31
Optical approaches to partial pressure measurement Dr. Stephan Putzke, PTB

Optical approaches to partial pressure measurement

  • Upload
    others

  • View
    5

  • Download
    0

Embed Size (px)

Citation preview

Page 1: Optical approaches to partial pressure measurement

Optical approaches to partial pressure measurement

Dr. Stephan Putzke, PTB

Page 2: Optical approaches to partial pressure measurement

Optical approaches to partial pressure measurement

Dr. Stephan Putzke, PTB

Page 3: Optical approaches to partial pressure measurement

Stephan Putzke – Optical approaches to partial pressure measurement

Outline

• Quadrupole mass spectrometers and partial pressures• Laser absorption spectroscopies

• Tunable diode laser absorption spectroscopy• Setup at PTB• Cavity-ringdown and Cavity-enhanced spectroscopy• Infrared-Fouriertransform spectroscopy

11.5. Laser Spectroscopy 437

amplitudes and records an exponential decay with thedecay time

τ1 = L/cT + A +αL

. (11.79a)

Without absorbing gas in the resonator (α = 0) the decaytime will be lengthened to

τ2 = L/cT + A

. (11.79b)

From the difference

∆τ = τ2 − τ1 = αL2/c(T + A)(T + A +αL)

= αL2/c(1− R)(1− R +αL)

, (11.80)

the product

αL = (1− R)∆τ/τ1 (11.81)

of the absorption coefficient α and cavity length L canbe determined as a function of the laser wavelength λ.The minimum detectable absorption is limited by thereflectivity R, the unavoidable losses A of the resonatorand by the accuracy of measuring the decay times τ1and τ2. This accuracy is limited by the attainable signal-to-noise ratio.

The physical reason for the large sensitivity is thelong effective absorption path, which is

Leff = L/(1− R + A) . (11.82)

EXAMPLE

For L = 1 m, R = 0.999, A = 0.001 ⇒ Leff = 500 m.

The experimental setup is shown in Fig. 11.47. Theincoming laser beam has to be mode-matched to thefundamental TEM00q resonator mode. Otherwise trans-verse modes are excited, which have much higherdiffraction losses (see Sect. 8.2). A careful alignment ofthe system is required to obtain optimum results. Whenthe laser wavelength is tuned across the absorption spec-trum the maxima of the absorption coefficient α atthe center of the absorption lines lead to decay timeminima and a computer program converts these mea-sured minima into the absorption coefficient α(λ). InFig. 11.48 the rotational spectrum of the overtone tran-sition in the HCN molecule, measured by Romanini

Tunable pulsedlaser

Detector

Ringdown-resonator

Mode matching

12 MM

Fig. 11.47. Experimental setup with mode-matching optics

and Lehman [11.26] is shown, illustrating the goodsignal-to-noise ratio.

The following conditions should be met to realizethe high sensitivity and spectral resolution:

1. Due to the spectral bandwidth of the laser pulsemany fundamental resonator modes within thebandwidth δωR can be excited. In order to resolvethe absorption lines with spectral width δωa thelaser bandwidth δωL should be smaller than theabsorption line width.

2. The relaxation time of the resonator must be longerthan that of the excited molecules. This demandshigh reflection coatings of the cavity mirrors (R >0.999) and careful alignment.

EXAMPLE

With a resonator length L = 0.5 m, a mirror re-flectivity R = 0.998, resonator losses per roundtripA = 0.001 and mirror transmission T = 0.0001 permirror, the decay time of the empty resonator isτ2 = 0.5/(3 ×108 × 0.002) = 8.33 ×10−7 s. With anabsorption coefficient α = 10−6/cm = 10−4/m the ab-

19,440 19,460 19,480 19,500 19,520

0.4

0.2

0.0

–0.2

I

1cm/ −

Fig. 11.48. Section of the rotational lines on the overtoneband (2, 0, 5) ← (0, 0, 0) of the HCN molecule, measuredwith CRDS [11.26]

2

Page 4: Optical approaches to partial pressure measurement

Stephan Putzke – Optical approaches to partial pressure measurement

Problems of partial pressure measurements using QMS

• QMSs do not measure partial pressures independently• Calibrations using pure gases do not hold for mixtures• Spectroscopic distinction → species-(un)specific detection• Long term stability (filament aging, detector sensitivity etc.)

