible as possible for a proof of principle prototyping, in order to get a marketable product it should be redesign to achieve the following features:
• The system used to fill the cuvette with the reagent should be changed in order to improve the loading accuracy. This could be achieved exploiting disposable cartridges of the reagents instead of the current method implemented with reservoir.
• The reduction of the analysis time. This objective could be achieved in two ways: one is by taking the reference fluorescent measurement with known concentration of analyte in parallel with the actual measurement, the other one is by increasing the temperature of the reaction chamber with the introduction of a temperature control to speed up the biochemical processes.
• Size reduction of the module by optimization of the design,
• Increasing of the number of analyte, this requires the synthesis of new proteins
• The identification of a suitable supply chain in order to make available to the market the reagents used for the fluorescent analysis. 

website: www.custom-project.eu
contacts: alberto.secchi@selex-es.com .


The partners of the project custom are listed below:

SELEX ES Italy
GASERA Finland
University of TURKU Finland
III-V Lab France
CNR IBP Italy
ENEA Italy
INSTM Italy
Aalto University Foundation Finland
Direction Nationale du Renseignement et des Enquêtes Douanières France
TECNALIA SPAIN


Figure 18: Company Logos

List of Websites:
The dissemination activities generated a flow of information about the objectives and the results obtained up to the end of the project, the contributions made to European scientific knowledge, the value of collaboration on a Europe-wide scale and benefits to EU citizens in general, but taking into consideration, each consortium members, business and trade secrets.
Four different dissemination levels have been identified in the dissemination activities.
1. Within the consortium of CUSTOM to ensure an information transfer between the members and between the commission
2. Towards the scientific community to inform scientist community of CUSTOM research results, identify suitable congress or seminars, promote research exchange and share knowledge, identify suitable collaboration projects.
3. towards the society to identify others stakeholders who would benefit of the knowledge acquired by the CUSTOM consortium,
4. Toward industrial actors to establish contact with industrial associations at national and European level.

The promotion of a common corporative identity for the project, in order to facilitate the identification of any CUSTOM material and result is done by using the project logo (reported in Figure 1: CUSTOM project logo) and making use of the common set of templates (in .doc and .ppt file extension) for documents and presentations to publish information internally and externally.
CUSTOM consortium has established an external link to a group of experts providing support for the identification of real-world operational constraints and requirements for the CUSTOM drug detection sensor. The selected experts are the customs and law enforcement authorities and the local units engaged in the fight against illicit drugs.
As potential users, the direct involvement of experts in the definition of technical requirements and performance assessment aims at strengthening the potential of an industrial and commercial exploitation for the CUSTOM results and reducing the time to market.

Figure 19: CUSTOM project logo
Every published papers containing results co-funded by the project have been coordinated and agreed by the consortium together with the Advisory Board.
Following the directives included in the GA every published results make explicit reference that the project CUSTOM is funded by the European Commission within the FP7.
The updated list of published papers, submitted abstracts, posters and conferences talks is reported below:

