Technical Report No. IRL-1062 November 1, 1967

A COMPUTER OPERATED
MASS SPECTROMETER
SYSTEM

Walter E. Reynolds Timothy B. Coburn
James C. Bridges Robert B. Tucker

 

Instrumentation Research Laboratory, Department of Genetics
Stanford University School of Medicine
Palo Alto, California
A COMPUTER OPERATED
MASS SPECTROMETER SYSTEM

Walter E. Reynolds Timothy B. Coburn
James C. Bridges Robert B. Tucker
Technical Report No. IRL-1062 November 1, 1967

Prepared under:

National Aeronautics and Space Administration
Grant NsG 81
National Institutes of Health
Grant FR 00151
Air Force Office of Scientific Research
Contract AF 49(638)-1599

Principal Investigator: J. Lederberg

Director, Instrumentation Research Laboratory: E. Levinthal

Instrumentation Research Laboratory, Department of Genetics
Stanford University School of Medicine

Palo Alto, California
ABSTRACT

A mass spectrometer-computer system has been devised to utilize
the decision-making capabilities of the modern digital computer. The
system described assists the researcher user by allowing a computer to
query the researcher for operating parameters. The computer translates
these into detailed control functions that operate the instrument. The
data acquired from the mass spectrometer is made available to the
researcher in an on-line system. The system employs a small digital
computer and an integer resolution quadrupole mass spectrometer. A
reference gas is valved into the mass spectrometer by computer control
to permit automatic calibration. Spectra processing of GLC effluent

was demonstrated; the means and results are given.

FC-43 REFERENCE FLUOROCARBON

AMPLITUDE (7)
Sp Sp 7p POP ipo

Pf .

7p 3

 

_f

 

 

p

 

 

MASS

iii
II.
IIl.

IV.

VI.

VII.

VIII.

Referenc

Appendix

CONTENTS

Introduction ..... + «+ « «
Theory of Operation. .....
Overall Physical Description .

The Software Structure ....

Description of the Interface Electronics

The Software and Its Concepts.
A. Calibration .....-
B. Data Acquisition. + « -
C. Data Manipulation ...

bD. Data Presentation ...

An Example of User-Computer Dialogue

Conclusion ..... + ««s«-.

CS « «© © © 28# # © © «© © &@ © ee «@

iv

+ 40

~ 42
Figure

6-10

Al

A2

A3

A4

Table

ILLUSTRATIONS

A block diagram of the computerized mass spectrometer .
system

The inlet system to the quadrupole mass spectrometer

Block diagram of the electronics of the IRL interface.
between the LINC computer and the EAI quadrupole mass
spectrometer

Systems flow chart

. . oe o «6 e . e . ee 7 e . e

Partitioning and flow of a routine under SPECTRUM.

Data presentation by a Calcomp plotter

. o ef e@ * «@ eo

The IRL interface with a local Model 33 Teletype
and a 103A Data Set

Data flow between the SEND and RECEIVE stations.

The IRL hardware modification which removes an
incoming parity bit from a coded character

Timing diagram of parity delete circuit. .......

TABLES

A list of programs in SPECTRUM .......-6.6..e.-s

Logic Outline of DISPLAY ...... 2... 2 ee eee

. 16

. 27

37-39
I. Introduction

The use of a mass spectrometer as a research tool has been hampered
by the slow development of fast and convenient methods of reducing the
output to numerical data. The output (time versus intensity) is related
to the mass spectrum (mass/charge versus intensity) by the characteristics
of the scanning circuit and the nature of the mass analyzer (or dis-
criminator). The theoretical relationship is marred, however, by the
practical difficulties of construction and the drifts and instabilities
which occur in operation. Therefore, it is difficult to establish a
unique relationship between the settings of the operating controls and

the output of the mass spectrometer.

A common method of mass-peak identification is to add a reference
compound to the unknown sample. Then the oscillograph recording contains
mass markers, corresponding to the known peaks of the reference compound.
The operator counts up or down from the mass markers to identify the
peaks of the unknown sample. The output of a mass spectrometer,
particularly when it is used in conjunction with a gas chromatograph,
provides the researcher with more information than he can process with

this approach.

Several people have conceived and demonstrated systems which
monitor and record digital data from a mass spectrometer at rapid
(1-10 sec.) scan rates. (See references 1, 2, 3 and 4.) The usable
scan rate is sometimes limited by the particular mass spectrometer
used. In high resolution spectrometry, it may be limited by the sample

time required to determine the mean time of arrival at the collector of

1
a statistically small number of ions. All of the systems described in
the references deliver their data to a modern digital computer (some in
real time) for processing. The system described in this report is one
of many efforts to enhance the potential of the mass spectrometer

through its use with a digital computer.

Our system employed the EAI QUAD 300 Quadrupole Mass Spectrometer
*
and a LINC computer (9). (An excellent description of the principle
of the quadrupole appears in reference (5), pages 34-38.) The QUAD 300

provides integer mass resolution from mass 10 to mass 500.

There are two distinctive features of our system, One is the high
level of operator involvement, making him an important system element.
Through a standard Model 33 Teletype, the operator engages the computer
in a conversational dialogue in which the computer requests input and/or
output parameters and some branching decisions in the software
algorithms. The dialogue also allows the operator to initiate output,
via a Calcomp 565 digital plotter, and exercise some control over its

manner of presentation. (The software is detailed later in this report.)

The human operator and the small computer perform together as a
larger computer might perform alone. However, the task of programming
the larger computer would be more complex. One would like to reduce

the duties of the operator to merely pushing a few buttons. A push-

 

*
The LINC computer in our laboratory is a small instrument computer

(2K memory with built-in tape unit), designed and built as a research
project at Washington University. It is very similar to the LINC 8
manufactured by the Digital Equipment Corporation (146 Main Street,
Maynard, Massachusetts), and the Micro-Linc put out by Spear, Inc.
(335 Bear Hill Road, Waltham, Massachusetts).
button system which would do as much as the system described in this

report is beyond our present capability.

The second distinctive feature of our system is its method of mass-
peak identification. Under the guidance of the operator, the computer
valves in a reference compound which it uses to calibrate the scan
circuit by relating the time versus intensity output to the known mass
peaks of the reference. Subsequently, the control voltage which the
computer interface feeds to the scan circuit is related to a mass index
stored in the computer memory. This allows the output of the mass spec-
trometer with an unknown sample to be recorded directly as mass/charge
versus intensity. In the other referenced systems, the computer
receives the data of the unknown sample as time versus intensity and

then converts to mass/charge versus intensity.

The spectra which are reproduced in this report show 256 as the
highest recorded mass. The primary reason for this limit is the
simplicity of programming effort which it afforded. Two hundred and
fifty six words corresponds to one "quarter" of the LINC core memory.
A quarter of LINC core may be read from or written on LINC tape with a

single machine instruction.

