ITT | OI Analytical, 2148 Pelham Parkway, Bldg 400, Pelham, AL 35124Omar Hadjar, Gottfried Kibelka, Scott Kassan, Chad Cameron, Ken Kuhn

Experimental Results

IntroductionMany industries rely on the tight control of chemical processes to profitably produce a high quality product. High sample throughput and a dependable analyzer enable real-time analyses from receiving inspection to post-production verification. Spatially dispersive, non-scanning mass spectrometers (MS) continuously provide full mass spectra, thus

enabling rapid analytical techniques with enhanced sensitivity.

Miniaturizing a sector-field MS and joining it with a detector array that combines high spatial resolution and linear response results in a transportable instrument offering speed and precision, the IonCam . Ion beams shaped and focused by the ion optical elements of the double-focusing MS are detected by an IonCCD in the focal plane of the

instrument. This allows the observation of many chemical processes in real time. Sequences of mass spectrometric frames up to 360 frames/s (< 3 ms temporal resolution) will greatly help understanding the dynamics of various processes.

Experimental Setup

References

Simulation: Design upgrade

The ions are

finally m/z dispersed

and energy focused on to a focal plane.

The m/z peak width is proportional to the

slit width, the magnetic radius and inversely proportional

to the electrostatic radius. The mass range of the analyzer, better

defined as Mmax/Mmin, is given by the analyzer geometry. The first parameter

is the blind plane length (lbp), defined as the distance between the ion entrance

on the magnet and the first pixel of the detector array. The second parameter is the focal plane

length (lfp) which is defined by the array detector length: Mmax/Mmin = (1 + lfp / lbp)2

1170 1200 1230 1260 1290 1320

-10

0

10

20

30

1170 1200 1230 1260 1290 1320

-10000

0

10000

20000

30000

10-1

100

101

102

103

102

103

104

105

106

+101

with averaging

103 with

out avera

ging

100 ion/dN

Ion

CC

D r

esp

on

se a

t 1

5 m

s in

t. t

ime

(dN

)

ion beam current (pA)

50 ion/dN

1 frame:

S/N=188/107=1.8

Ion

CC

D s

ign

al

(dN

)

pixel number

200 frames:

S/N=188/10.7=18

Ion

CC

D s

ign

al

(dN

)

pixel number

The IonCCD1, 2 is designed around

the MH-MS analyzer, hence made of

a single array of pixels (one-

dimension) spanning 51 mm length,

matching the focal plane length (lfp)

and 1.5 mm pixel height matching

the magnet air gap. The pixel width is

21 m and insulating gap is 3 m

providing a pitch or spatial resolution

of 24 m. The result is a total of

2126 pixels with 88 % pixel area

ratio (PAR). The detector operates

from atmospheric pressure to high

vacuum as no high voltage is used.

The IonCCD analogue voltage signal is digitized

by a 16 bit ADC so all data are expressed in digital

numbers (dN). The IonCCD has a response of 100

ions/dN/pixel.2 The noise floor when not cooled

and frame averaged is 9 dN/pixel. In terms of ion

current density, the IonCCD has a limit of

detection (LOD) of 0.5 fA/pixel.2 The detector

provides up to 360 frames or spectra per second

providing about 3 ms temporal resolution.

Imin(A/pixel)=R*3*n*q/t

Temporal resolution: 3 ms

Integration time: t=83 s-5s

Frame rate: 360 Hz

Spatial resolution (pitch): 24 m

Array size: 51 x 1.5 mm2

Total pixels: 2126

Noise floor: n=9dN

Response: R=100 ions/dN/pixel

Single-frame dynamic range: >1000

Detector size: 105 x 67 x 30 mm3

Operation Pressure: 103 to 10-8 torr

Operation bias voltage: 0 to 3 kV

0 200 400 600 800 1000 1200 1400 1600 1800 2000

10

100

1000

10000

Ion

CC

D s

ign

al (

dN

)