List of publication
1. Public Service review - Home Affairs: ISSUE 22: "Detecting Drug Precursors Developing Fluorescent Sensors and LASER Photoacoustic Sensors"
2. Public Service review - Home Affairs: ISSUE 23: "CUSTOM equipment aiding the detection of smuggled items"
3. Antonio Varriale, Sabato D’Auria, “An immuno-based surface plasmon resonance biosensor for ephedrine detection”, Security + Defence Conference Prague 19 -22 September (2011), SPIE paper 8189A-21
4. P. Sievilä, N. Chekurov, J. Raittila and I. Tittonen, ”Sensitivity-improved silicon cantilever microphone for acousto-optical detection”, submitted and under review for publication in Sensors and Actuators A: Physical, January 2012
5. Varriale, M. Staiano, VM. Marzullo , M. Strianese, S. Di Giovannni, G. Ruggiero, A. Secchi, M. Dispenza, AM. Fiorello and S. D'Auria, “A surface plasmon resonance-based biochip to reveal traces of ephedrine”, Analytical Methods, 2012, Advance Article DOI: 10.1039/C2AY25231G
6. J. Uotila, J. Lehtinen, T. Kuusela, S. Sinisalo, G. Maisons, F. Terzi, I. Tittonen (2012). “Drug precursor vapor phase sensing by cantilever enhanced photoacoustic spectroscopy and quantum cascade laser”. In D. H. Titterton & M. A. Richardson (Eds.), (Vol. 8543, pp. 85450I–85450I–13). Presented at the SPIE Security + Defence, SPIE. doi:10.1117/12.974508
7. C. B. Hirschmann, J. Lehtinen, J. Uotila, S. Ojala, R.L. Keiski (2013). "Sub-ppb detection of formaldehyde with cantilever enhanced photoacoustic spectroscopy using quantum cascade laser source". In Applied Physics B: Lasers and Optics, 111(4), 603–610. doi:10.1007/s00340-013-5379-4
8. Marco Calderisi, Alessandro Ulrici, Sauli Sinisalo, Juho Uotila, Renato Seeber, “Simulation of an experimental database of infrared spectraof complex gaseous mixtures for detecting specific substances.The case of drug precursors”, Sensors and Actuators B 193 (2014) 806– 814
9. Christophe Daniel, Vincenzo Venditto, Rosa Califano, Gianluca Fasano, Anna Borriello, Gaetano Guerra; Aerogels with Controlled Crystalline Nanopores, Biofoams2011, 21-23 Settembre 2011, Capri (NA), Italy
10. Stefano Di Giovanni, Antonio Varriale, Vincenzo Manuel Marzullo, Giuseppe Ruggiero, Maria Staiano, Alberto Secchi, Luigi Pierno, Anna Maria Fiorello and Sabato D'Auria. Determination of benzyl methyl ketone – a commonly used precursor in amphetamine manufacture Anal. Methods, 2012,4, 3558-3564
11. Secchi Alberto, “A Surface Plasmon Resonance approach for trace detection of ephedrine”, Polaris Innovation Journal

Conference Proceedings
1. Ulrici, M. Calderisi, R. Seeber, J. Uotila, A. Secchi, A.M. Fiorello, M. Dispenza, A feature selection strategy for the development of a new drug sensing system, Lecture Notes in Electrical Engineering, 2014, 162, 183-187, doi: 10.1007/978-1-4614-3860-1_32.
2. F.Terzi A.Ulrici R.Seeber A.Secchi A.M. Fiorello, M. Dispenza, J.C. Antolín, T. Kuusela, A. Varriale, S. D'Auria, I. Tittonen, F. Colao, I. Menicucci, M. Nuvoli, P.Ciambelli V. Venditto, J. Uotila, G. Maisons, MCarras, Toward a compact instrument for detecting drug precursors in different environments, Lecture Notes in Electrical Engineering, 2014, 162, 89-93, doi: 10.1007/978-1-4614-3860-1_14.
3. M. Calderisi, A. Ulrici, L. Pigani, A. Secchi, R. Seeber, Experimental design-based strategy for the simulation of complex gaseous mixture spectra to detect drug precursors, Proc. SPIE 8545, 85450B, 2012, doi: 10.1117/12.971494.
4. Ulrici, R. Seeber, M Calderisi, G Foca, J. Uotila, M. Carras, A.M. Fiorello, A feature selection strategy for the analysis of spectra from a photoacoustic sensing system, Proc. SPIE 8545, 85450K, 2012, doi: 10.1117/12.970432.
5. Secchi, A.M. Fiorello, M. Dispenza, S. D'Auria, A. Varriale, A. Ulrici, R. Seeber, J. Uotila, V. Venditto, P. Ciambelli, J.C. Antolín, F. Colao, T. Kuusela, I. Tittonen, P Sievilä, G. Maisons, Drugs and precursor sensing by complementing low cost multiple techniques: overview of the European FP7 project CUSTOM, Proc. SPIE 8545, 85450G, 2012, doi: 10.1117/12.973756.
6. Antolín-Urbaneja, J. C., Eguizabal, I., Briz, N., Dominguez, A., Estensoro, P., Secchi, A., Varriale, A., et al. (2013). Compact and cost effective instrument for detecting drug precursors in different environments based on fluorescence polarization. Optical Sensor 2013. Proceedings of SPIE 8774-56 (Vol. 2, pp. 1–11). 2013. doi:10.1117/12.2017257