In this system, a QUAD 300 Mass Spectrometer was used to monitor
and identify the output of a Varian gas chromatograph. In a typical
experiment, amino acids were coupled with an optically active reagent
to give diastereoisomers resolvable by gas chromatography. The mass
spectrometer was then used to identify and determine the optical purity

of the stereoisomers (6,7). In a second application (8), amino acid
mixtures were derivatised to give compounds separable by gas
chromatography. The QUAD 300 was then used to confirm the structure

of the derivatives and to identify the volatile by-products of the

reaction.
II. Theory of Operation

During spectrum data acquisition, the computer directs the mass
analyzer to a program-selected mass position and reads the output
intensity of the mass spectrometer. The mass analyzer is not swept in

the conventional sense. It is controlled by a voltage (V) such that

m/e = £(V)

where m/e is the mass/charge ratio and f(V) is a monotonic function
characterized by the mass spectrometer. V is the output of a computer
directed digital-to-analog converter. For every M (M = m/e) to be
passed by the mass analyzer, the computer identifies (according to the
calibration program) a digital number N. Calibration programs assist

the operator in determining the empirical relationship between N and

M.
The procedure to measure the intensity at any M is as follows:

(1) The number Ny which corresponds to the selected mass/charge
ratio (M,) is loaded from the computer into the digital-to-analog con-
verter. This sets the control voltage of the mass analyzer. The output

of the mass spectrometer is proportional to the quantity of M, present

i
in the sample.

(2) An analog circuit, reset and released by the computer,
integrates the output of the mass spectrometer. The output of the
integrator is sent to an analog-to-digital converter in the input

circuitry of the computer.
(3) Several milliseconds after the integrator is released, (the
choice of integration time is supplied by the user upon program request),
the computer samples the output of the converter. That digital value is

stored as the intensity of M,-

Steps (1) through (3) are repeated to acquire a complete spectrum.
The fundamental restraint upon this system is the drift of the
mass spectrometer following calibration. Our experience with the
QUAD 300 is that this drift causes an error in N of less than 1/8 the

value from one N entry to the next in a one hour period. This is

sufficiently small to allow unambiguous mass identification.
III. Overall Physical Description

The system elements referred to or mentioned in this section are

explained in detail in the following sections.

Figure 1 is a block diagram of the system. The ground floor
connections of the teletype and Calcomp to the LINC have been used for
several years in other applications. (See Technical Report IRL-1055)
They are duplicated near the site of the mass spectrometer research
activity for the convenience of the principal users of the system.
Except for sample injection, all input and output are through the

teletype keyboard and Calcomp plotter.

The LINC computer was made available to the IRL under an NIH
grant. Designed primarily for instrumentation, the LINC provides hard-
ware inputs and outputs to several buffered registers, as well as pulse
outputs and input level-sense lines. It has a 2K core memory and a
built-in tape unit. "Tape Read" and "Tape Write" functions are single

machine instructions.
The cable connecting the two terminal boxes contains:

(1) Twelve lines to carry the twelve bits of the LINC accumulator.

(2) A pulse line, OPR 10, to transfer the accumulator contents
into a register which controls the mass position on the
quadrupole.

(3) OPR 11, which clocks the contents of the accumulator into a
register that controls the mass sweep, reference gas inlet,

integrator reset and integrator release.
 

GLC.

 

 

 

 

EAI QUAD MASS

 

SPECTROMETER

 

 

 

 

d

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

A block diagram of the computerized
mass spectrometer system.

 

 

 

 

 

CABLE
TERMINAL
BOX

 

 

 

 

REF. GAS
ELECTRIC
VALVE
cRO
Y AXIS IRL COMPUTER-MASS SPECTROMETER INTERFACE eee
1 (DIGITAL-TO-ANALOG CONVERTER, REMOTE DIGITAL
X AXIS REGISTERS, SIGNAL CONDITIONERS, ETC,) PLOTTER
3300A
CABLE
TERMINAL TELETYPE
BOX
THIRD FLOOR OF MEDICAL CENTER
l~—— ABOUT 200 FT. OF CABLE
GROUND FLOOR OF MEDICAL CENTER
Figure 1

TELETYPE

 

 

 

 

 

 

 

 

 

 

 

CALCOMP
565
DIGITAL
PLOTTER

 

 

 

LINC

COMPUTER

 

 

 

 
VACUUM
THERMOCOUPLE
GAUGE

 

QUAD 300
CONTROLLER

 

 

 

 

TO A4 MELCOR
OF THE ———— 1619
INTERFACE FET OR AMP. A3

 

 

@

 

 

 

 

RESERVOIR

REF. COMPOUND FC-75,
A LIQUID MANUFACTURED
BY 3-M

 
  
 

 

ELECTRICA
VALVE

 

Figure 2

OPERATED

  

LLY

 

 

 

 

VACUUM

 

  

METERING VALVE
3° TAPER

The inlet system to the quadrupole mass spectrometer.
portion of the diagram enclosed in dotted lines was constructed

by the Instrumentation Research Laboratory.

EAl
SYSTEM

  

 

 

 

 

 

 

—

QUADRUPOLE
MASS
FILTER

 
   
 

 

 

 

   

BIEMANN
SEPARATOR

 

 

GLC.

 

 

 

The

BAYER-ALPERT TYPE
(ONIZATION VACUUM
GAUGE

VACUUM
THERMOCOUPLE
GAUGE

    
 

ROUGHING
VACUUM
PUMP
(4) A pulse line, OPR 0, to control the Calcomp.

(5) A coaxial line to carry analog spectrum information to the
LINC.

(6) Teletype communication lines.

(7) A signal line which carries the instruction to terminate

data acquisition.

The inlets for the quadrupole are diagrammed in Figure 2. The
gas chromatograph effluent is fed through a Biemann separator and
enricher into the ionization chamber of the quadrupole. A roughing
pump is used for the separator. The gas chromatograph also feeds a
conventional flame detector whose output goes to a strip chart recorder.
The recording is used only to monitor the run. For data processing, the
equivalent output is reconstructed from the summation (total ion count)

of mass spectra.

The second inlet to the ionizer is the vapor from a reference com-
pound, stored in a 500 milliliter flask at 0.5 - 1 Torr. The pressure
in the flask is maintained by manually operating the valves from the
liquid reservoir or roughing pump. The flow of the reference gas into
the mass spectrometer is enabled by a solenoid-actuated valve which is

turned on or off by the computer.

The interface built by the Instrumentation Research Laboratory
includes the digital register, digital-to-analog converter, and other
electronics necessary for the computer to control all of the third floor

equipment. It accepts computer commands to control mass position, inlet

10
of the reference gas, the teletype and the Calcomp plotter. In addition,
it provides a suitable signal for displaying the mass spectrum on an

oscilloscope. The interface has some manual controls for servicing.

ll
IV. The Software Structure

The objectives of the software are to (1) operate the controls of
a mass spectrometer, (2) acquire data from the mass spectrometer, and
(3) to process and present that data in a manner useful to the researcher.
The approach was to isolate the researcher (as much as possible) from
the apparatus in order that he might spend most of his time and effort

on sample analysis.