IonCCD pixel number

10 20 30 40 50 280 290 300 310 320 330 340

10

100

Ion

CC

D s

ign

al (

dN

)

IonCCD pixel number

To achieve atomic hydrogen detection, the standard magnetic sector (1.04 T) used for routine mass window detection of [8,

128] u is replaced by a weaker one (0.38 T) allowing mass window detection of [1, 16] u & [4, 64] u. In the case of the 0.38 T

magnet, the first and second window are achieved by a 1000 and 250 V acceleration respectively. The mass spectrum shown

below is produced with a 1000 V acceleration and the signal is expressed in a logarithmic scale. The orange and black spectra

are taken with a He sampling bag on and off respectively. The atomic hydrogen ion, or proton, originates from the residual gas

and is visible on both spectra. The ion peak at mass 2 u in the graph bellow is produced by doubly charged He ( alpha particle)

and or molecular hydrogen ions. The zoom around mass 1 u and 2 u is shown to the right. The zoom clearly shows the base line

separated double peak structure demonstrating that both species are detected and nicely separated.

-1 0 1 2 3 4 5 6 7 8 9 10 48 49 50 51

0.0

0.5

1.0

1.5

2.0

2.5

m=

71

u,

FW

HM

= 1

19

m

m=

69

u,

FW

HM

= 1

12

m

m=

70

u,

FW

HM

= 1

12

m

m=

68

u,

FW

HM

= 1

08

m

m=

72

u,

FW

HM

= 1

21

m

m=

67

u,

FW

HM

= 1

14

m

m=

6 u

, F

WH

M=

46

m

m=

5 u

, F

WH

M=

45

m

m=

4 u

, F

WH

M=

49

m

m=

3 u

, F

WH

M=

61

m

m=

2 u

, F

WH

M=

92

m

last active pixel

Tra

nsm

issi

on

(2

4

m b

in h

isto

gra

m)

IonCCD array (mm)first active pixel

1.000 T 1.045 T

1.2 kV

100 m slit

50 mm ESA

m=

1 u

, F

WH

M=

19

7

m

High Mass Dynamic Range Magnetic Assembly

0 1 2 3 4 5 6 66 67 68 69 70 71 72 73

0.0

0.5

1.0

1.5

2.0

2.5

M=

72

.03

9 u

, M

= 0

.30

8 u

M=

71

.02

9 u

, M

= 0

.30

1 u

M=

70

.00

0 u

, M

= 0

.28

2 u

M=

68

.98

7 u

, M

= 0

.27

9 u

M=

67

.96

7 u

, M

= 0

.26

7 u

M=

66

.95

1 u

, M

= 0

.28

0 u

M=

6.0

37

u,

M=

0.0

33

u

M=

5.0

07

u,

M=

0.0

29

u

M=

3.9

86

u,

M=

0.0

28

u

M=

2.9

79

u,

M=

0.0

30

u

M=

1.9

82

u,

M=

0.0

37

u

1.045 T

last active pixel

Tra

nsm

issi

on

(2

4

m b

in h

isto

gra

m)

m (u)first active pixel

1.000 T

1.2 kV

100 m slit

50 mm ESA

M=

1.0

14

u,

M=

0.0

56

u

High Mass Dynamic Range Magnetic Assembly

3rd

order polynomial fit using 12 peaks

The mass range of the present MH-MS appears to be acceptable when it is tuned for a minimum mass of

C+, resulting in a mass window of about 200 u. However for hydrogen applications, the mass range drops

to a very narrow mass window of 16 u. One can consider scanning the energy to produce different mass

windows which then can be stitched. This would, first, defeat the whole purpose of the non-scanning

instrument. Secondly, it would decrease the duty cycle drastically. Calibration of masses must be

performed for every energy. Also, the signal must be corrected for the dependence of ion transmission on

the energy. Last but not least, more complex software is needed to orchestrate all of the steps above. To

attack this problem, one could think of an increase in the dynamic mass range providing a acceptable

detectable mass window at single acceleration. This is achieved either by increasing the array detector

length (or butting two IonCCDs) and increasing the magnet size or reducing the blind plane length (lbp).