Page 5: Optical approaches to partial pressure measurement

Time in sFrequency (width≙1 cm-1 at 2300 cm-1)

Abs

orpt

ion

in O

D

Stephan Putzke – Optical approaches to partial pressure measurement

T. Rubin, Diploma thesis, FU Berlin, 2009

Laser absorption spectroscopy

4

Page 6: Optical approaches to partial pressure measurement

Stephan Putzke – Optical approaches to partial pressure measurement

Laser absorption spectroscopy

5

Lambert-Beer-Law

Page 7: Optical approaches to partial pressure measurement

Laser absorption spectroscopy• Extremely sensitive for single species, no cross-sensitivities (in brackets)• Best under high-vacuum (10-3-10-6 mbar), works also at atmosph. pressure• Lambert-Beer-Law

➡ α is a species-specific constant and it is known➡ Method is (virtually) calibration free!

• Pressure from I/I0 (need to know the gas temperature)• Laser diodes are cheap!• Long absorption path → multipass/longpath cells

Stephan Putzke – Optical approaches to partial pressure measurement

Tunable Diode Laser Absorption Spectroscopy (TDLAS)

LaserDetector

Computer

11.5. Laser Spectroscopy 437

amplitudes and records an exponential decay with thedecay time

τ1 = L/cT + A +αL

. (11.79a)

Without absorbing gas in the resonator (α = 0) the decaytime will be lengthened to

τ2 = L/cT + A

. (11.79b)

From the difference

∆τ = τ2 − τ1 = αL2/c(T + A)(T + A +αL)

= αL2/c(1− R)(1− R +αL)

, (11.80)

the product

αL = (1− R)∆τ/τ1 (11.81)

of the absorption coefficient α and cavity length L canbe determined as a function of the laser wavelength λ.The minimum detectable absorption is limited by thereflectivity R, the unavoidable losses A of the resonatorand by the accuracy of measuring the decay times τ1and τ2. This accuracy is limited by the attainable signal-to-noise ratio.

The physical reason for the large sensitivity is thelong effective absorption path, which is

Leff = L/(1− R + A) . (11.82)

EXAMPLE

For L = 1 m, R = 0.999, A = 0.001 ⇒ Leff = 500 m.

The experimental setup is shown in Fig. 11.47. Theincoming laser beam has to be mode-matched to thefundamental TEM00q resonator mode. Otherwise trans-verse modes are excited, which have much higherdiffraction losses (see Sect. 8.2). A careful alignment ofthe system is required to obtain optimum results. Whenthe laser wavelength is tuned across the absorption spec-trum the maxima of the absorption coefficient α atthe center of the absorption lines lead to decay timeminima and a computer program converts these mea-sured minima into the absorption coefficient α(λ). InFig. 11.48 the rotational spectrum of the overtone tran-sition in the HCN molecule, measured by Romanini

Tunable pulsedlaser

Detector

Ringdown-resonator

Mode matching

12 MM

Fig. 11.47. Experimental setup with mode-matching optics

and Lehman [11.26] is shown, illustrating the goodsignal-to-noise ratio.

The following conditions should be met to realizethe high sensitivity and spectral resolution:

1. Due to the spectral bandwidth of the laser pulsemany fundamental resonator modes within thebandwidth δωR can be excited. In order to resolvethe absorption lines with spectral width δωa thelaser bandwidth δωL should be smaller than theabsorption line width.

2. The relaxation time of the resonator must be longerthan that of the excited molecules. This demandshigh reflection coatings of the cavity mirrors (R >0.999) and careful alignment.

EXAMPLE

With a resonator length L = 0.5 m, a mirror re-flectivity R = 0.998, resonator losses per roundtripA = 0.001 and mirror transmission T = 0.0001 permirror, the decay time of the empty resonator isτ2 = 0.5/(3 ×108 × 0.002) = 8.33 ×10−7 s. With anabsorption coefficient α = 10−6/cm = 10−4/m the ab-

19,440 19,460 19,480 19,500 19,520

0.4

0.2

0.0

–0.2

I

1cm/ −

Fig. 11.48. Section of the rotational lines on the overtoneband (2, 0, 5) ← (0, 0, 0) of the HCN molecule, measuredwith CRDS [11.26]

Laser frequency

Abs

orpt

ion

(arb

. u.)