The software for this system is titled "SPECTRUM". SPECTRUM is a
central pivot for a collection of several programs. Its sole function
is to allow the user to call an operational program. The operational
programs are listed -~- along with a brief description of what each does --

in the SPECTRUM User's Guide, Table 1.

A program is called by typing its name on the teletype keyboard.
The operator then gives instructions to the computer (via the teletype)
to control the mass spectrometer, acquire data or present the data.
Each program requests at least one decision from the user which will
affect its course of action. Most of the programs also request input
and/or output parameters. As an example, a flow diagram of the program

called DISPLAY is given in Table 2.

12
(1)

(2)

(3)

(4)

(5)
(6)

(7)

(8)

(9)

(10)

(11)

(12)

(13)

Table 1

A List of Programs in SPECTRUM

LOCATE: Creates a new "N table". Octal numbers are accepted as "N" entries.
The user must give the corresponding mass (m/e) value as M. When two pairs
have been accepted, LOCATE computes the 254 other values by straight line
interpolation. This "N file" is filed under a name of the user's choice.

DISPLAY: Allows visual spectra presentation. Loads an N file from tape,

accepts either M or N as input, and positions the mass spectrometer gate
accordingly. It also causes an auxiliary sweep about this mass position to

cause a display upon an oscilloscope.

MEASURE: Types amplitude values at or about a mass peak. Loads an N file
from tape, accepts M or N, measures the amplitude for time T, and types
out. In multiple mode the amplitudes of the 8 or so N values on either side
of the specified N value are also measured.

TYPE: Allows inspection of various numeric files. Requests N file, reference
file, or spectrum identification and types the contents from entry 1 to 256 by
increments of 1, or any arbitrary variation of these three parameters.

GAS: Informs user of reference gas status, accepts on or off change if desired.

TWEAK: Adjusts an N file made by LOCATE, or previous TWEAKS, by accepting
a set of number pairs consisting of mass and the adjustment (error) at that
mass; linear interpolation is used between the mass numbers entered.

TAKE ONE: Takes a spectrum (only one), sampling for time T in milliseconds
at each mass number as specified by an N file. Resulting spectrum is stored
in a file from 0 to 256.

GOGO: Takes up to 256 spectra in succession over a period of up to 20 to 40
minutes, depending onthe time spent integrating at each mass; each of the
spectra is stored on tape starting in file l.

SUM: Sums the individual spectra produced by GOGO, stores the result, a
summation per entry, in file 0. The appearance of minus values in a spectrum
indicate an End-of-File block. The resulting sums are roughly equivalent to
the point values of a gas chromatograph curve as a function of time.

SORT: Searches the spectra produced by GOGO, picking up a specified mass x.
The resulting series of values represents the variation of mass x with time.
Five separate mass numbers may be selected per SORT.

PLOT: Plots any or all of the files of spectra in either a bar graph or line
plot. The plots are scaled so that the maximum value in the files selected
equal 100%. All other values are correspondingly smaller.

?: Indicates that the user is confused. A list of all programs available
under SPECTRUM is typed out.

WIPE: Indicates that the user is despondent. The computer, acting without
hesitation, wipes out all software relating to the mass spectrometer. The
tape will retain timing and block marks so that it may be used for other
purposes,

13
Step

Table 2

Logic Outline of DISPLAY

Description

Type < DISPLAY> (Note: <...> indicates a message typed
out by the computer.)
Turn on display sweep relay. Go to step B.

Type < FILE NAME = >, Then accept responses:
a. +,Disable display sweep relay and
return to SPECTRUM,
b. ‘less than four characters';, go to C.
c. ‘four or more characters';, then
1. if known file identifier, go to D.
2. if unknown file identifier, go to C.

Type < IMPROPER ENTRY> and return to B.

Known file identifier. Read the indicated N file into core.
Go to E,

Type <M OR N?> Accept the following responses:
a. t, go to B.
b. M;, note M state is true and go to F.
c. N3;, note M state is false and go to G.
d. anything else, go to E.

Type < M=> Accept:
a. +, go to E.
b. A decimal number within the range of
1 to 256. Look up the corresponding
entry in the N table. Consider this
the 'N value". Go to H.
c. Anything else, go to F.

Type < N=> Accept:
a. t, go to E.
b. An octal number within the range of
0 to 7777. Consider this the "N value".
Go to H.
c. Anything else, go to G.

Type the N value as <N= 'N value'> in octal notation. Also

put this N value in the external register for D-A conversion.
If M state is true, then go to F, else go to G.

14
Vv. Description of the Interface Electronics

A block diagram of the interface electronics appears in Figure 3.
Most of the circuitry is housed in a short relay rack, occupying about
18 inches of vertical space. An additional 10 inches of rack space is
used for the power supplies (+10 volts, +15 volts and 3.6 volts for the

integrated logic) and a precision reference voltage for the digital-to-

analog converter.

The logic is a mixture of "A" series circuit cards manufactured by
the Digital Equipment Corporation and low cost, epoxy, Fairchild
integrated circuits. The 12 bit digital-to-analog converter was con-
structed of DEC cards. The Amelco 2110 BE FET was used for the electronic

switches. The operational amplifiers were manufactured by Melcor; most

are of the economy line.

The LINC has a "fantail cage" designed to hold custom electronics.

The following input/output circuits were installed:

(1) A set of gates enabled by Relay Register 3 (in the LINC).
These transmit the contents of the accumulator to the remote

interface. Relay Register 3 is turned on or off by a machine

instruction.

(2) A set of drivers for the operate pulses. These are OPR 10,

OPR 11 and OPR O (LINC outputs).

(3) A resistor through which the signal current is passed and
converted to a voltage suitable for the LINC's analog-to-

digital converter.

