Automatic sorting of polymer materials on the basis of the fluorescence decay time of the intrinsic fluorescence of the polymer

Abstract

The invention relates to the identification of polymer materials on the basis of the fluorescence decay time of the intrinsic fluorescence of the polymer materials for definite sorting in a completely separated manner. The invention further relates to marking with fluorescent dyes, which, because of the specific fluorescence decay times of the fluorescent dyes, can further increase the sorting reliability by means of redundancy and can be used to identify particular batches.

Claims

1. A method for detecting polymers comprising: determining individual fluorescence decay times of intrinsic fluorescences of dye-free plastics; identifying the dye-free plastics based on the decay times of the intrinsic fluorescences of the dye-free plastics; and separating the identified dye-free plastics.

2. The method of claim 1 wherein the separating comprises sorting the identified dye-free plastics in a completely separated manner.

3. The method of claim 1 wherein pulsed light sources are used for fluorescence stimulation.

4. The method of claim 1 wherein pulsed light sources are used for fluorescence stimulation, said light sources being at least one of laser diodes, light diodes, gallium nitride lamps, and gas discharge lamps.

5. The method of claim 1 wherein said dye-free plastics are detected with phase-sensitive detectors.

6. The method of claim 1 further comprising using a fluorescence decay time for detecting dye-free plastics by stimulating fluorescence with periodic pulsed light sources and accumulating and averaging a fluorescence signal.

7. The method of claim 1: wherein the separating provides a sorted separation in a completely separated manner for recycling; and wherein the dye-free plastics comprise a material comprising at least one of thermoplastics, polyoxymethylene (POM), polystyrene (PS), polyamide (PA), polyethylene terephthalate (PET), polycarbonates (PC), polyethylene (PE), polypropylene (PP), polyvinyl chloride (PVC), polyetherketone (PEK), aramids, KAPTON® (polyimide), and polysulfone.

Description

BRIEF DESCRIPTION OF THE DRAWINGS

(1) FIG. 1. Fluorescence spectra of polymer materials and the optical stimulation at 356 mm. Delrin: Dashed line, Luran; Dotted line. Ultramid: Solid line

(2) FIG. 2. Fluorescence decay behavior of Luran. Dotted line: time dependent intensity curve of the stimulation lamp. Solid, noisy curve: Time curve of the intrinsic fluorescence with deconvoluted function as Balancing curve (mostly covered by the Measured curve). On the right side: Detail of the fluorescence decay curve and balancing function with logarithmic scale.

(3) FIG. 3. Fluorescence decay behavior of Delrin. Dotted line: time dependent intensity curve of the stimulation lamp. solid, Noisy curve: Time curve of the intrinsic fluorescence with deconvoluted function as Balancing Curve (mostly covered by the Measured curve). On the right side: Detail of the fluorescence decay curve and balancing function with logarithmic scale.

(4) FIG. 4. Fluorescence decay behavior of Ultramid. Dotted line: time dependent intensity curve of the stimulation lamp. Solid, Noisy curve: Time curve of the intrinsic fluorescence with deconvoluted function as Balancing Curve (mostly covered by the Measured curve). On the right side: Detail of the fluorescence decay curve and balancing function with logarithmic scale.

(5) FIG. 5. Overview of the time-dependent intensity curve off the stimulation lamp parent open dotted line, left her in close. And from left to right the deconvoluted functions of Dehrin (solid line), Ultramid (Dotted line) and Luran (dashed line).

(6) FIG. 6. from left to right: UV/VIS—absorption—(each on the left) and fluorescence spectra (each on the right) of perylene ester 1 (buttons), the peryleneimide 2 (solid) and of terryleneimide 3 (best) in chloroform.

(7) FIG. 7. Fluorescence decay behavior of the ester 1 in Luran; fluorescence stimulation at 490 mm, detection at 573 nm. Dotted line: time-dependent intensity curve of the stimulation lamb. Solid, noisy curve: Time curve of the intrinsic fluorescence with deconvoluted function as Balancing curve (mostly covered by the Measured curve). On the right side: Detail of the fluorescence decay curve and balancing function with logarithmic scale.

(8) FIG. 8. Fluorescence decay behavior of the ester 1 in Delrin; fluorescence stimulation at 490 mm, detection at 573 nm. Dotted line: time-dependent intensity curve of the stimulation lamb. Solid, noisy curve: Time curve of the intrinsic fluorescence with deconvoluted function as Balancing curve (mostly covered by the Measured curve). On the right side: Detail of the fluorescence decay curve and balancing function with logarithmic scale.