Oral Presentations
1. M. Calderisi, A. Ulrici, R. Seeber, Feature selection strategy on experimental design simulation of gaseous mixture spectra, VIII Colloquium Chemiometricum Mediterraneum, June 30 - July 4, 2013, Bevagna (PG).
2. M. Calderisi, A. Ulrici, L. Pigani, A. Secchi, R. Seeber, Experimental design-based strategy for the simulation of complex gaseous mixture spectra to detect drug precursors, SPIE - Security + Defence 2012, September 24–27, 2012, Edinburgh, Great Britain.
3. Ulrici, R. Seeber, M Calderisi, G Foca, J. Uotila, M. Carras, A.M. Fiorello, A feature selection strategy for the analysis of spectra from a photoacoustic sensing system, SPIE - Security + Defence 2012, September 24–27, 2012, Edinburgh, Great Britain.
4. Secchi, A.M. Fiorello, M. Dispenza, S. D'Auria, A. Varriale, A. Ulrici, R. Seeber, J. Uotila, V. Venditto, P. Ciambelli, J.C. Antolín, F. Colao, T. Kuusela, I. Tittonen, P Sievilä, G. Maisons, Drugs and precursor sensing by complementing low cost multiple techniques: overview of the European FP7 project CUSTOM, SPIE - Security + Defence 2012, September 24–27, 2012, Edinburgh, Great Britain.
5. Ulrici, M. Calderisi, R. Seeber, Metodi chemiometrici per lo sviluppo di un nuovo sensore per l'identificazione di precursori di droghe in fase gassosa, Workshop di Chemiometria 2012 del Gruppo divisionale Chemiometria - Divisione di Chimica Analitica della SCI, May 21-23, 2012, Pavia.
6. F. Terzi, A. Ulrici, R. Seeber, A. Secchi, A.M. Fiorello, M. Dispenza, J.C. Antolín, T. Kuusela, A. Varriale, S. D'Auria, I. Tittonen, F. Colao, I. Menicucci, M. Nuvoli, P. Ciambelli, V. Venditto, J. Uotila, G. Maisons, MCarras, Toward a compact instrument for detecting drug precursors in different environments, Convegno Nazionale Sensori, February 15-17, 2012, Roma.
7. J. Uotila, J. Raittila, I. Kauppinen and J. Kauppinen, “Sensitive analysis of ambient green house gases by using cantilever enhanced photoacoustic cell combined with a quantum cascade laser”, PITTCON 2011, March 2011.
8. J. Uotila, J. Raittila, I. Kauppinen and J. Kauppinen, “Sensitive analysis of trace gases by using cantilever enhanced photoacoustic cell combined with a quantum cascade laser” 6th International Conference on Advanced Vibrational Spectroscopy (ICAVS6), June 2011.
9. Uotila, J. Raittila, I. Kauppinen, “Drugs, drug precursor and hazardous chemical sensing by quantum cascade laser and cantilever enhanced photoacoustic spectroscopy” Oral presentation, PITTCON 2012, Orlando, March 2012.
10. Uotila, J. Lehtinen, T. Kuusela, S. Sinisalo, G. Maisons, F. Terzi, I. Tittonen, “Drug precursor vapor phase sensing by cantilever enhanced photoacoustic spectroscopy and quantum cascade laser” oral presentation, SPIESecurity + Defence, Edinburgh, September 2012.
11. Marianna Loria, Vincenzo Venditto, “Drug precursors sorption in nanoporous polymers” - Forum dei Giovani Ricercatori di Scienza e Tecnologia dei Materiali 2012", Padova (Italy), 28-30 maggio 2012
12. Marianna Loria, “Drug precursors sorption in nanoporous polymers”, ISOPSC 2012, Salerno (Italy), 2-7 settembre 2012
13. Christophe Daniel, Vincenzo Venditto, Rosa Califano, Gianluca Fasano, Anna Borriello, Gaetano Guerra; Aerogels with Controlled Crystalline Nanopores, Biofoams2011, 21-23 Settembre 2011, Capri (NA), Italy
14. Vincenzo Venditto, Co-crystalline and Nanoporous Polymer Phases, 43rd IUPAC World Polymer Congress, Polymer Science in the Service of the Society, Scottish Exhibition and Conference Centre, Glasgow, GB, 11-16 July 2010, G23_O27
15. Vincenzo Venditto, Alexandra Romina Albunia, Anna Borriello, Guerra Gaetano, NANOPOROUS CRYSTALLINE POLYMERS WITH SULFONATED AMORPHOUS PHASES: HIGHLY EFFICIENT VOCs SORBENT MATERIALS, XII Congresso Nazionale di Chimica dell’Ambiente e dei Beni Culturali; Taormina, Italy, 26-30 Settembre 2010