HCN

6

Page 8: Optical approaches to partial pressure measurement

Stephan Putzke – Optical approaches to partial pressure measurement

TitleHerriott-type multipass cell

• Spherical mirrors• Path lengths from cm to several 100 m• Robust!

Page 9: Optical approaches to partial pressure measurement

Stephan Putzke – Optical approaches to partial pressure measurement

TitleHerriott-type multipass cell

• Spherical mirrors• Path lengths from cm to several 100 m• Robust!

Page 10: Optical approaches to partial pressure measurement

Stephan Putzke – Optical approaches to partial pressure measurement

Herriott-type multipass cell

Simulation software developed at PTB

x-position in mm x-position in mm

y-po

sitio

n in

mm

8

Page 11: Optical approaches to partial pressure measurement

aerodyne.com

Stephan Putzke – Optical approaches to partial pressure measurement

Multipass-cell designs

9

Page 12: Optical approaches to partial pressure measurement

aerodyne.com

Stephan Putzke – Optical approaches to partial pressure measurement

Multipass-cell designs

Sentinel Photonics

9

Page 13: Optical approaches to partial pressure measurement

aerodyne.com

Stephan Putzke – Optical approaches to partial pressure measurement

Multipass-cell designs

Sentinel Photonics

KAPITEL 3. DIE MULTIPASS-ZELLE 18

Anordnung liegen die Punkte auf den Spiegeln immer in einer Ebene.Der Strahl wird hier vom oberen Ende des einzelnen Spiegels B in die White-Zelle ein-gekoppelt und verlaßt diese am anderen Ende unter dem gleichen Winkel. Der Verlaufinnerhalb der Spiegelanordnung ist einfach zu berechnen, da alle Spiegel den gleichenKrummungsradius haben. Die konfokale Anordnung hat zur Folge, daß der Strahl auf Aund A’ immer die beiden gleichen Punkte a und a’ triÆt. Auf B entstehen mehrere Punk-te, in denen der Strahl jeweils fokussiert wird. Da die Spiegel A und A’ im Abstand Daufgestellt sind, liegen die Reflexionspunkte auf B auch im Abstand D auseinander.

A

a

a

B

1

2

3

4

A

Lichtquelle

D

d

Abbildung 3.1: White-Zelle mit 8 Durchgangen

In diesem Bild kann man auch direkt die Verlust-Bedingung fur divergente Strahlung er-kennen. Falls der gesamte Strahl vollstandig Spiegel A triÆt und dann uber B komplett aufA’ abgebildet wird, treten Intensitatsverluste nur durch BeugungseÆekte an den Randernund durch die beschrankte Reflektivitat der drei Spiegel auf.Bei der Aufstellung der drei Spiegel zueinander muß man folgende Punkte beachten. Alserstes mussen die Spiegel parallel stehen. Dann ist im weiteren der Abstand von A und A’zu Spiegel B gegeben durch den Krummungsradius. Der kritische Punkt in der Justierungliegt in der Trennung der Krummungszentren von A und A’, denn dieser Abstand Dbestimmt die Separation der Punkte auf B. Der Quotient aus dem Durchmesser von B undD begrenzt die Anzahl der Durchgange. Wenn die beiden Spiegel A und A’ symmetrischum den Mittelpunkt von B justiert sind, haben die Punkte auf B den gleichen Abstandwie die Krummungszentren von A und A’. Fur nur einen Spot auf B ergeben sich damit 4Durchgange, fur 3 Punkte genau 8 (siehe Abb. 3.1). Andere, nicht durch 4 teilbare Zahlensind nicht moglich.Wenn A oder A’ horizontal oder vertikal ein wenig geneigt sind, wandern die Punk-te auf B in der entsprechenden Ebene auseinander. Sind die beiden Spiegel nicht mehrsymmetrisch zu B angeordnet, liegen die Spots auf B nicht mehr aquidistant, sondernfallen jeweils paarweise zusammen. Jedoch treten hier bei kleinen Verstellungen keine

White-type cell(Dipl. th., I. Möller, RU Bochum ’98)

9

Page 14: Optical approaches to partial pressure measurement

aerodyne.com

Stephan Putzke – Optical approaches to partial pressure measurement

Multipass-cell designs

Jouy et al., Analyst, 2014,139, 2039

Sentinel Photonics

KAPITEL 3. DIE MULTIPASS-ZELLE 18

Anordnung liegen die Punkte auf den Spiegeln immer in einer Ebene.Der Strahl wird hier vom oberen Ende des einzelnen Spiegels B in die White-Zelle ein-gekoppelt und verlaßt diese am anderen Ende unter dem gleichen Winkel. Der Verlaufinnerhalb der Spiegelanordnung ist einfach zu berechnen, da alle Spiegel den gleichenKrummungsradius haben. Die konfokale Anordnung hat zur Folge, daß der Strahl auf Aund A’ immer die beiden gleichen Punkte a und a’ triÆt. Auf B entstehen mehrere Punk-te, in denen der Strahl jeweils fokussiert wird. Da die Spiegel A und A’ im Abstand Daufgestellt sind, liegen die Reflexionspunkte auf B auch im Abstand D auseinander.

A

a

a

B

1

2

3

4

A

Lichtquelle

D

d

Abbildung 3.1: White-Zelle mit 8 Durchgangen

In diesem Bild kann man auch direkt die Verlust-Bedingung fur divergente Strahlung er-kennen. Falls der gesamte Strahl vollstandig Spiegel A triÆt und dann uber B komplett aufA’ abgebildet wird, treten Intensitatsverluste nur durch BeugungseÆekte an den Randernund durch die beschrankte Reflektivitat der drei Spiegel auf.Bei der Aufstellung der drei Spiegel zueinander muß man folgende Punkte beachten. Alserstes mussen die Spiegel parallel stehen. Dann ist im weiteren der Abstand von A und A’zu Spiegel B gegeben durch den Krummungsradius. Der kritische Punkt in der Justierungliegt in der Trennung der Krummungszentren von A und A’, denn dieser Abstand Dbestimmt die Separation der Punkte auf B. Der Quotient aus dem Durchmesser von B undD begrenzt die Anzahl der Durchgange. Wenn die beiden Spiegel A und A’ symmetrischum den Mittelpunkt von B justiert sind, haben die Punkte auf B den gleichen Abstandwie die Krummungszentren von A und A’. Fur nur einen Spot auf B ergeben sich damit 4Durchgange, fur 3 Punkte genau 8 (siehe Abb. 3.1). Andere, nicht durch 4 teilbare Zahlensind nicht moglich.Wenn A oder A’ horizontal oder vertikal ein wenig geneigt sind, wandern die Punk-te auf B in der entsprechenden Ebene auseinander. Sind die beiden Spiegel nicht mehrsymmetrisch zu B angeordnet, liegen die Spots auf B nicht mehr aquidistant, sondernfallen jeweils paarweise zusammen. Jedoch treten hier bei kleinen Verstellungen keine

White-type cell(Dipl. th., I. Möller, RU Bochum ’98)

9

Page 15: Optical approaches to partial pressure measurement

Stephan Putzke – Optical approaches to partial pressure measurement

TDLAS setup at PTB

Gas inlet PumpsManometer

Laser-interferometer

Herriott cell (1)Vacuumchamber

Laser diode Chopper

Herriott cell (2)

MCT-Detectors

Elektronics:LaserdriverLock-in amp.ComputerEtc.

Mass spectrometer

T. Rubin, PTB

Path length: 3-100 m

10

Page 16: Optical approaches to partial pressure measurement

Stephan Putzke – Optical approaches to partial pressure measurement

CO2 in nitrogen

0,90

0,95

1,00

1,05

1,10

1,15

1,20

1,25

1,30

0 200 400 600 800 1000

Rel

ativ

e in

dica

tion

Total pressure in Pa

TDLAS QMS

CO2 partial pressure constant at 20.2 Pa N2 added to total pressure as indicated

11

Page 17: Optical approaches to partial pressure measurement

Stephan Putzke – Optical approaches to partial pressure measurement

Laser-absorption spectrum of water

Line

stre

ngth

(cm

/mol

ecul

e)

Laser frequency (cm-1)Laser frequency (cm-1)Laser frequency (cm-1)Laser frequency (cm-1)Laser frequency (cm-1)

Path length: ~30 mLaser power: 10 mWppart,H20: 10-5 mbarRoom temperatureFrom Hitran-database

12

Page 18: Optical approaches to partial pressure measurement

Stephan Putzke – Optical approaches to partial pressure measurement

H2O in nitrogen

13

Page 19: Optical approaches to partial pressure measurement

Stephan Putzke – Optical approaches to partial pressure measurement

Other light sources

• optical parametric oscillators (OPO)• frequency combs• multiplexing of several laser diodes• ...