While the first option is very costly with analyzer size increase, the second option is straight forward with

no instrument size increase, allowing the use of the same detector. This last option calls only for a slight

change of the magnet shape, referred to as High Dynamic Mass Range Magnet (HDMR). The question is

how would the peak shapes and mass calibration suffer from such a change? The simulation above shows

the scenarios where lbp is shortened from 17 to 7 mm, theoretically increasing the dynamic mass range

from 16 to 69. A dynamic mass range of 72 is simulated. This is a welcome byproduct of the reduced lbp

(lower B-field at low mass due to the lack of magnetic material). The mass calibration, in this case, would

be performed using both B values for low (blue) and high mass (green), introducing no serious calibration

issues, especially at such unit mass resolution. However this outcome needs to be proven experimentally

in the near future. The result is a full mass range as observed in the above simulated spectra.

Mass calibrated spectrum showing the mass resolution of the

instrument over the detected mass window [1, 72] u.

Conclusions

Beam profile of the dispersed m/z ions with energy and direction

focus along the 51 mm long focal plane IonCCD array detector.

• Non-scanning MH-MS instrument is used in combination with a focal plane array detector IonCCD

• The IonCCD shows a 103 dynamic range within a single frame with a 0.5 fA/pixel sensitivity

• The IonCCD acquisition speed of 360 Hz is key for fast GC and GCxGC (< 100 ms peak) applications

• Preliminary experimental data show promising results with Hydrogen to Oxygen detection

• Baseline separation between He2+ & H2+ (2 u nominal mass) is achieved at 1000 V and 0.38 T

• Simulation suggests a possible dynamic mass range increase with no loss of instrument performance

At 70 eV, the electron impact ionization cross section3 of H2 is about 1*10-16 cm2, smaller than

those of N2 and O2 which are about the same, 2.6*10-16 cm2. The tight focus of H+ and H2+ due

to the short magnetic radii compensate the latter effect. Despite the fact H+ and O+ are

detected with a lag of 3 s at 1000 V (SIMION 8.0), the simultaneous detection statement

remains correct at 99.9 %, considering the 3 ms temporal resolution of the existing camera (83

s minimum integration time). In the best case scenario, with the ultimate camera system

matching the chip’s maximum frame rate of 2 kHz (500 s temporal resolution), the

simultaneous detection will only drop to 99.4 %. The dynamic mass range increase (see next

part) will increase the above detection lag to about 5 s, which is of no consequence (99 %).

1. M. P. Sinha and M. Wadsworth. Miniature focal plane mass spectrometer with 1000-pixel modified-CCD detector array for direct ion

measurement. Rev. Sci. Instr.76 025103(2005).

2. O. Hadjar et al. IonCCDTM for direct position-sensitive charged-particle detection: from electrons and keV ions to hyperthermal biomolecular

ions. J. Am. Soc. Mass Spectrom. Online (2011) April issue cover.

3. W. Hwang and Y.-K. Kim. New model for electron-impact ionization cross sections of molecules. J. Chem. Phys. 104 2955(1996).

Monitoring Hydrogen and Gaseous Fuels

using a Double-Focusing Mass Spectrometer

Pittcon 2011, Atlanta

The gas phase sample is introduced through a fused silica capillary to the ionization chamber. The

molecules are ionized by 70 eV electron impact generated by a hot rhenium filament, not

shown below for clarity. The electron impact (EI) ion source is biased

at 1000V. The double focusing analyzer of Mattauch-Herzog geometry

(MH-MS) is operated at ground potential. The ion energy is

then defined by the EI source bias. The ions are

extracted through a 100 m object slit and

directionally focused through

a 31.8o electrostatic

analyzer.