Poster Presentations
1. Ulrici, M. Calderisi, R. Seeber, Develpoment of algorithms for the optimization of drug precursors detection with a laser photoacoustic sensing system, Giornata del Gruppo Sensori della Societa' Chimica Italiana e SIOF, September 19-20, 2013, Sestri Levante (GE).
2. Ulrici, M. Calderisi, R. Seeber, A wavelet-based procedure to detect sharp peaks in laser photoacoustic spectra of gas mixtures, VIII Colloquium Chemiometricum Mediterraneum, June 30 - July 4, 2013, Bevagna (PG).
3. A. Ulrici, M. Calderisi, F. Terzi, R. Seeber, Chemometric methods for the development of a new sensing system for the identification of drug precursors in gaseous phase, XXIII Congresso della Divisione di Chimica Analitica della Società Chimica Italiana, September 16-20, 2012, Isola d’Elba (LI).
4. A. Ulrici, M. Calderisi, R. Seeber, J. Uotila, A. Secchi, A.M. Fiorello, M. Dispenza, A feature selection strategy for the development of a new drug sensing system, Convegno Nazionale Sensori, February 15-17, 2012, Roma.
5. Kauppinen, J.Uotila J. Kauppinen, T. Kuusela, G. Maisons, X. Marcadet, “Detection of drug precursors by tunable quantum cascade laser combined with cantilever enhanced photoacoustic spectroscopy” Poster presentation, 16th international conference on photoacoustic and photothermal phenomena (ICPPP16), Merida, November 2011.
6. Antolín-Urbaneja, J. C., Eguizabal, I., Briz, N., Dominguez, A., Estensoro, P., Secchi, A., Varriale, A., et al. (2013). Compact and cost effective instrument for detecting drug precursors in different environments based on fluorescence polarization. SPIE Optics+Optoelectronics 2013. 8774-56 (Vol. 2, pp. 1–11). Prague, Czech Republic.
7. Loria, M.; Venditto, V.; Vaiano, V.; Sannino, D.; Ciambelli, P.; Secchi, A. Syndiotactic polystyrene as concentrator for drug precursors molecules, Frontiers in Polymer Science, Sitges, Spain, 21-23 May 2013, P1.161
8. Loria, M.; Venditto, V.; Guerra, G. Selectivity and uptake capacity of nanoporous aerogels of syndiotactic polystyrene towards VOCs traces, EPF2013, Pisa, Italy, 16-21 June 2013, P2.98.
Web Site
The web site (www.custom-project.eu) has been continuously updated and integrated with the activities targeted during the project. The homepage (see Figure 20) includes project and FP7 logos as well as the European Community flag.

Figure 20: CUSTOM Homepage
The public area contains the presentation of the project, the list of contacts, the project achievements, the component datasheets, links to external documents, the list of publications, etc.
It is useful to improve public awareness and target potential users. It includes also Social networking features, Statistical information and a forum area.
The password restricted area with upload facility allows the exchange of information between partners. It includes reserved documents (drafts, slides presentation, deliverables) as well as large size public documents.
The list of conferences and workshops relevant to the project is kept updated in the web site and CUSTOM repository. Usually, each partner circulates the information of details on conference/workshop opportunity to be present through the submission of papers and posters in the most relevant conferences.
Furthermore a network of several subjects involved in different FP7 projects relying in chemical sensing have been created.

A formal meeting has been held in Parma with the representatives from the FP7 projects Dirac, DOGGIES, SNIFFER, EMPHASIS LOTUS, CLARITY, INEX, BONAS, MIRIFISENS and several end users coming police departments of Greece, france and spain
The FP7-Security project DOGGIES (http://www.fp7-doggies.eu/) launched 10 months ago, aims at demonstrating an operational movable stand alone sensor for an efficient detection of hidden persons, drugs & explosives and addresses trace detection by relying on the combination of two “orthogonal” technologies, mid-Infrared spectroscopy technology (widely tuneable QCL-based integrated MIR source coupled with a miniature photo-acoustic cell) and Ion mobility spectrometry (with a non-radioactive source) coupled to the use of specific pre-concentrators and advanced data processing.
The idea behind this workshop is to share experiences, ideas and results with end users and to create links between some European projects focusing on the drugs, explosives and hidden persons detection.
Part of the information from the technical deliverables have been considered and used for scientific dissemination.
A number of project meetings, conference call and technical workshops have been periodically arranged between the whole consortium or part of it to focus on specific issues.

Figure 21: Network of the FP7 projects relying in chemical sensing.