near infrared and raman spectroscopy for in-line ... · near infrared and raman spectroscopy for...

Institute of Polymer Research Dresden

Near Infrared and Raman Spectroscopy for in-line Monitoring of Polymers -

Applications in the Polymer Industry

µ 1. Applications with NIR-diodearray-processspectrometer SentroProc

PP/EVA, PA/FSM

µ 2. Application with the combined NIR- and Raman-spectrometer

PP/EVA (Raman), PE/PS (NIR)

µ 3. Application of in-line NIR- und Raman-spectroscopy in a mixer

control of the reaction PS/DGEBA + PS/MDEA and DMP 30

µ 4. Summary


Process monitoring – comparison of in-, on-, and off-line


NIR Process Analysis

Modern Product Philosophies• Just-in-time manufacturing• Total Quality Management• Zero defect product

Aim is reduction of production cycle time from the raw material to the final product and maintenance of quality.

Extrusion is one of the most important processes in plastic engineering for the production of polymer blends and polymer/additive systems (fillers, UV-stabilizer, etc.)

Process analysis in the polymer industry has a significant growth because of • increased demands on the quality of polymeric products• increased legislative demands• reduction of production costs by avoidance of polymers of insufficient quality • environmental protection


NIR Process Analysis

These demands require rapid, reliable non-invasive and cost effective analytical methods for process control

---> Spectroscopic methods like IR (ATR), NIR and RAMAN

On-line systemextracts the sample continuously (bypass) and presents it to the instrumentadvantage: separated from the process stream, so temperature and

pressure can be controlleddisadvantage: time lag between sampling and measurement

In-line systemmeasurement is done directly in the process lineadvantage: no sampling delays disadvantage: measurement is limited by temperature and pressure variations

the probe can interfere with the main process stream


Extrusion monitoring of polymer melt streams on a twin screw extruder with ATR-IR, NIR und Raman probes


Extrusion monitoring Equipment and experimental conditions

Single and twin screw extruders

melt pressure: 20 - 270 bar (4000 psi) melt temperature: 120 - 280 °C

a melt-at-die interface is adapted on the end of the extruders

NIR probes are in the interfaces, the probes are connectedvia optical fibers to the NIR spectrometer

NIR-probes: Transmission probes (Axiom) Diffuse reflectance probe (BTO)

NIR-spectrometer: diodearray (InGaAs) – process spectrometer SentroProc (Sentronic), 900 - 1700 nm (5800 – 11000 cm-1)with an internal drift correction for longtime stability andno moving parts inside for mechanical stability


Extruder with

NIR- and Raman-probe

Raman-probe

Adapter

melt

NIR-probe


Determination of the content of Vinylacetate (VA)in PP/EVA-blends by in-line NIR-spectroscopy

EVA is a random copolymer. The incorporation of VA lead to better flexibility and clarity and to better mechanical properties, e. g. impact strength.

Applications for PP/EVA blends are:# hot melt adhesives # resins# coatings# textiles

PP: EVA:CH2 CH CH2 CH2 CH2 CH

O

C

CH3

OCH3



Aim was the determination of the efficiency of NIR-spectroscopy for the quantitative limit of detection and the accuracy and reproducibility of this blend in real time during extrusion.

Measure time: 5 ms / scan (1000 scans in 5 sec for one spectrum)Measurement in Transmission (5 mm pathlength)25 different PP/EVA mixtures with a VA-content from 0 to 10 % were extruded

ChemometricsPLS (GRAMS32/PLSplus/IQ (Galactic))21 mixtures for calibration - 168 spectra (8 spectra for every mixture)4 mixtures for validation

Wavelength region: 1100 bis 1300 nm (9090 – 7690 cm-1)R2 = 0,999 (6 main components)



.05

.1

.15

.2

.25

.3

.35

.4

1100 1200 1300 1400 1500 1600

Absorbance / Nanometers Overlay Y-Zoom CURSOR

File #17 : PP-EVA_11_2001.08.21_12;21;19 Res=None

Probe 25 PP-EVA 10,0%

In-line spectra



.1

.15

.2

.25

.3

.35

1140 1160 1180 1200 1220 1240 1260 1280

Absorbance / Nanometers Overlay Y-Zoom CURSOR

File #14 : PP-EVA_11_2001.08.21_11;04;34 Res=None

Probe 21 PP-EVA 6,0%

C-H 2nd overtone



process.tdf,10 (R² = 0.999109134)process.tdf,10 (R² = 0.999109134)