(9) FIG. 9. Fluorescence decay behavior of the ester 1 in Ultramid; fluorescence stimulation at 490 mm, detection at 573 mu. Dotted line: time-dependent intensity curve of the stimulation lamb.

(10) Solid, noisy curve: Time curve of the intrinsic fluorescence with deconvoluted function as Balancing curve (mostly covered by the Measured curve). On the right side: Detail of the fluorescence decay curve and balancing function with logarithmic scale.

(11) FIG. 10. Fluorescence decay behavior of S-13 (2) in Luran; fluorescence stimulation at 490 mm, detection at 573 nm. Dotted line: time-dependent intensity curve of the stimulation lamb. Solid, noisy curve: Time curve of the intrinsic fluorescence with deconvoluted function as Balancing curve (mostly covered by the Measured curve). On the right side: Detail of the fluorescence decay curve and balancing function with logarithmic scale.

(12) FIG. 11. Fluorescence decay behavior of S-13 (2) in Delrin; fluorescence stimulation at 490 mm, detection at 573 nm. Dotted line: time-dependent intensity curve of the stimulation lamb. Solid, noisy curve: Time curve of the intrinsic fluorescence with deconvoluted function as Balancing curve (mostly covered by the Measured curve). On the right side: Detail of the fluorescence decay curve and balancing function with logarithmic scale.

(13) FIG. 12. Fluorescence decay behavior of S-13 (2) in Ultramid; fluorescence stimulation at 490 mm, detection at 573 nm. Dotted line: time-dependent intensity curve of the stimulation lamb.

(14) Solid, noisy curve: Time curve of the intrinsic fluorescence with deconvoluted function as Balancing curve (mostly covered by the Measured curve). On the right side: Detail of the fluorescence decay curve and balancing function with logarithmic scale.

(15) FIG. 13. Fluorescence decay behavior of the terrylene (3) in Luran; fluorescence stimulation at 490 mm, detection at 667 nm. Dotted line: time-dependent intensity curve of the stimulation lamb. Solid, noisy curve: Time curve of the intrinsic fluorescence with deconvoluted function as Balancing curve (mostly covered by the Measured curve). On the right side: Detail of the fluorescence decay curve and balancing function with logarithmic scale.

(16) FIG. 14. Fluorescence decay behavior of terrylene (3) in Delrin; fluorescence stimulation at 490 mm, detection at 667 nm. Dotted line: time-dependent intensity curve of the stimulation lamb. Solid, noisy curve: Time curve of the intrinsic fluorescence with deconvoluted function as Balancing curve (mostly covered by the Measured curve). On the right side: Detail of the fluorescence decay curve and balancing function with logarithmic scale.

(17) FIG. 15. Fluorescence decay behavior of terrylene (3) in Ultramid; fluorescence stimulation at 490 mm, detection at 667 nm. Dotted line: time-dependent intensity curve of the stimulation lamb. Solid, noisy curve: Time curve of the intrinsic fluorescence with deconvoluted function as Balancing curve (mostly covered by the Measured curve). On the right side: Detail of the fluorescence decay curve and balancing function with logarithmic scale.

DETAILED DESCRIPTION

(18) We have found that technical polymer raw materials (here as examples Luran, Dehrin and Ultramid) astonishingly show large intrinsic fluorescences if they are stimulated optically for example in the UV-region, in particular in the UVA-region, and also in the short-wave visual range; see FIG. 1. One can now try an identification via the intrinsic fluorescence spectra but would then still have the problem to sense the wavelength dependency of the fluorescence and provide an identification with that. We have found in contrast that the fluorescence decay time is astonishingly heavily dependent from the used materials; see FIGS. 2 to 4 and table 1. The time constant of the fluorescence decay time can be used as “fingerprint” of the respective polymer material; the time constant is only slightly dependent of the stimulating wavelength; see FIG. 5. Preferably the optical stimulation is performed in the UVA region with any light sources, preferably in the range of 365 nm; here are of interest the mercury vapor discharge lamp line at 365 mm or the hydrogen-Balmer lines at 397, 389 and 384 nm or the overtones of neodym YAG-laser, like for example at 355 nm or also light emissions of various semiconductor lasers (for example based on gallium-nitride), which are provided in wavelength ranges up to for example 405 nm and which suit very well as pulsed light sources. One obtains with the fluorescence stimulation at 365 mm for example by detecting at 573 mm (this wavelength is preferred because one can cover a plurality of plastics and markers) for the technical plastics Ultramid 1.96 ns, for Delrin 0.74 ns and for Luran 3.53 ns decay time (the short time component is preferred which can be distinguished very well and which enables therefore a high detection reliability); see table 1.

