Control and traceability of ISPM 15 heat treatment IFQRG 12th annual meeting, September 8-12, 2014 Gabriel ROBERT, FCBA FCBA Institut technologique Forêt Cellulose Bois construction Ameublement
Context The French Food, Agriculture and Forestry Departement would like to provide border control teams with portable tool (device) to control the ISPM15 heat treatment. Some manufacturers mark the wood packaging without heat treatment It s too expensive It s too long They don t have the capacity to treat all their output Some wood packagings with ISPM 15 mark have pest and represent a risk for the forestry. Page 2
Which changes on the wood after heat treatment? A low moisture content? Like a dried wood (artificially or naturally dried) A wood structure modification? ISPM 15 Treatment temperature is too low, under 150 o C which is commonly considered to be a minimal temperature for wood modification and degradation. What about extractives rate? The content of free sugars and the amount of lipophilic extractives varied significantly (Lambertz and Welling, 2010) Page 3
Which changes on the wood after heat treatment? Volatil Organic Compounds (VOC) During kiln drying, monoterpenes are rejected and their rates in the wood decreases (Englund and Nussbaum, 2000) Alpha-pinene and beta-pinene are the main terpenes extract during drying process (Manninen et al, 2002) (Johansson et al, 1998) VOC concentration is higher at the beginning of drying (24 hours) and is correlated with temperature (FCBA/ADEME 2008) Page 4
Previous study 2005: PHYTOVER Partners: CIRAD, CEMAGREF, WTC, FCBA Funding: COFORD (National Council for Forest Research & Development) Identify potential methods that could be used to test individual pallets for compliance with ISPM 15 with literature review. Identify potential application methods and costings where these methods are fully developed. Identify necessary R&D to finalize development (or adapt a method used for other purposes to test pallets/packaging) where methods are not fully developed Page 5
7 potential methods identified 2 groups: 1. Nothing is added on the wood package 2. Something is added on. Traceability is improved but the quality system must also be tightened up Group 1 Electronic nose Bio-Chemistry-Luminescence NIRS (Near Infrared Spectroscopy) Group 2 Microparticles added during treatment RFID tags Polymers Irreversible thermo-chromium paintings Page 6
Electronic Nose Principle: identifies the specific components of an odour and analyses its chemical makeup to identify this odour. Potential problems: Discrimination of wood preservatives Differences between species (training) Is the difference between treated and not treated significant/ persistant? Costs of device (40 K - 90K ) Page 7
Bio Chemi Luminescence Principle: Measuring any molecule or enzyme in real-time Owing to the appropriate reagent: production of light in proportion of the search molecule Device: Luminometre (portable) Page 8
Bio Chemi Luminescence Page 9
Bio Chemi Luminescence Bio bois palette 16000 14000 12000 att 10000 om ole 8000 6000 D1/0h D12/6h 4000 2000 0 ATP [ATP] NA 14000000 12000000 Sample signature 10000000 RL U 8000000 6000000 D 1/0h D12/6h 4000000 2000000 0 15 60 105 150 195 240 285 330 375 420 Time in second (10 s interval) Page 10
Bio Chemi Luminescence No sampling protocol and few samples few things can be asserted (poor reproductibility of the results) Difference between treated and untreated wood regarding AN and signatures the chemical composition of wood seems to be modified (oxidasing agents ) A luminometre: from 8,000 to 12,000 depending on the model Cost of analysis: about 2-3 per analysis (reagent) Subcontracting of the analysis: from 10 to 20 Page 11
Near Infrared Spectroscopy Surface analyses with near infrared spectorscopy 2 experiences are carried out Follow-up experiment Separate experiment Page 12
Near Infrared Spectroscopy Sampling follow-up experiment 8 green blocks wrapped in cellophane heat treated at 70 C successively during 0, 5, 20, 40, 60, 240 mn 4 NIR spectra Page 13 The measurement spot is the same for each treatment duration
Near Infrared Spectroscopy Sampling separate experiment 13 green blocks and 28 green boards wrapped in cellophane heat treated at 70 C Two samples per dimension were treated during 0, 20, 120, 180, 240, 300, 360 mn spectra on RT and RL plane the measured spot is different because the heat treatment duration and the spot location are different. Page 14
Near Infrared Spectroscopy PLS calibration follow-up experiment Treatment Duration CV Predicted 300 250 200 150 100 50 R 2 = 0.970 8 Latent Variables RMSEC = 12.0981 RMSECV = 14.5341 0 5 20 40 60 240 RMSEC : standard error of calibration RMSECV : standard error of cross-validation Question : Drying effect? 0-50 0 50 100 150 200 250 Treatment Duration Measured Page 15 Water concentration difference between duration?
Near Infrared Spectroscopy PLS Validation follow-up experiment Heat treatment Duration Predicted (mn) 240 220 200 180 160 140 R 2 = 0.069 11 Latent Variables RMSEC = 16.186 RMSECV = 19.2926 RMSEP = 101.6176 NIR calibration tested on a separate test set : Not exploitable 120 50 100 150 200 250 300 350 400 Heat treatment Duration Measured (mn) Page 16
Near Infrared Spectroscopy PLS calibration separate experiment 500 Heat Treatment Duration predicted (mn) 400 300 200 100 0-100 R 2 = 0.212 12 Latent Variables RMSEC = 66.4063 RMSECV = 110.8723-200 0 50 100 150 200 250 300 350 400 Heat Treatment Duration Measured (mn) Question again : Drying effect? Water concentration difference between duration? Page 17
Near Infrared Spectroscopy Conculsion 1. Calibration with good prediction performances were observed 2. But drying effect could explain the results obtained which is not exploitable for the heat treatment traceability 3. Nevertheless relatively good calibration can be developed omitting water effect on the spectrum 4. Improvement of the calibration performances should encompass exhaustive choice of training samples, specific data pre-processing and adapted chemometrics tools. The choice of this method should be based on large calibration and validation samples set. Page 18
Work in progress IVALSA / CNR Poster presentation at COST action FP0904 meeting in 2011(Biel, Switzerland) Effect of wood treatment in low temperatures on the near infrared spectra (J.Sandak and al.) 1. The wood exposed to thermal treatment in moderate temperature changes. 2. FT-NIR technique seems to be sensitive enough to detect such changes. 3. Even if the visible variations of spectra are negligible, the spectra recorded the chemical effects of treatment. 4. 2D spectral correlation might be a tool for revealing the spectra bands related to treatment. 5. Partial Least Squares can be applied for estimation of the treatment temperature. Page 19
References Gerda Lambertz & Johannes Welling (2010) Changes in extractives of Scots pine (Pinus sylvestris L.) after ISPM 15 heat treatment and their effect on fungal discolouration, Wood Material Science & Engineering, 5:2, 67-72. Englund F., Nussbaum R.M., Monoterpenes in Scots Pine and Norway Spruce and their emission during kiln drying, Holzforschung, 2000, 54, 5, 449-456. Manninen A.M., Pasanen P., Holopainen J.K., Comparing the VOC emissions between air-dried and heat-treated Scots pine wood, Atmosph. Environ., 2002, 36, 1763-1768. Johansson A., Rasmuson A., The release of monoterpenes during convective drying of wood chips, Drying technology, 1998, 7, 1395-1428. Page 20