15
oT

42 MANUAL
SWITCHES

 

 
 

  
 

 

 

 

 

 

 

  

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

    

 

 

 

 

 

 

 

 

 

  
 
     
        
 

 

  
   

SUMMING
MANUAL AMPLIFIER
LEVEL O-a
LINC ACC. (12 BITS)
Ce eed CONVERTER
Pl
LINC OPR 10 ULSE a
Ar 5x2 SWEEP WIDTH
(HP-33008)
@
MASS POSITION /7\
CONTROL TO Ji (BNC CONNECTOR ON REAR OF
2UOBE THE SWEEP 1S ONLY FOR QUAD CONTROLLER)
FET ENABLE VISUAL DISPLAY
esi/ AMELCO | SWEEF
SWITCH
Kl J TO CONTROL SOLENOIC SENSE SWITCH 5 ON THE LINC
LINC ACC. (6 BITS) ——*] 3 FLIP-FLOP ---/- OF REFERENCE GAS CONSOLE
REGISTER INLET SYSTEM toe 4
LINC OPR Ii 4 0 GO
2N08E
FET RESET
es2| AMELCO | SWITCH
SWITCH
sta
SENSING RESISTOR
y wv IN THE QUADRUPOLE
SIGNAL
MONITOR
fa X AXIS
| EvectRometer = | Y AXIS
| MOUNTED NEAR | CRO
MULTIPLIER |
CONVERTER,
| 1¢ Vv TOL
ELECTRON MULTIPLIER ls
OUTPUT OF THE
QUAD 300
mn
| FET OP AMP.
Lo J BUFFER ~ ate
— AMPLIFIER INTEGRATOR
INVERSION
& OFFSET
SAM 14 (LINC A-D CONV)
CALCOMP INTERCONNECTING CABLES WITH
LINC ACC. (6 BITS) ————+] 6.2-neut bt THE QUADRUPOLE
pe
LINC OPR O eH NAND OrGiTAL
PLOTTER

 

 

 

 

Figure 3

Block diagram of the electronics of the IRL interface between the
LINC computer and the EAI quadrupole mass spectrometer.
OPR 10 clocks all 12 bits (N) of the accumulator into the jam
register which feeds the digital-to-analog converter. The purpose of
this jam loading is to minimize the transients caused by changing N
values. If the switch Sl is in the manual position, the digital-to-
analog converter is under the control of 12 manual switches mounted on
the front panel of the relay rack. Above each switch is an indicator

lamp. These lamps monitor the contents of the converter at all times.

OPR 11 clocks the 6 rightmost bits of the accumulator into a 3 bit

jam register. The 3 bits of this register:

(1) Turn the visual display sweep of mass analyzer control voltage
on or off;

(2) Permits or blocks the entrance of the reference gas into the
mass spectrometer;

(3) Reset the analog integrator, AS.

OPR O transmits the Calcomp commands from the 6 rightmost bits of

the accumulator.

The high impedance output of the digital-to-analog converter is
buffered by amplifier Al. If the switch S2 is closed, either manually
or by the Amelco FET, Al sums the output of the digital~to-analog con-
verter with the output of a triangular wave generator. This provides a
visual display of the masses either side of the mass which is selected

by the output of the d-to-a converter.

The ground plane in the QUAD 300 Controller was unavoidably a few

millivolts different than the interface. Therefore, the output of Al is

17
converted to a current (A2) before it is sent to the control circuit of

the mass analyzer. Any ground loop ripple (60 Hz.) feeds the common

mode of A2 and, therefore, is rejected.

The current from A2 is reconverted to a voltage by a sensing
resistor installed in the quadrupole. This was the only internal

modification made to the mass spectrometer,

A3 is an FET operational amplifier (Melcor 1619) which is used as
an electrometer. It is mounted as close as possible to the electron
multiplier output of the quadrupole. In the configuration shown, it
has a bandpass exceeding 10 KHz. with 10 megohms feedback. The feedback
resistor is selected to be 1, 10 or 100 megohms by a manual switch.

The normal setting is at 10 megohms. A4 picks up the electrometer out-
put at the relay rack in the differential mode, again to reject 60 Hz.
ground plane ripple. A4 drives the vertical axis of a Tektronic 502
Oscilloscope and feeds an analog integrator (A5). A Hewlett-Packard
3300A Low Frequency Function Generator sweeps the horizontal axis of

the 502 at about 5 Hz. The 3300A is also the source of the scan for the

control voltage of the mass analyzer during visual presentation.

A5 integrates the output of A4 (across a 0.01 microfarad capacitor).
Its output is converted to a current and sent over about 200 feet of

cable to the LINC analog-to-digital converter.

A front panel push button, labeled "STOP GOGO", enables the operator
to stop the data gathering program by that name at any time during its

run.

18
VI. The Software and Its Concepts

The software for the computer-mass spectrometer system: (a) allows
the user to acquire mass spectra of samples or a reference gas, and
(b) provides information on the stability and precision of the instrument
itself. The latter consideration stimulated the writing of software
that might not be required for chemical application of the system but is
useful in determining the reliability of the system and necessary to the

development and maintenance of the instrumentation.

This mass spectrometer system utilizes the interaction, in real
time, of the three major components. These components are: the computer,
an instrument, and a human operator. It is apparent that the human
operator is the slowest component, and in a sense the limiting factor in
acquiring spectra. However, he has the greatest range of decision-making

ability. This system takes advantage of this ability.

The computer-spectrometer system does not demand the contiguous
location of all three components. As noted before, the computer was
located three stories below the mass spectrometer. Most of the software
could be operated from either location or indeed a third location.
Though the mechanics of introducing the unknown chemical sample were
not automated, the reference gas was valved into the instrument under

computer control.

There exist two levels of communication between the operator and
the control system. The first (SPECTRUM level) simply accepts any one of

a set of eleven subprograms which then specifies the further operation

19
 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

SPECTRUM
GAS worst LOCATE
v ™,
‘ ~
\ ~~
\ “SA
\ ™S
\ AL
\ ~S N File
1 .
\ ~S
\ _
\ “>
‘ .
‘ be
‘
‘
\
‘
\
DISPLAY MEASURE
Calroration
Error information on Error information on
the N File tne N Frie
TWEAK
TAKE ONE
GO 60
Uata
Acaursition
SUM SORT
Det.
Manioulation
Chromategram)
TYPE PLOT
ata
Presentation =]
Listing Listing
of of Mass
N File Spectra
tL

 

 

 

 

Listing of

Gas Chro-
matoqram

Mass Spectra

Bar Graphs

 

FIGURE 4

SYSTEMS FLOW CHART

 

20

 

 

 
of the system. Control is returned to this level upon the completion

of any one of the subprograms.

Level two is that of the subprograms. It is within these sub-
programs that the necessary parameters are accepted for the performance
of the individual tasks. These eleven subprograms can be considered in
four groups according to the functions they are related to: Calibration,
Data Acquisition, Data Manipulation, and Data Presentation. These four
groups are indicated to the left of the systems flow chart shown in
Figure 4. The figure illustrates the way the calibrate programs are
used, The flow is under operator control; the lines indicate the normal
paths. The result of the calibration is an accurate list of 256 numbers
(an N file). These are related in a one-to-one fashion to mass numbers,
so that when a number in the N file is gated into the D-A converter
the resulting analog voltage allows the corresponding mass to pass
through the mass analyzer. Thus a signal is available at the electrom-
eter proportional to the abundance of that mass/charge ratio. The 256

elements in the N file correspond to mass/charge ratios 1-256.

A. Calibration

GAS allows the operator to turn on or off the reference gas. It is

this reference gas that is used to develop and adjust the N file.