(19) Since the fluorescence decay curves can be in general described exactly with exponential functions, a similar procedure for the detection of various polymer materials is possible. It is not required at all to capture the complete exponential curve but rather two punctual or also integrated intensity measurements [5] (cumulation—summation—of single measurements, each over a defined time period) at different times are more than sufficient. When having a bi-exponential curve one requires three intensity measurements, if having pure plastics however the mono-exponential parts are relevant.

(20) TABLE-US-00001 TABLE 1 fluorescent decay times of pure polymers, markers In a Chloroform-solution and of the doped polymers. Probe τin ns.sup.a) τ.sub.bi in ns.sup.b) λ.sub.ex.sup.c) λ.sub.em.sup.d) Delrin 0.74.sup.e) 5.78 365 573 Luran 3.53.sup.e) 8.42 365 573 Ultramid 1.96.sup.e) 7.83 365 573 Ester 1 in CHCl.sub.3 3.53 442 485 S-13 (2) in CHCl.sub.3 4.06 490 573 Terrylen 3 in CHCl.sub.3 3.66 598 687 Delrin-Ester (1) 3.90 442 485 Delrin-Ester (1) 3.92 490 573 Delrin-S-13 (2) 3.74 490 573 Delrin-Terrylen (3) 3.31 598 667 Delrin-Terrylen (3) 3.34 490 667 Luran-Ester (1) 4.08 490 573 Loran-S-13 (2) 4.56 490 573 Luran-Terrylen (3) 3.53 598 667 Luran-Terrylen (3) 3.96 490 667 Ultramid-Ester (1) 2.44.sup.e) 7.53 442 485 Ultramid-Ester (1) 2.26.sup.e) 5.75 490 573 Ultramid-S-13 (2) 1.83.sup.e) 5.37 490 573 Ultramid-Terrylen (3) 2.34.sup.e) 6.98 598 667 .sup.a)fluorescence decay constant; .sup.b)additional bi-exponential part; .sup.c)stimulating wavelength in nm; .sup.d)detection wavelength in nm; .sup.e)evaluated bi-exponentially.

(21) Integration over each defined time period, preferably a measurement before the first half-value period and a measurement after the first half-value period is in particular advantageously because the signal-to-noise-ratio can be significantly enhanced (the fluorescent light of the probe is used more efficiently). Starting from the usual fluorescence decay times of approximately 5 ns, then this requires two measurements in time distance of 1 to 2 ns with integration times of also approximately 1 to 2 ns. Measurements with such a time resolution do not pose a problem electronically, however the measurement procedure can be further simplified by stimulating the plastic probe not only once but periodically. It is assumed that after approximately ten half-value periods the optical stimulation is reduced to such an extent that another stimulation can be carried out undisturbed; when starting from an unfavorable case having 10 ns fluorescence decay time this is achieved after approximately 70 ns. One can optically stimulate the probe periodically with a pulse sequence of 70 ns, therefore with repeating frequency of approximately 15 MHz. The two measurements for the determination of the fluorescence decay times then do not need to have to lie within single decay time, but can lie within two subsequent pulses and the requirements for the electronic components for the evaluation are further lowered.

(22) TABLE-US-00002 TABLE 2 Test for reproducibility of the measurement of fluorescent decay times and the method by repetition of single tests by differently shaked and therefore differently oriented granulate probes respectively. Probe τin ns.sup.a) τ.sub.m, in s.sup.b) λ.sub.ex.sup.c) λ.sub.em.sup.d) Delrin-Ester (1) 3.92 442 485 Delrin-Ester (1) 3.91 442 485 Delrin-Ester (1) 3.95 442 485 Delrin-Ester (1) 3.91 442 485 Delrin-Ester (1) 3.90 3.92, 0.02 442 485 Delrin-S-13 (2) 3.78 490 573 Delrin-S-13 (2) 3.77 490 573 Delrin-S-13 (2) 3.79 490 573 Delrin-S-13 (2) 3.79 490 573 Delrin-S-13 (2) 3.73 490 573 Delrin-S-13 (2) 3.74 3.77, 0.03 490 573 Delrin-Terrylen (3) 3.35 598 667 Delrin-Terrylen (3) 3.45 598 667 Delrin-Terrylen (3) 3.31 3.37, 0.07 598 667 Luran-S-13 (2) 4.55 490 573 Luran-S-13 (2) 4.56 490 573 Luran-S-13 (2) 4.43 4.51, 0.07 490 573 Ultramid-Ester (1) 2.43 442 485 Ultramid-Ester (1) 2.44 2.44 442 485 .sup.a)Fluorescence decay constant; .sup.b)Average, standard deviation s; .sup.c)Stimulating wavelength in nanometers; .sup.d)Detection wavelength in nanometers.

(23) More productive is the measurement with a fluorescence decay process by detecting the two required integrated signals with periodic stimulation in that is time-shifted and triggered by the stimulating pulse; here a separation can be performed, for instance with two parallel running phase-sensitive detectors (PSD) with which the intensities are measured integrating over two different time ranges of the decay curve. Finally it is not required to determine the absolute decay time, because device specific raw data can be used as long as they are sufficiently reproducible; for all used devices a very good reproducibility of the measured values and also of the raw data (for instance not corrected by deconvolution) was obtained. There have been found very small standard deviations s of 0.02 to 0.07 nanoseconds, which document a clear discrimination between various probes; see table 2. The method is significantly further simplified because of the unproblematic use also of raw data.

(24) When considering, that a recycling-flake is in the worst case of a maximum size of 10 mm and for reliability reasons a distance of 20 mm is imposed between two flakes then at 15 MHz pulse sequence and a form feed velocity of 500 m/s (because of technology reasons one should stay below the velocity of sound) more than 200 stimulating pulses per flake are provided. When averaging these one can enhance the signal-to-noise-ratio significantly and the detection reliability can be further increased. Assuming a mass of 25 mg for a recycling-flake (the value was obtained by averaging flakes from commercially available technical recycling material), one can sort 1.5 tons of material per hour with a sorting line. This can be regarded as a realistic technological continuous sorting power, which can be managed by the described detection method. In many cases such high sorting power is not required; when requiring small sorting power, requirements of electronics and mechanics are significantly lower.

(25) The method described here uses the intrinsic fluorescence of the materials. One can dope the polymer materials additionally with fluorescence markers; When having dope concentrations below 0.02 ppm the markers are invisible for the human eye. Markers require a minimum of light fastness; when shreddering plastics prior to the detection, this is not a serious problem because of the formation of fresh areas of fracture. For light fastened fluorescence dyes perylene derivatives and other peri-arylenes, like for example 1 to 3 can be used; for the UV/VIS-absorption- and fluorescence spectra in chloroform-solution see FIG. 6.

(26) ##STR00004##

(27) We have used as short-wave, visual fluorescing perylene derivate the perylene tetracarboxylictetraisobutylester 1 [6], for middle-wave wavelengths the bisimide 2 [7] and for longer wavelengths the terrylene derivative 3 [8]—see FIG. 12—, which are very different in their fluorescence spectra—see FIG. 6—, and we have therefore achieved an astonishingly clear additional marking option. When irradiating with a shorter wavelength into the marked materials, therefore for example at 365 nm, but also up to 405 nm, then one obtains the intrinsic fluorescence of the polymer. At 365 nm no dye will be practically optically stimulated and we have experimentally nearly only found the dominant intrinsic fluorescence of the polymeric materials; the fluorescence of the markers is negligible. When irradiating with a longer wavelength into the long wavelength absorption spectra of the dyes—here one can use also 490 nm for all marker dyes—then one only founds practically the fluorescence of the markers. This is technologically interesting, because one can identify in this way the polymer basic material by stimulating it with a short wavelength via its intrinsic fluorescence and then to sort particular charges via the fluorescence of added markers (for instance 1 to 3). Since the markers provide fluorescence in well distinctable spectral ranges, a plurality of markers can be simultaneously present and independently of each other be detected. The coding of the charges can therefore be also provided binary for example by the presence of one marker or another or also both, etc. Therefore one obtains 2.sup.n−1 coding possibilities with n markers (if undoped material is not considered). In this concrete case one obtains seven coding possibilities, which can correlate to seven distinctive charges of the respective polymers. Here the differentiation of the markers is performed based on the fluorescence wavelengths. Moreover the markers can be distinguished also in their fluorescence decay times, which are characteristic for each marker-polymer combination; See FIG. 7 to 15. One can therefore either sort based on wavelengths or based on fluorescence decay times. When using both criteria one obtains a redundancy which further enhances the sorting reliability; this can obtain a particular meaning for a special charges of high-quality materials. Finally one can dope undoped materials afterwards with fluorescence dyes, for instance by diffusing from the liquid phase or via the gas phase or also via melting processes; with that, for instance by redundancy, sorting reliability can be in particular significantly enhanced for recycling materials.

(28) The describes method here has been primarily developed for the recycling of plastics, it can however be used in general for instance manifold marking purposes, for example by replacing a barcode or QR-code markings; for the latter usually optically readable markings are required whereas fluorescence decay times can be determined also from arbitrarily light scattering objects. The marking can be attached hidden or partially hidden and can therefore for instance be used as a marking against product counterfeiting. When introducing the marker into plastics of a plastic bottle a particular reliability regarding their identification can be obtained since neither the label nor the Inscription are necessary, which can be lost.

CONCLUSION

(29) Using the fluorescence decay time of the intrinsic fluorescence polymer materials can be clearly identified and in this way can be sorted in a completely separated manner automatically. Because of the high velocity of the detection process possibilities for sorting devices with high material throughput are opened. Because of the use of phase-sensitive detectors and integrative measurements electronic costs for detection units are beneficial. Furthermore the polymer materials can be marked for their detection with fluorescence dyes, like for example perylene ester, peryleneimides, and terryleneimides, which are detected in different spectral regions via their fluorescence, or also via their individual fluorescence decay times in various polymer materials. Due to a binary combination of the marker a multiplication of the detection possibilities is provided. The identification can be realized via their spectral regions as well as via their individual decay times in that respective polymers.

LITERATURE

(30) [1] (a) E. Nemeth, G. Schubert, V. Albrecht, F. Simon, Aufbereitungs Tech. 2005, 46, 35-46. (b) E. Nemeth, F. Simon, V. Albrecht, G. Schubert, Ger. Patent (2006), DE 102004024754 B3 (12 May 2004); Chem. Abstr. 2006, 144, 392348. (d) U. Gohs, V. Albrecht, K. Husemann, E. Reinsch, R. Schuenemann, F. Simon, Ger. Offen. (2009), DE 102007055765 A1 (11 Dec. 2007); Chem. Abstr. 2009, 151:57663. [2] H. Langhals, T. Schmid, M. Herman, M. Zwiener, A. Hofer, Int. J. Environm. Engin. Sci. Technol. Res. 2013, 7, 124-132. [3] (a) H. Langhals, T. Schmid, M. Herman, M. Zwiener, A. Hofer, Ger. Offen. DE 1020120127723 (22 Jun. 2012). (b) General Electric (Erf. S. Hubbard, R. Potyrailo, P. Schottland, V. Thomas), US-Pat. 2005/0095715 (5 May 2005). [4] H. Langhals, T. Schmid, M. Herman, M. Zwiener, A. Hofer, Ger. Ojfen. DE 1020120127723 (Jun. 22, 2012); Chem. Abstr. 2013, 160, 63983. [5] (a) R. M. Ballew, J. N. Demas, Anal. Chem. 1989, 61, 30-33. (b) R. J. Woods, S. Scypinski, L. J. Cline Love, H. A. Ashworth, Anal. Chem. 1984, 56, 1395-1400. (c) R. J. Meier, L. H. Fischer, O. S. Wolfbeis, M. Schäferling, Sensors and Actuators B 2013, 177, 500-506. [6] S. Alibert-Fouet, 1. Seguy, J.-F. Bobo, P. Destruel, H. Bock, Chem. Europ. J. 2007, 13, 1746-1753. [7] S. Demmig, H. Langhals, Chem. Ber. 1988, 121, 225-230. [8] (a) H. Langhals, S. Poxleitner, Eur. J. Org. Chem. 2008, 797-800. (b) H. Langhals, A. Walter, E. Rosenbaum, L. B.-A. Johansson, Phys. Chem. Chem. Phys. 2011, 13, 11055-11059.