Actual Concentration ( C1 )Actual Concentration ( C1 )

Pre

dict

ed C

once

ntra

tion

( F6

C1

)P

redi

cted

Con

cent

ratio

n ( F

6 C

1 )

-1

2

5

8

11

0 3 6 9 -1

2

5

8

11

0 3 6 9

234567891011121314151617181920212223242526272829303132

33343536373839404142434445464748495051525354555657585960616263

66676869707172 7374757677787980

8182838485868788

8990919293949596

979899100101102103104 105

106107108109110111112

113114115116117118119120

121122123124125126127

130131132133134135136

137138139140141142143144

145146147148149150151152

153154155156157158159160

161162163164165166167

-1

2

5

8

11

0 3 6 9

SECV = 0.09 %


In-line real time determination of the contentof VA in PP/EVA-blends

blend with 0,9 % VA

0,80

0,85

0,90

0,95

1,00

1,05

1,10

0 50 100 150 200 250 300

time [ s ]

VA-c

onte

nt [

% ]

accuracy (average error) = 0.08 %


blend with 0,9 % VA (2 x extruded)

0,80

0,85

0,90

0,95

1,00

0,00 50,00 100,00 150,00 200,00 250,00 300,00 350,00 400,00

time [ s ]

dete

rmin

ed V

A-co

nten

t [%

]

In-line real time determination of the content of VA in PP/EVA-blends

accuracy (average error) = 0.02 %


Determination of the content of the flame retardant melamine cyanurate (MC) in polyamide 12 (PA) during extrusion

Flame retardants are chemicals that are added to a large varietyof materials to increase their fire safety.

Many plastics are easy to ignite, therefore they are providedwith flame retardants.

The increasing use of plastics in application areas that have special demands for flame protection provides a wide arena for their use.

We can find flame-protected products in the most varied of areas:

• Buildings (insulation material, water pipes, facade facings) • Electronics (monitor housings, cables, plugs, fuse boxes, circuit boards) • Automotive (seats, roof liners, sun shades) • Airplanes/trains (seats, textiles/carpets, interior cladding, drive heads)



The demands placed on modern flame retardants have changed considerablyover the past years. In addition to effectiveness in the case of fire, environmental considerations also play a role in all phases of the product life cycle - from production, processing, application and recycling, which increasingly plays an important role.

Gas formation by melamine break down



PA is widely used in industry and for household products. It is flammable and so it is necessary to add a flame retardant.The content of the flame retardant vary according to application.

Measure time: 5 s / scan (10 scans in 50 sec for one spectrum) Measurement in Diffuse Reflectance

24 different PA12/MC mixtures with a content from 5 to 16 % MC were extruded

ChemometricsPLS, GRAMS32/PLSplus/IQ (Galactic)21 mixtures for calibration - 126 spectra (6 for every mixture)3 mixtures for validation

Data pretreatments: without pretreatment, MSC, SNV (standard normal variation)Wavelength region: 1132 – 1622 nm (8830 – 6165 cm-1) R2 = 0.999 (3 main components)



NIR-probe

RAMAN-probe



Inte

nsity

In-line spectra

Wavelength (nm)



calc

ulat

ed c

onte

ntM

C (%

)

true content MC (%)

R2 = 0.999

SECV = 0.06 %


Determination of the content of the flame retardant melaminecyanurate (MC) in polyamide 12 (PA) during extrusion

0

2

4

6

8

10

12

14

16

calc

ulat

ed c

onte

nt M

C (%

)

no pretr. MSC SNV

no pretr. 7,52 10,49 13,44MSC 7,61 10,5 13,48SNV 7,47 10,47 13,45

1 2 3= 7,5 % = 10,5 % = 13,5 %true content MC


robust, fast and combined spectrometers for NIR and Raman with embedded computers

NIR- and RAMANprocess spectrometerSentroProcNIRRAMAN(Sentronic)

Development of robust, fast and miniaturized process systems for NIR and Raman


robust probes to400 °C and 700 bar(10 000 psi), insensitive to vibration and shock

NIR-probe Ultrasonic-probes

Raman-probes

Development of robust, fast and miniaturized process systems for NIR and Raman


Determination of the composition of PE/PS-blendsduring extrusion by NIR

11 different PE/PS blends with 0 – 100 % PS were extruded in a twin screw extruder (Micro-Leistritz) at 200 °C and 70 bar.

NIR-diodearray(InGaAs)-spectrometer (SentroProc, Sentronic GmbH),900 - 1700 nm, measure time 50 s (10 Spectra); NIR-DR-probe (BTO),fibers: 400 und 600 µm.

110 Spectra (10 spectra for every mixture)

Quantitative Analysis - NIR

Chemometrics (1132 bis 1622 cm-1)PLS - R2 = 0,997 (6 components)Standard error (SECV) = 0.97 %


NIR spectra of PE/PS-blends during extrusion


Validierung der NIR-Messungen der PE/PS-Blends (40 u. 65% PS)

35

40

45

50

55

60

65

70

75

1 2 3 4 5 6 7 8 9 10

Messung

Geh

alt P

S [%

]

peps40 peps65 wahrer Wert


Determination of the composition of PP/EVA-Blends During extrusion by in-line Raman-Spectroscopy

24 different PP/EVA Blends from 0 to 6,65 % Vinylacetate-content were extruded in a single screw extruder (Viskosystem, Reifenhäuser) at 200 °C and 100 bar

Raman-Spectrometer (Holoprobe, Kaiser Optical Systems Inc.)Laser: 785 nm, CCD-diodearray detectorRaman high temperature and high pressure probemeasure time: 300 ms/scan, 40 scansfibers: 50 und 100 µm

20 blends for calibration, 4 blends for an independent validation

240 Spectra (10 spectra for every mixture


PP/EVA - in-line RAMAN-Spectra


PP/EVA - in-line RAMAN-Spectra (630 cm-1)


Determination of the composition of PP/EVA-Blends During extrusion by in-line Raman-Spectroscopy

Quantitative Analysis

I. Chemometrics – PLS (350 bis 1600 cm-1)

R2 : 0,9937 (3 components)detection limit: 0.19 % Vinylacetate

II. Integration of the band area at 630 cm-1 (590-665 cm-1 )

detection limit: 0.095 % Vinylacetat


Determination of the content of hemp fibers in PP/hemp-fiber composite materials by NIR

Investigation of PP/hemp-mixtures with hemp contents from 10 to 30 %

.4

.5

.6

.7

.8

6000 6500 7000 7500 8000 8500

cm-1

First overtone of OH of Cellulose


Determination of the content of hemp fibers in PP/hemp-fiber composite materials by NIR

R2 = 0.98 SEP = 1,26Determination of independentValidation samples

Original Property / Predicted Property

All Spectra

True Property hanf

Pred

icte

d Pr

oper

ty h

anf

10 15 20 25 30

10

15

20

25

30

Validation Spectra f(x)=0.9888x+0.2301 r=0.980147Calibration Spectra f(x)=0.9733x+0.5340 r=0.986560Validation SpectraCalibration Spectra

User Spectra

sample

True

value

Calc.

value

1

12 %

12,4 %

2

18 %

20,0 %

3

20 %

21,5 %

4

28 %

27,9 %

Nirc

al:p

p-ha

nf-2

001.

nir5

/7,s

moo

th, 4

440-

9000

. 31.

07.0

2 15

:11:

20 A

dmin

istra

tor


In-line NIR- und Raman-Spectroscopy in a Mixer

On-line NIR and Raman Spectroscopy of the Reaction PS/DGEBA with PS/MDEA and DMP 30

Mixer: Haake Rheomix 600p (Roller-Rotors R600). temperature for all experiments: 164 °C rotation speed: 60 rpm.

NIR measurements:spectrometer SentroProc NIR (Sentronic GmbHDresden, Germany) and with a diffuse reflectance fiber optic probe with a sapphire window.wavelength region: 1030 - 1660 nm time for one spectrum (50 accumulations): 1 min

Raman measurements: spectrometer HoloProbe (Kaiser Optical Systems, Inc., USA) and with a fiber optic probe with sapphire window.measure time for one spectrum (20 accumulations): 1 min



Change of the baseline of the NIR spectra at 1250 nm(reaction 50PS/50DGEBA + 73PS/27MDEA, 07.11.01)

0,4

0,6

0,8

1

1,2

1,4

1,6

1,8

2

0 600 1200 1800 2400 3000 3600 4200 4800 5400 6000 6600 7200

time in s

abso

rban

ce a

t 125

0 nm



.7

.8

.9

1

1.1

1200 1400 1600

1144 (reference)

1425 (OH)1497 (NH2)

X-Zoom CURSORAbsorbance / NanometersFile #17 : EPOXY+PS_123456_2001 Res=None

Figure 5: online NIR spectra of the reaction 50PS/50DGEBA + 73PS/27MDEA, 07.11.01



. Change of relative absorbance for the NIR bands at 1425 nm and 1497 nm (reference at 1144 nm)

0,9400

0,9900

1,0400

1,0900

1,1400

1,1900

0 10 20 30 40 50 60 70 80 90 100 110 120

time in min

rela

tive

abso

rban

ce

NIR band at 1425 nmNIR band at 1497 nm

OH

NH2

Figure 6: Experiment A: reaction 50PS/50DGEBA + 73PS/27MDEA, 07.11.01



Spectrum after 1 minSpectrum after 45 minSpectrum after 90 min

epoxy band

.

5200

5400

5600

5800

960 940 920 900 Arbitrary Y / Raman Shift (cm-1)File # 1 = IN0006 50PS/50DEGBA + 73 PS/27 MDEA 1 20

Y-Zoom CURSOR13.11.01 07:40 Res=None

Figure 7B: Raman spectra, experiment A: reaction 50PS/50DGEBA + 73PS/27MDEA



Spectrum after 1 minSpectrum after 45 minSpectrum after 90 min

.

6600

6400

6200

6000

5800

1300 1250Arbitrary Y / Raman Shift (cm-1)File # 1 = IN0006 50PS/50DEGBA + 73 PS/27 MDEA 1 20

Y-Zoom CURSOR13.11.01 07:40 Res=None

Figure 7C: Raman spectra, experiment A: reaction 50PS/50DGEBA + 73PS/27MDEA



.

Figure 9: Experiment A: reaction 50PS/50DGEBA + 73PS/27MDEA,

Predicted conversion [%] calculated with calibration model (PLS)

0

20

40

60

80

100

0 10 20 30 40 50 60 70 80

time in min

pred

icte

d co

nver

sion

[%]

predicted conversionvalues, determined by INSA


Summary - NIR spectroscopy for in-line process analysis

Advantages- transmission and reflectance measurements (for opaque melts)

- In-line determination in real time of the components of the melt

- Adaptation of robust high pressure and high temperature probes

by fibers up to 1000 m between probe and spectrometer

Disadvantages - calibration with a precise reference method necessary - chemometric method for quantitative evaluation necessary- limit of detection is dependent from the investigated polymer system (0.02 - 0.5%)


Summary - NIR spectroscopy for in-line process analysisdiscussion of important points

Calibration and Validation data sets versus Cross Validation

Background-Spectrum: Spectralon-Standard; time between two backgrounds

Diffuse Reflectance:* the investigated systems should be homogenous * all calibration mixtures should have the same stray properties like the investigated

real system in the process (transfer of calibrations from the lab to the process is problematic)

Temperature and pressure have to be checked at the position of the probe

Optimisation of the connection probe - fiber – fiber interface in the spectrometer is very important

Use of fiber bundles or single fibers


Summary - Message

Near infrared (and RAMAN) spectroscopy are powerful tools for process analysis, especially for monitoring polymers melts in extrusion processes.

In all investigations, emphasis should be given on the securing of the calibration during modifications of the spectrometer and the optimization of the fiber link between spectrometer and the probe.

The quantitative analysis carried out with multivariate methods can determine the composition of multicomponent mixtures.

near infrared and raman spectroscopy for in-line ... · near infrared and raman spectroscopy for...

Documents