LOCATE creates a first approximation to an appropriate N file. It
accepts two sets of coordinates each consisting of a mass number and its
corresponding N value. By simple linear interpolation (and extrap-

olation) the remaining 254 elements in the N file are filled. The

21
N values corresponding to adjacent masses differ by about 15. This
allows sufficient sensitivity in setting the mass analyzer to insure

sensing the mass peak at its maximum rather than on the shoulders or

sides.

DISPLAY provides a method for the operator to visually examine
discrepancies between the N file entries created by LOCATE and the
real position of the reference gas peaks. By entering a mass number,
the corresponding point of the spectrum is displayed on the center of
an oscilloscope. (The user can also display a point by entering any
N value.) Thesurrounding portions of the spectrum are displayed under
the control of a potentiometer. If the peak is not precisely in the
center of the display, the N value calculated by LOCATE is not quite
correct. The increment of error can be determined by typing in new N
values until the display is correct, or by the MEASURE routine. Table 2

is a detailed flow diagram of the DISPLAY dialogue.

MEASURE samples and digitizes the electron multiplier output at a
specified point or series of consecutive points. These amplitudes are
typed out with their corresponding N values. The highest amplitude

will indicate the center of the peak.

TWEAK accepts corrections to any N file. This may be a new file
created by LOCATE or it may be one used previously that is in need of
minor adjustments. These corrections can be entered for any or all of
the reference gas peaks. The N file is then updated by the entered
values at the designated points and by interpolation between the

designated points.

22
Calibration of the system has been achieved using the four preceding
programs in roughly fifteen minutes. An N file thus developed has
been used with only minor modifications for a number of days. It appears

to be as accurate as the resolving power of the mass spectrometer.

TAKE ONE provides a final check of the calibrating procedure. It
uses the N file created by the operator to position the sensing gate
at sequential mass positions from 1 to 256. The fon current is sampled
and digitized for each position. The resulting spectrum can be plotted

or typed; it should match the known reference gas spectrum.

B. Data Acquisition

GAS is used to turn off the reference gas used in the calibration

procedure and provide a delay for the residual to be pumped off.

GOGO starts taking spectra at the operator's signal, continues tak-
ing spectra at intervals specified by the operator (5 second nominal),
and stops on a signal from the operator or at the end of 256 spectra.
Each spectrum consists of amplitude measurements at each mass value,
1+256. These spectra are stored on tape sequentially. No attempt was
made to automate the starting or stopping of the data gathering, since
the variability of the diffusion rate through the gas chromatograph
for different substances made operator control desirable. The time
scale of the operation, 10 to 25 minutes, was also appropriate for

human interaction.

C. Data Manipulation

SUM adds up the amplitudes of each spectrum to produce a record

23
(a set of up to 256 values) of total ion current. This record is in

fact a reconstructed gas chromatograph recording. It provides a quick
check on the total ion peaks seen by the spectrometer and compares
closely with output from the flame detector of the gas chromatograph.
When the values are plotted, the scale on the horizontal axis constitutes
an index to the tape locations of the individual spectra taken during

the data gathering (see Fig. 6).

SORT searches through each stored spectrum picking out the amplitude
at a specified mass position. The result may be plotted to show the

variation of the ion current at a specific mass with time.
D. Data Presentation

PLOT is a special purpose plotting routine for displaying results
as a line or bar graph. Any spectrum or the results of the SORT or SUM

programs may be plotted (see Figs. 6 through 10).

TYPE is a general purpose listing routine. It lists as four digit

integers any of the files or spectra produced by other programs.

This second level of communication between the operator and computer
involves the entry of parameters for the specific programs. The
parameters are requested individually by the program in a prompting
fashion. This type of demand-response entry of parameters is very
convenient from the user's point of view, although it is tedious to
program. It allows a relatively inexperienced user to make use of the
system without memorizing long sequences of instructions. We found that

chemists with no previous computer experience could, and indeed would,

24
make use of the system without hesitation. Since the chemist is
expected to play an integral part in the feedback system, the acceptance
of his role is important. In the following section, an example of

computer-operator dialogue is given.

The programming of instrument control is quite simple with the LINC
computer. One instruction commands are available for the digital out-
puts (OPR n), and for the analog input (SAM n). Sensing the level of
the digital input is a simple skip instruction (SXL n). Appropriate
timing loops between these instructions are all that is required. This
part of the software would require modifications for use with other

systems, since it is designed specifically for the hardware involved.

Since the signal is converted to digital form by an 8 bit converter
it is desirable to take full advantage of these 8 bits without saturating
the converter. We achieved this by trial and error with each unknown sample.
Appropriate manipulation of the integrating time by the computer could

achieve the same result in a much more elegant manner.

More flexible control over the amplitude of the peaks could have
been achieved by using the analog integrator in a different fashion.
The amplitude of the signal seen at the computer varies with (a) the
gain setting on the electrometer, (b) the amount of material injected
into the spectrometer, and (c) the length of time the integration is
continued. Since unknown samples can be of unknown concentration the
amount of material injected is inherently variable. Altering the gain

in a digital fashion would be difficult. The length of time for

25
integrating can be varied with great precision by examining the signal

at frequent intervals.

At the instruction level of the computer, the programs consist of
short mainline routines with an assortment of 50 or 60 subroutines.
These subroutines vary in length from 10 instruction error checks to the
512 instruction arithmetic routines. The memory of the LINC is quite
limited, 2048 twelve bit words, of which only 1024 words are programmable.
The software was organized to allow extensive overlaying. The resident
portion of the program requires 200 words of memory. The remaining
12,000 words of the system are read in from tape as needed. In order
to improve the efficiency of the running system each of the subprograms
is self-contained. This necessitated duplication of certain subroutines
for each program, but allowed the majority of memory to be available when
needed for high speed processes. Although this system required about
12,200 instructions on the LINC computer, it might well be duplicated
in 8000 instructions on a computer where the subroutines (without
duplication) could remain in core. An incentive to conserving the core
memory used in this system was the desirability of extending the mass
range to 500. This extension was not attempted with the quadrupole

spectrometer, but the LINC memory would not have been a limiting factor.

The following diagram of Figure 5 illustrates both the partitioning
and flow of a hypothetical routine under SPECTRUM. The program select
table, teletype routines, and parameter table are resident in core. The
number of partitions within programs vary from one to four; the program
in the diagram has two partitions.

26
 

Start Resident <
Portion

|

Read Tape

 

 

 

 

Computer-User
Communication

Partition 1

 

 

Read Tape

Computer Instrument
Partition 2 Communication

 

Write Data Tape

 

 

 

 

Return to Resident
Portion

Figure 5

Partitioning and Flow of a Routine under SPECTRUM

27
VII. AN EXAMPLE OF USER-COMPUTER DIALOGUE

In this section a typical dialogue between the user and the computer
system is reproduced. The example will probably give the reader a
quicker feel for the flow of the communication. It is presumed that this
is a "cold start"; there was no prior calibration. The example goes
through the complete calibration phase and a data acquisition from a gas
chromatograph run.

The teletype listing is reproduced on the left. Normally most lines
have some output, but here lines are skipped to keep the comments in
step with the teletype listing. The comments are on the right.

All computer output is left as it appeared. The user's typing is
underscored for identification. One other user key used was the < rubout>.
This leaves no character on the paper; the program simply ignores what was
on that line and advances the paper and repeats. Where rubout was used,
the legend '< rubout>" has been inserted.

Teletype Listing: Comments:
SPECTRUM "SPECTRUM" was typed by the computer. This indicates it is
ready to execute any program stored in SPECTRUM.
BAS3 The program selected by the user is "GAS".

REFERENCE GAS IS OFF)

DO YOU WANT IT ON 7YESS In this case the user indicates he wants the reference pas

WAIT 15 SECS. valved in. The computer gates it into the mass ionizer
and instructs the user to wait for the gas pressure to
stabilize in the analyzer source.

SPECTRUM

LOCATEs The next program the user calls is "LOCATE". LOCATE
illustrates the method of initially establishing an "N
Table". The computer must be given the known N value and

LOCATE m/e identification for two known masses. LOCATE also
enables the display of the spectrum on the oscilloscope,
centered about the last N which was entered.

N= 16003 The first guess is 1600. (This can also be done manually
by switches.) This causes the spectrum to be displaved
centered about 1600.
Next the computer asks for '"M’, the m/e value. Appar-

Mel ently the operator is not ready to identify the peak, or
the peak is not centered.
Thus the operator answers with an upward arrow (+).

N= 16163 This causes the program to ask the previous question,

Me r Ne!

28
M= 693

N= 490@3

M= ft

N= SAggs

M= 449903 IMPROPER ENTRY
Met

N= 44003

Me ¢

N= 4509903

M= 4600: IMPROPER ENTRY
Ms ft

Ne 46003

M= ft

N= 45663

M= fF

N= 45703

Me *

N= 45733

M= 457"

Ne 45723

M= 1693

:

 

 

 

 

 

 

N= 1646 M= 6869
N= 4572 M= 6169

ACCEPT? YES3

FILE NAME = REPORTS3

SPECTRUM

DISPLAY3

DISPLAY
FILE NAME = REPORTS;

M OR N2MS
M= 693 N= 1606

N= 4572

M= 1003 N= 2537

The operator guesses 1610 and then is satisfied with 1606.
The entry 1606+ is rejected by the program as meaningless.
It again asks for N.

Finally at N=1606 the operator is satisfied that this is
the best value for m/e = 69; thus he answers 69 to the
query M=.

The guess and try method identifies 4572 as the position
for mass 169.

Mass 4400 is beyond the range of the mass spectrometer.

After the second identified M, the program types the
parameters for verification and asks for acceptance. If
the user had indicated non-acceptance, the program would
have backed up to the beginning and allowed the work to
be redone.

The program computes the other 254 N values from the
relationship, N = k*m/e + b. k and b are computed from
the two known points. This file of 256 values mist be
stored in a manner so that it may be called for use
whenever desired. The program asks for a FILE NAME.
The user responds with "REPORTS".

This N Table is now available under the identifier,
"REPORTS".

The program returns the user to "SPECTRUM", level one.

The user now will call DISPLAY to verify, or determine the
errors in N Table. DISPLAY only displays the spectrum on the
oscilloscope. The display center is the N value output

from the computer. The manual width control is typically

set so that integer peaks have a separation of one centimeter.
This is the user call.

The computer responds. DISPLAY program is now active.

What file? User types "REPORTS", and the program loads that
table.

Query. Does the user wish to direct by N entry or by M
entry? (The computer looks up the N entrv.) User answers

that he wishes to direct by mass entry.

He asks for mass 69. The program looks up the table, finds
1606 and outputs it to the mass spectrometer interface. It
then types the N value for operator information.

Same with 169. Note that these are the two points piven

in LOCATE.

Now mass 100 is tried. Apparently the user is not too
happy with this; the peak does not seem to be exactly
centered.

29
s

{«

m OR NINS

we 25423

ws 254493

we or

N

No

= 2542
23544

So now upward arrow causes the program to back up a
question,

The user indicates by '"N" that he wishes to enter N
numbers directly.

In two more tries he finds that mass 100 is at N entry
2544.

Backs up again, and this time writes himself a note about

™ OR N? NOTE 2544 GODD AT Mass 199 mass 100,

m OR N?MS

33 <rubout>

FILE NAME = REPORTS3

 

New
NS
Ae
N=
N=

N=

Ns

Ne

N=

N=

<

N=

N=

N=

N=

New

sSyv

MW OR N?P
FILE NAME =
SPECTRUM
DISPLAY3
DISPLAY
m OR N?7Ms
m= 1003
m= 383

= a5
m= S65
m= 695
m= 1003
ms 1193
= 1313
me 1563
ms 1693
m= St
mx 1813
m= 1973
= 2193
M= 2313
ma 24aTe
M= +

m OR N?L

FILE NAME =

SPECTRUM

Lt

2537
@426
1067
i145
1606

2537

3208

346 7

4136

4572

>
$e6e@
$444
6164
6452
7836

+

This is meaningless to the computer, so the program
retypes "M OR N"?. This cime the user answers M.

Now let us recap what happened. The N entry at mass 100
is 5 off. (All N numbers are in octal.) The error is
+5, since it is necessary to add it to the entry in the
REPORTS N Table.

In typing this note, the user overran the buffer. The
SS is meaningless. Doing a<rubout> does no good.

The user can recover. The program responds to + by going
one level nearer to SPECTRUM.

Three upward arrows get the user clear back to SPECTRUM.

Now DISPLAY is called again. The program is reread from the
tape. The user goes through the list of reference peaks,
checking each one.

Between 169 and 181 he makes a typing error, but notices it
before typing a semicolon. A<rmbout>, one of the teletype
keys, cancels the line.

Presumably the user has noted the position errors, if any,
for each listed mass. These errors will be needed soon.
The actual process, similar to that shown for mass 100,
takes about an hour for a new setup. For day to day check-
ing, it takes only about 10 to 20 minutes.

Upward arrows bring the user back to SPECTRUM.

30
MEASURE

 

MEASURE
FILE NAME =
SINGLE? NOPE$3
M= 1813 ~
= 253

N AMPs
5031 agae
545¢ AaTa4
5451 @161
5052 B207
5953 B223
5054 A236
5855 @241
5056 8235
5057 a217
S869 9173
S61 0134
5962 ai14
5963 O25?
5964 0036
5465 0023
5066 2233
5867 4011
5a70 8214
5107 8207
Mat
M OR N?P
SINGLE? +

FILE NAME =

 

 

SPECTR
TWEAKS
TWEAK
M= 283
S= 1123
= 473
Ms 593
Fe 153
M= 1063
k= 53
Ms 1193
R= 103
M= 1313
k= 73
M= 1593
E= 33
M= 1693
R= G3
Mz 1813
R= 233
M= 1973
E= ~73
M= 2193
R= -153
M= 2313
f= -173
M= 24s
R= 7243
M= 2000

3

M

wo

 

"MEASURE" may also be used to check the N Table.
Here it is called and the N Table "REPORTS" is requested.

Multiple readings are desired.
At mass 181.
Each reading is to be integrated for 25 milliseconds.

These are the results. For a span equal to about that
between two integer mass positions, each N number is tried.
Also two points, 1.5 mass divisions up and down, are
included to give a good base reference. The results are
typed as shown. Inspection here shows that mass 181 is
actually centered about N=5055. This is in contrast with
the N#5060 in the REPORTS table. Hence the error is -3.

Again, upward arrows bring the user back to SPECTRUM.

The program TWEAK is now called to correct the N file.
The user types in pairs of numbers, masses and errors that
he has determined at those masses.

Mass 100.
Error in the table of REPORTS for mass 100 is +5.

When the user has completed his list he signals that fact
by typing 2000.

31
ERROR

28

aT

5@
180
119
131
1598
169
181
197
219
231
247

ENTRIES

8112
8832
@815
8095
00198
aee7
8863
F777
7772
77798
7762
7760
7753

ACCEPT? YES;

FILE NAME=REPORTSs

CALIBRATION COMPLETE

COMPLETED The program notes this and

recypes the entries for verification.

The program then gives the user a chance to accept or
reject. Upon acceptance, the program prepares to use
these points to make corrections to an N Table.

It asks what N table to correct. REPORTS is the answer.

Then the N table and the errors are merged to form a new
corrected table. Mass positions between calibrated points
are corrected by straight line interpolation of the error
function.

The riew corrected file must be put somewhere. This time the
query "FILE NAME=" is for this purpose. The user responds
with "SUPER". Apparently this file name has already been
used,

FILE NAME=SUPERs OV ERWRITE?GQO AHEADs

 

 

 

SPECTRUM
DISPLAY3
DISPLAY

FILE NAME =
Mm OR N?M3

Mz 283 N=
M= 323 Ne
M= 405 N=
M= 505 N=
M= 2313 Nex
M= 693 Ne
M=x Tf

M OR N?NS

N= 6143 N=
N= 6163 N=
N= 77@3 N=
N= 7623 Na
N= 6

N= 7643 N=
N= 7653 N=
N= ft

M OR N22
FILE NAME =
SPECTRUM

SUPER;

8548
@622

8770
1162
6 433
1617

B6t4
0616
6778
8762

@764
@765

£

Reuse of this name causes whatever is in SUPER to be
erased. The program will not do this unless expressly
directed to. That is why it asks "OVERWRITE?". The
answer is affirmative, so old SUPER is erased. Now the
new file is in SUPER.

The program gives control back to SPECTRUM.

The user has recalled DISPLAY.

And che file SUPER.

He is investigating the behavior around masses 28 to 50.

He goes back to SPECTRUM. He apparently has noted more
errors.

32
 

ERROR ENTRIES COMPLETED

17777
28 7777
32 7773
49 7774
$08 7777

255 7777

ACCEPT? YESs
FILE NAME=SUPER3
CALIBRATION COMPLETE

 

Calls TWEAK again.
Note that he says no error at 1 and 28. This will insure

that no changes are made at those inclusive masses.

He did note errors at 32 and 40.

But 50 through 255 is pegged by specifying these errors
to be zero.

And this is typed out for verification. (The notation is
octal, one's complement.)

The correction is made to SUPER.

FILE NAME=SUPER3 OV ERWRITE?YESs

SPECTRUM
DISPLAY3

DISPLAY

FILE NAME = SUPERS
M OR N?M3

M= 283 N= 8548

M2 323 N= @6 16
M= 503 N® 1162
M= 403 N= 9765
M= 1313 N= 3476
M= 2193 Ne 6147
Mat

M OR N?S

FILE NAME = ¢
SPECTRUM

TAKE ONES

TAKE ONE

FILE NAME = ft

T= 173

FILE NAME = SUPER:

SPECTRUM FILED AT 1.
REPEAT? NO3

The new table is filed under SUPER. (The old one has
been erased.)

Now DISPLAY is called for a final verification.

All seems well.

Return to SPECTRUM.

TAKE ONE will read each mass point from 1 to 256. The time
for integration at each peak is remembered from the last
time TAKE ONE was used. To change this parameter, upward
arrow must be used here. The time is set to 17 milliseconds
per peak.

The N Table to be used is SUPER. Now the spectrum is

taken.

In about 3 seconds it is complete.

The user may elect to take another.

Since the reference gas has been left on all this time, the

plotted results of this would be similar to first illustra~-
tion in this report.

33
SPECTRUM

GAS3

REFERENCE GAS IS ON

Now GAS is called.

It reports its condition. And the user requests that it be
turned off.

DO YOU WANT IT OFF YES PLEASE;

WAIT 15 SECS.

SPECTRUM

GOGO;
GOGO

MOUNT FABER UE EEE

MSEC PER PEAK=173

SPECTRUM

SuUMs

After the due wait for the system to pump out, che program
returns the user to SPECTRIM.
The system is now fully calibrated.

GOGO is the program that will take a sequence of mass
spectra during a gas chromatograph run.

The N file must be specified.
At this point the user may mount a new tape if he wishes
to retain the data on magnetic tape.

And last the milliseconds per peak integration is

specified. Actual data acquisition starts with the semi-
colon after the "17". This continues for about 18 minutes.
The program would have come to a normal halt at the end

of about 21 minutes. However, the operator was watching a
strip chart recording of the signal from the flame

detector of the GLC. He noted that the last peak of interest
has passed and pressed the terminal button.

Control is passed back to SPECTRUM.

SUMMING TAKES TIME WAIT 5 OR 10

FILE# @. #OF POINTS=@217 SUM is the program that adds up each individual spectrum.

SPECTRUM

 

x
o
-—
a 8
z
e
a
i

 

X DIMEN.= 163

Y DIMEN.= 73

BAR GRAPH? NO3

-A

There were 217 spectra. The sum of each was put in file 0.

PLOT causes Calcomp plots of the results.

This call of PLOT plots the result of SUM.

This and the next entry are in error.

Typist correcting his error.

The error is corrected.

The plot is to be 16 inches wide.

And 7 inches high.

Since this is a total ion plot, it is the equivalent of a
GLC chart. The numbers on the x axis refer to the spectra.
Since a spectrum was taken every 5 seconds, this axis is
also a measure of time. Also, the plot is a good index to
the spectrum of interest.

The plot is shown in Figure 6.

34
SPECTRUM
PLOTs

NAME: NCS- ALANINE METHYL ESTERS

FILE #Si 163

J TO K BY [.

J= 203

Ke 25903 ©

Iz 15

X DIMEN.= 155
Y DIMEN .= 63
BAR GRAPH? YES}

This is the first plot of a mass spectrum of this group.
As can be seen in Figure 6, file 16 indexes the first GLC
peak. The operator, from his knowledge of the GLC plot,
identifies and titles it.

This is shown in Figure 7,

SPECTRUM
PLOTS
NAME: NCS-METHIONINE METHYL ESTERS

FILE #S: 1563
J TO K BY I.

J= 283
K= 2563
I=

X DIMEN.= 153

Y DIMEN.= 63
BAR GRAPH? YUSPS

SPECTRUM
SORTS

SORT SPECTRA FOR MASS

MASSC!)= 1593
MASS(2)= 1433
MASS(3)= 1453
MASS(4)= 2933
MASS(5)= 2653
WAIT...

STORE IN FILES 8507s

SPECTRUM
PLOTS

NAME: MA
FILE #S: 5093

J TO K BY I.

Js is

K= 2403

I= i

X DIMEN.= 163

Y DIMEN .= 65

BAR GRAPH? NOPES

SPECTRUM
PLOT

This is another of the 217 spectra available. The program

asks the user for certain items. First is the Name. This

becomes the title of the plot. Then the range of interest

(J to K). The size of the plot and if it is to be a bar or
connected line plot. This "YUSP" was interpreted correctly
as "YES". See Figure 8.

SORT will go through the spectra and get the time history
of the amplitude on any 5 mass peaks.

In this case mass 159, 143, 145, 203 and 205 are requested.

@508. 6509, 0510, @S511.

>
File 509 contains the time history of mass 145 (see above).
This, as can be seen in Figure 7, appears to be the
molecular peak of NCS-Alanine Methyl Ester. The connected
line graph is shown in Figure 9.

NAME: MASS 203 (NCS-ASPAR] <rubout> The user corrected himself.

NAMES MASS 293 (NCS-ASPARTIC ACID DIMETHY) ESTERDS
FILE #S: 5193 This is a similar plot of mass 203 shown in Figure 10.

35
J TO K BY I.

J= 1s

K= 2403

I= ts

X DIMEN.= 163

Y .DIMEN.= 63

BAR GRAPH? NOT NOWs

SPECTRUM

About 30 to 45 minutes have elapsed since the end of the
GLC run. SUM took the longest time, about 12 minutes.
The plots took about 4 to 5 minutes each.

All of this control, listing, and plotting took place in
the laboratory right next to the mass spectrometer.

36
GLC OF NCS-AMINO ACID METHYL ESTER MIXTURE

AMPLITUDE (3)
ape

FILE 0

 

. i me tbe tin the ihe” 100 ake” ibe ste tbe abe aha zie oe 2 48
TIME

Figure 6

Each unit on the abscissa represents five seconds. There
was one spectra taken during each such unit. The ordinate
is the sum of all mass peaks or total ion current. Hence,
this is an index to the mass spectra recorded and a GLC
representation.

NCS-ALANINE METHYL ESTER

AMPLITUDE (7)

 

hi Lube at
b oe 1be 110 120 190 140 1p 10 178 108 100 20 710 220 238 248 260
MASS

Pp pp op ape

 

FILE 16

Figure 7

The spectra of the 16th time slot of
Figure 6, the first GLC peak.

37
FILE 156

FILE S09

AMPLITUDE (3)

AMPLITUDE (5)

NCS-METHIONINE METHYL ESTER

 

 

 

&
+
t
ta
+
by
+
Ey
+
a
MASS
Figure 8
A spectrum from the third peak from the end. As
can be expected from the longer retention times,
this fraction has a higher molecular weight.
x MASS 145 (NCS-ALANINE METHYL ESTER)
a
2
i
3
3
+
%
cs
*

 

 

Figure 9

The occurrence of mass 145 versus time.
The time scale is the same as Figure 6.

38
FILE 127

AMPLITUDE (5)

FILE 510

AMPLITUDE (3)

MASS 203 (NCS-ASPARTIC ACID DIMETHYL ESTER)

 

 

 

  

%

a

3

By

2

2

x

x

=

%

ob ba SS NS le ile the ake ime ile ike ate ile” le che ain” che” De” we
TIME
Figure 10
A time plot of mass 203.

. NCS-ASPARTIC ACID DIMETHYL ESTER
&

2

&

FS

BS

+

PF

2

>

4

 

 
 

MASS

Figure 11

Another spectra indexed by Figure 6.

39
VIII. Conclusion

This system demonstrates the use of a computer in the operation
of a mass spectrometer. By thoughtful programming techniques, a complex
arrangement of hardware is made easily accessible to a user. The
operator-computer dialogue is particularly useful in the processing and
presentation of data. The software of the system significantly reduces
the time required for sample analysis, releasing the researcher to

explore the potential of the mass spectrometer..

The calibration of a mass spectrometer provides a reference for
measuring the precision and stability of the instrument. Computer
monitoring of the operating parameters of a mass spectrometer will pro-
vide data needed to improve these characteristics. We are working on
ways to make the calibration more automatic, limited only by the time
constants of the mass spectrometer. This would involve using a given
N file to take a spectrum of the reference gas. The computer would then
compare the spectrum data to a manually entered spectrum of the
reference gas which would be filed on tape. After comparing the two,
adjustments would be made to the N file. This would permit a calibration

before every GLC run.

It is feasible to build a limited purpose computer and a
quadrupole mass spectrometer as one unit, forming a compact instrument
with features similar to the system presented in this report. Some of
the software could be embodied in hardware circuits. Alternatively,
the system could be interfaced with a larger time-shared computer.

We expect to investigate this possibility.

40
(1)

(2)

(3)

(4)

(5)
(6)

(7)

(8)

(9)

(10)

References

R. A. Hites and K. Biemann, "Computer Recording and Processing of
Low Resolution Mass Spectra", International Mass Spectrometry
Conference, Berlin, September 1967.

R. A. Hites and K. Biemann, "A Computer-Compatible Digital Data
Acquisition System for Fast-Scanning, Single-Focusing Mass
Spectrometers", Anal. Chem., 39, 965, 1967.

H. C. Bowen, D. J. Shields, and H. M. Stanier, "Experience with
Fast Processing of High-Resolution Mass Spectra using Direct
Transfer of Data to an On-Line Digital Computer Without Inter-
mediate Recording", International Mass Spectrometry Conference,
Berlin, September 1967.

C. Merritt, Jr., P. Issenberg, and M. L. Bazinet, "Design and
Performance of A Direct-Coupled, