Hoffman D.M., Singh B., Thomas J.H. (Eds). Handbook of Vacuum Science and Technology

Подождите немного. Документ загружается.

370 Chapter 3.3: Practical Aspects of Vacuum System Mass Spectrometers

3.3.5.4 Transfer Standards

Considering the total number of mass spectrometer users, users seldom directly

calibrate most instruments by sending the instrument transducer and electronics

to an organization such as NIST. Users that do require some quantitative preci-

sion in their work most often employ a transfer standard, e.g., a spinning rotor

gauge or ionization gauge that has been previously calibrated by one of the afore-

mentioned methods. This instrument is then added to and remains as an integral

part of the system. Periodically a gas (previously employed as calibration gas) is

backfilled to the desired pressure range, and the mass spectrometer output is then

compared to that pressure. The resulting sensitivity is thus determined as a mass

spectrometer signal current out divided by the pressure as indicated by the cali-

brated SRG or ionization gauge, e.g., in (A/torr). The SRG calibration is good

down to /7 ~ 1 X 10"^ ton* and provides a two-decade overlap with the mass

spectrometer range. The resulting sensitivity is then assumed to be a constant and

independent of pressure for use in the lower pressure ranges. Direct comparison

to a calibrated ionization gauge covers a much broader pressure range (1 X 10""*

10 "^^

torr), but it is less palatable as a transfer standard, because the gauge

has the same problems as does the ion source in the mass spectrometer. Another

benefit of transfer standards is that frequent checks on the mass spectrometer sen-

sitivity can be conducted as desired without the problems of exposing the instru-

ment to atmospheric pressure, physically moving the instrument, and the vagaries

of subsequent bakeouts.

3.3.6

SOME APPLICATIONS

3.3.6.1 Industrial Applications

The following section represents examples of the types of information obtained

from vacuum system mass spectrometers in industrial environments.

LEAK DETECTION

The basic information desired from a vacuum system mass spectrometer is, of

course, very much dependent on the application. In general, all users first exam-

ine the spectrum to see if the vacuum system is leak tight and whether there is

a normal spectrum. It, therefore, is initially a leak detector. If the spectra show a

high atomic nitrogen fragmentation peak at mie = 14 amu, a high H2O peak at

m/e =18 amu, or a high molecular oxygen peak at m/e = 32 amu, then there is

3.3.6 Some Applications 371

most probably a leak. If the leak is too severe to continue, then the mass spec be-

comes a leak detector and normal leak detection procedures are implemented. If

not, then the user proceeds to normal operation procedures. In the case of a leak,

most often, the instrument is tuned to He {mie = 4 amu) and the search to locate

the leak initiated. The output of the He peak can be visually studied on an oscillo-

scope, strip chart recorder or on a computer by repetitive scans over mIe = 4 to

indicate when the leak is found. If it is a very small leak, it may be sealed by low-

vapor-pressure epoxies. If the leak partially or fully seals, the spectrum inmiedi-

ately shows a reduction in the aforementioned peaks.

VAPOR DEPOSITION

The acceptable pressure in which one can conduct thin film depositions without it

being compromised by an unwanted residual gas contaminant, is a function of the

application. The partial pressure of the contaminant species is directly related to

the incident collision rate, v, with the surfaces or volume of interest.

P (29)

{Ipmk^T)''^

where/7 is the partial pressure, m is the mass, k^ is Boltzmann's constant, and Tis

the gas temperature. The well-known rule of thumb that a monolayer of surface

contamination can be formed by an incident contaminant species (sticking co-

efficient of 1) in 1-2 seconds at approximately 1 X 10"^ torr illustrates the level

of the problem.

EVAPORATION

In deposition by evaporation or physical vapor deposition, the source material is

heated (usually by an electron bombardment) to a point where the resulting vapor

pressure provides a sufficient deposition rate to coat the substrate. This is done in

a vacuum environment that ordinarily has an ultimate pressure of ~

"^ torr

and degrades to

"^ torr during the deposition. If the partial pressure of the

contaminant gas is too high (CO, for example), the purity, growth, and properties

of the thin film may be severely altered by the dissociating carbon and oxygen.

Basically, the C and O nucleate new phases that alter the crystallinity, electronic,

mechanical, and chemical properties [33].

SPUTTER DEPOSITION

Sputter deposition is inherently a faster deposition rate than evaporation, but

more of a contamination problem because the system is backfilled to pressures

several orders of magnitude higher, p ~ 1 X

10""*

torr or greater, which gives a

372 Chapter 3.3: Practical Aspects of Vacuum System Mass Spectrometers

proportionally higher contamination rate (often the vacuum system is not as good

as those employed in an evaporation system). In addition, the purity of the back-

fill gas can create other contamination problems.

Mass spectra in these systems give an indication of what the species are as well

as what their partial pressure is, therefore providing the corresponding contami-

nation rate. This information, will ultimately provide some insight into the result-

ing film quality. Not all high background pressures are detrimental to the process,

however. For example, one might have a high background of He (perhaps from

permeation through elastomer 0-rings or a glass envelope), which will play no

part in contamination at all. It is therefore essential to know the gases that com-

prise the total background pressure.

In these last 2 deposition methods it is possible that inert gases such as He and

Ar are physically trapped in the deposit. Although chemically the film is unal-

tered, it may have different physical characteristics due to the entrapment.

CHEMICAL VAPOR DEPOSITION

This application is primarily one of process control. The application can vary

from atmospheric pressure down to the low-pressure regime and include plasma-

enhanced methods. In general, the method requires that the mass spectrometer be

differentially pumped through a capillary or an orifice [34] to allow normal oper-

ation of the instrument (<

10 ""^

torr). The various gases then can be observed

on the input and output sides of the reaction cell to control the ultimate deposi-

tion. Contaminants such as oxygen- and carbon-bearing gases are controlled to

the ppm level and can be monitored in real time to allow maximum refinement of

the process and therefore minimize contaminant-related effects.

3.3.6.2 Research Applications

As more sophisticated system requirements evolve and more demand is made on

device performance, lower and lower pressures are desired. Today, research labo-

ratories as well as many other scientific and industrial disciplines require ultra-

high vacuum (p < 1 X

10 ~^

torr). In the research discipline of surface science,

even a background partial pressure of 1 X 10"^^ torr may be too high. At this

pressure, according to our rule of thumb, a monolayer of contamination will oc-

cur in 1.5 h. Some experiments can be seriously affected with less than 1% of a

monolayer and may require experiment times longer than a 1.5 h to conduct. This

would mean that even a partial pressure of

~'^ torr is not sufficient! Mass

spectra at the operational pressure of the system are invaluable in assessing the

potential contamination to a given experiment, thus providing an indication of

whether the experiment can proceed and what effects it may have had on the re-

3.3.6 Some Applications 373

suits of that work. In research systems (where there are ion guns, electron guns,

X-ray sources, cylindrical and hemispherical analyzers, and a variety of other in-

struments and hardware), the resulting partial pressures are most certainly deter-

mined by the different construction materials contained therein, their total area,

temperature, and history as well as numerous other factors that control the out-

gassing. Mass spectra provide revealing information as to the relative state of the

system cleanliness. A figure of merit often quoted in research papers is the ulti-

mate pressure and residual gas species in the experimental system employed, be-

cause, in part, it basically reflects the quality of the work.

Mass spectrometers are used in a variety of different ways for research studies.

Some examples are as follows:

TEMPERATURE DESORPTION SPECTROSCOPY [TDS]

The mass spectrometer looks at the desorption of gases adsorbed on surfaces that

have been rapidly heated (2-10°C/s) with a specific temperature ramp (usually

linear).

The gas desorbs directly into the ion source and is subsequently analyzed.

The analysis of the rate of gas evolution allows activation energies for the process

to be determined and give some insight into the adsorption state proceeding the

desorption [35].

SECONDARY ION MASS SPECTROSCOPY [SIMS]

The mass spectrometer analyzes ions (positive and negative) and neutrals that are

sputtered by an incident ion beam of some inert gas (e.g.. He, Ar) or some reac-

tive ion (e.g., O2, Cs). The sputtered ions are energy filtered to narrow the energy

spread and the ions analyzed with a mass filter (e.g., quadrupole) [36].

CHEMICAL KINETICS ANALYZER

The mass spectrometer ion sources that are pulsed ~ 10k Hz and the ions pulsed

into a long time-of-flight (TOF) drift tube allow kinetic changes of various reac-

tions (conducted in a reaction chamber) to be analyzed for ms chemical changes.

The arrangement of the reaction cell adjacent to the ion source where the ions are

pulsed into the long flight tube allows high-resolution TOF analysis and permits

chemical changes to be monitored at the rate of better than

full spectrum per

time intervals.

TIME-OF-FLIGHT [TOF] ENERGY ANALYZER

The neutral atomic and molecular kinetic energies can be analyzed by TOF meth-

ods where the neutrals are ionized by electron impact perpendicular to the axis of

374 Chapter 3.3: Practical Aspects of Vacuum System Mass Spectrometers

the drift tube and then chopped or gated into the drift tube. The measure of the

transient time then gives the kinetic energy [37].

GAS CHROMATOGRAPH-MASS SPECTROMETRY [GCMS]

The gas chromatograph columns initially separate different specific gas species

by using selective adsorbents in a column the egress of which ultimately is mea-

sured by a mass spectrometer with much higher accuracy and sensitivity [38].

REFERENCES

F. A. White, Mass Spectrometry in Science and Technology (Wiley, NY, 1968); F. W. Aston,

Mass Spectra and Isotopes (Longmans, Green, New York, 1942); R. W Kiser, Introduction to

Mass Spectrometry and Its Applications (Prentice-Hall, Englewood Cliffs, NJ, 1965).

P. H. Dawson, Quadrupole Mass Spectrometry and Its Applications (Elsevier, Amsterdam,

1976/American Institute of Physics Press, New York, 1995).

M. J. Drinkwine and D. Lichtman, Partial Pressure Analysis, American Vacuum Society Series

(American Institute of Physics Press, New York, 1990).

H. Semat,

Introduction

to Atomic and Nuclear Physics (Rinehart, New York, 1958).

A. O. C. Nier, Advances in Mass Spec. ,6(1981)212.

6. D. G. Torr, Photochemistry of Atmospheres, edited by J. S. Levine (Academic Press, Orlando,

FL,

1985).

G. E Weston, Vacuum, 25 (1975) 469.

8. W. G. Perkins, J.

Vac,

Set TechnoL, 10 (1973) 543.

9. R. A. Outlaw and M. R. Davidson, J.

Vac.

Sci. TechnoL, A12 (1994) 854.

10.

Handbook of Chemistry and Physics, 56th ed. (CRC Press, Cleveland, 1975).

11.

L. Lieszkovsky, Handbook of Vacuum Technology (Academic Press, Boston, 1997), Chapter 3.2.

12.

Mass Spectral Data, American Petroleum Institute Research Project #44 (Carnegie Mellon Uni-

versity, Pittsburgh, PA, 1962). See also National Institute of Standards and Technology, Stan-

dard Reference Data Base, Gaithersburg, MD, for current mass spectral library of compounds.

13.

S. Rezaie-Serej and R. A. Outlaw, /

Vac.

Sci. TechnoL, A12 (1994) 2814.

14.

P. A. Redhead, J. P. Hobson and E. V. Kornelsen, The Physical Basis of Ultrahigh Vacuum

(Chapman and Hall, Lx)ndon, 1968/American Institute of Physics Press, New York, 1993).

15.

W. H. Kohl, Handbook of Materials and Techniques for

Vacuum

Devices (Reinhold, New York,

1967/American Institute of Physics Press, 1995).

16.

P A. Redhead, Can. J. Phys., 42 (1964) 886.

17.

D. Menzel and R. Gomer, J. Chem. Phys., 41 (1964) 3311.

18.

M. L. Knotek and R J. Feibleman,

Surf.

ScL, 90 (1979) 78.

19.

P R. Antoniewicz, Phys. Rev., B21 (1980) 3811.

20.

T. E. Madey and J. T. Yates, Jr., J.

Vac.

ScL TechnoL, 8 (1971) 525.

21.

M. J. Drinkwine and D. Litchman, Prog.

Surf.

ScL, 8 (1970) 78.

22.

H. A. Engelhardt and D. Menzel,

Surf.

ScL, 57 (1976) 591.

23.

A. von Engel, Ionized Gases (Oxford University Press, London, 1965/American Institute of

Physics Press, New York, 1994).

24.

J. B. Hasted, Physics of Atomic Collisions, 2nd ed. (Butterworths, London, 1972).

References 375

25.

T. A. Flaim and P. D. Ownby, J.

Vac.

ScL TechnoL, 8 (1971) 661.

26.

J. F. Hennequin and R. L. Inglebert,

Intl.

J. Mass Spec. Ion Pliys., 26 (1978) 131.

27.

R. Young, Rev. Sci. Instrum., 37 (1966) 1414.

28.

L. C. Beavis, Partial Pressure Analyzers, American Vacuum Society Education Committee,

New York, 1996.

29.

F. W. McLafferty, Interpretation of Mass Spectra (Benjamin, New York, 1967).

30.

American

Institute

of Physics Handbook (McGraw-Hill, New York, 1972).

31.

A. Berman, Total Pressure

Measurements

Vacuum Technology

(Academic Press, New York,

1985).

32.

R. E. Ellefson and D. Cain, J.

Vac.

Sci. Technol., A5 (1987) 134.

33.

R L. Swart and H. Aharoni, J.

Vac.

Sci. Technol., A3 (1985) 1935.

34.

N. Muller, G. Rettinghaus, and G. Strasser; J.

Vac.

Sci. Technol., A8 (1990) 2822.

35.

J. B. Hudson, Surface Science: An Introduction (Butterworth-Heinemann, Boston, 1992), p.292;

R. J. Madix,

Surf.

Sci., 299/300 (1994) 785.

36.

A. Benninghoven,

Surf.

Sci., 299/300 (1994) 246.

37.

G. Samorjai, Sutf Sci., 299/300 (1994) 849.

38.

I. Morisako, I. Kato, and Y. Ino,

Intl.

J. Mass Spec. Ion Phys., 48 (1983) 19.

CHAPTER

3.4

Mass Flow Measurement

and Control

Hinkle

MKS

Instruments,

Inc

3.4.1

GENERAL PRINCIPLES OF MASS FLOW MEASUREMENT

3.4.1.1 Units of Measure

The measurement of the mass of gaseous material as it enters a vacuum system

is critical to much of the work related to vacuum technology. Analogous with the

scale on which substances are weighed in a chemistry lab, the mass flow meter

provides the ability to control the mixture of gaseous substances as they enter the

process chamber. The necessity for a mass flow meter is not universal, however,

and other techniques for measuring the velocity or volumetric flow of a gas are

also quite common and useful in certain applications.

The true units of mass flow are not often used in vacuum practice, probably due

to the difficulty in picturing, for example, a kilogram of

gas

going by in a second.

On the other hand, volumetric flow units are intuitively simple —

liters

per minute

or cubic centimeters per minute. The problem with volumetric flow is that it

usually does not provide enough information. Since gases are compressible, the

amount of material in a certain volume depends on the pressure and temperature.

So without specifying temperature and pressure, the units of volumetric flow can-

not be used effectively for measuring gas flows.

$25.00 All rights of reproduction in any form reserved.

376

3.4.1 General Principles of Mass Flow Measurement 377

What has evolved over time is a system of practical units based on volumetric

flow relative to a standard condition temperature and pressure, 0°C and 1 atm re-

spectively. The most commonly used forms are seem (standard cubic centimeters

per minute) and slm (standard liters per minute). This convention assumes the

ideal gas law and uses the fact that one mole of an ideal gas occupies 22.4 liters of

volume. Therefore, standard volumetric flow relates directly to molar flow and

then to mass flow by the molar mass. For example,

500 seem of nitrogen

= 0.5 slm of nitrogen

= 0.5 slm/22.4 standard liters per mole = 0.0223 moles per minute of nitrogen

= 0.0223 moles per min X 28 grams per mole of nitrogen = 0.624 grams per

minute.

Other standard volumetric units (with an implied reference temperature of 0°C)

can be applied with the proper conversions as shown here:

1 mbar liter/second = 59.2 seem

1 Pa m-^/second = 592 seem

3.4.1.2 Methods of Measure

For continuous mass flow measurement of gases into vacuum systems, two gen-

eral approaches can be conmionly found in modern instrumentation: pressure

based and thermal based. Of

these,

the majority of instruments are thermal based.

Flow measurements may be made by measuring pressures upstream and down-

stream of a laminar flow element, an orifice plate, a Venturi element, or a molecu-

lar flow element. Also, simple mechanical devices known as rotameters (variable

area, constant differential pressure) can serve as convenient volumetric flow indi-

cators. As mentioned in the section of this handbook that discusses gas flow

equations, each of these embodiments must also involve temperature, gas proper-

ties,

and geometrical considerations to properly determine mass flow. The use of

pressure-based flow instrumentation has typically been in areas where some as-

pect of a thermal-based device, such as temperature or required pressure drop,

conflicts with the application.

The principle used in thermal mass flow sensors is that the thermal transfer

from upstream to downstream relates to the flow of thermal mass, which in turn

relates to the "true" mass flow by the specific heat c^ of the gas. For monatomic

and diatomic gases, the specific heat is relatively constant as a function of pres-

sure and temperature; the relationship between thermal transfer and mass flow is

nearly independent of the gas pressure and temperature. For more complex mo-

lecular gases, the dependence of

on temperature and pressure may have an ef-

378

Chapter 3.4: Mass Flow Measurement and Control

feet on the linearity and pressure sensitivity of the measurement. Note that there

is an inherent gas type dependence in that each gas has its own

value.

3.4.2

OVERVIEW OF THERMAL MASS FLOW

CONTROLLER TECHNOLOGY

A thermal mass flow meter (MFM) is often combined with a control valve and

electronics to form an instrument known as a mass flow controller (MFC). A rep-

resentation of these components and their typical arrangement in an MFC is

shown in Figure 1.

3.4.2.1 Sensing Techniques

The construction of the thermal flow sensor consists of a stainless steel capillary

tube with heaters and temperature sensors arranged symmetrically on the outer

surface. The ends of the tube are clamped, both mechanically and thermally, to

the body of the instrument such that the thermal environment is symmetric

around the center of the tube. Therefore, any imbalance detected by the tempera-

ture sensors is interpreted as a mass flow within the tube, the gas absorbing heat

from the upstream region and transferring heat to the downstream as it flows. A

typical sensor is designed with two matched windings of high-temperature coeffi-

cient resistance wire connected in a Wheatstone bridge circuit as shown in Fig-

Rg.l.

Control

Valve

Thermal MFC overview.

3.4.2 Overview of

Thermal

Mass Flow Controller Technology

379

Fig.

Heat Sink j^

Heat Sink

Flow

Resistance wire with high

temperature coefficient

~ V

Thermal flow sensor technique based on differential temperature.

ure 2. In this arrangement, each winding performs the heating and temperature-

sensing functions. The resulting bridge voltage represents the windings' resistance

difference and thus the temperature difference. Other methods of excitation and

sensing have been developed based on a "constant temperature, differential volt-

age"

approach or variations of these techniques.

The electronics that detect and condition the output of the sensor bridge must

amplify, linearize, and "speed up" the signal. It should be noted that this condi-

tioning results in a flow signal output in the case of an MFM; in an MFC, it also

serves as an input to a controller circuit (discussed in Section 3.4.2.4). A typical

sensor output exhibits 3-10% of full-scale nonlinearity and a time constant of

between 0.5 and 5 seconds. Various standard methods of linearization are used to

adjust the instrument output to within the typical 0.5% of full-scale linearity

specification, while an anticipation circuit, based on an active differentiator, is ef-

fective in making the output indicate the actual flow in real time.

3.4.2.2

Bypass Ranging

Most thermal flow sensors have a maximum flow limit of between 5 and 20 seem,

above which they become too nonlinear to correct for in the electronics. Since

many applications for flow measurement require higher rates, a laminar bypass is

placed in parallel with the sensor such that a fixed ratio of gas is diverted away

from the sensor. This is analogous to a shunt resistor used to select the range of an

anmieter. The bypass technique allows measurement range selection from

seem

to greater than 100 slm—all with one sensor. The standard ranges divide each

decade into three fairly equal regions (on a logarithmic scale) 10, 20, 50, 100,

200,

etc. Calibration of the meter must occur after the bypass has been installed

to account for the slight variations in the conductance from bypass to bypass and

sensor tube to sensor tube. Recall from Chapter 1.1 that the conductance for a