14

Page 20: Optical approaches to partial pressure measurement

Stephan Putzke – Optical approaches to partial pressure measurement

Laser absorption spectroscopies

+ High sensitivity down to 1 ppb+ Indepentend from intensity fluctuations

- Pulsed lasers➡ complex, repetition rate, bandwidth

- optical alignment

Laser Detector AD

Computer

Cavity-ringdown spectroscopy

15

Page 21: Optical approaches to partial pressure measurement

Stephan Putzke – Optical approaches to partial pressure measurement

Laser absorption spectroscopies

+ High sensitivity down to 1 ppb+ Indepentend from intensity fluctuations

- Pulsed lasers➡ complex, repetition rate, bandwidth

- optical alignment

Laser Detector AD

Computer

Cavity-ringdown spectroscopy

15

Page 22: Optical approaches to partial pressure measurement

Stephan Putzke – Optical approaches to partial pressure measurement

Laser absorption spectroscopies

+ High sensitivity down to 1 ppb+ Indepentend from intensity fluctuations

- Pulsed lasers➡ complex, repetition rate, bandwidth

- optical alignment

Laser Detector AD

Computer

Cavity-ringdown spectroscopy

15

Page 23: Optical approaches to partial pressure measurement

Stephan Putzke – Optical approaches to partial pressure measurement

Laser absorption spectroscopies

Cavity-enhanced spectroscopy

+ High sensitivity down to 1 ppb+ Indepentend from intensity fluctuations

- Pulsed lasers➡ complex, repetition rate, bandwidth

- optical alignment

+ More simple setup➡ time-integrated detection➡ uncritical alignment

+ Requires less (CW) laser power➡ laser diodes

- slightly lower sensitivity compared to CRDS (depends...)

Laser Detector AD

Computer

Cavity-ringdown spectroscopy

Laser Detector

Computer

15

Page 24: Optical approaches to partial pressure measurement

Stephan Putzke – Optical approaches to partial pressure measurement

Fourier-transform infrared spectroscopy (FTIR)

AD

S

FT

Source

Detector

Computer

Broadband!

ir-spektroskopie.de

16

Page 25: Optical approaches to partial pressure measurement

Stephan Putzke – Optical approaches to partial pressure measurement

Fourier-transform infrared spectroscopy (FTIR)

AD

S

FT

Source

Detector

Computer

Broadband!

ir-spektroskopie.de

Retardation

0

In

tens

ity

16

Page 26: Optical approaches to partial pressure measurement

Stephan Putzke – Optical approaches to partial pressure measurement

Fourier-transform infrared spectroscopy (FTIR)

AD

S

FT

Source

Detector

Computer

Broadband!

ir-spektroskopie.de

16

Page 27: Optical approaches to partial pressure measurement

Stephan Putzke – Optical approaches to partial pressure measurement

Fourier-transform infrared spectroscopy (FTIR)

ir-spektroskopie.de

17

Page 28: Optical approaches to partial pressure measurement

Stephan Putzke – Optical approaches to partial pressure measurement

Fourier-transform infrared spectroscopy (FTIR)

ir-spektroskopie.de

17

Page 29: Optical approaches to partial pressure measurement

Stephan Putzke – Optical approaches to partial pressure measurement

Summary

• Laser absorption spectroscopy can accurately measure individual partial pressures, even in gas mixtures, traceable

• Absorption length is important, several techniques are available• Quadrupole mass spectrometers can be calibrated using TDLAS

18

Page 30: Optical approaches to partial pressure measurement

Stephan Putzke – Optical approaches to partial pressure measurement

Question:

Why don’t you implement a laser-absorption based partial pressure measurement?

It may be the right tool!

19

Page 31: Optical approaches to partial pressure measurement

Stephan Putzke – Optical approaches to partial pressure measurement

Cavity-enhanced dual-comb (FT) spectroscopy

Bernhardt et al., Nature Photonics 4, 55 - 57 (2010)

Time domainFrequency domain

Fourier transformation

20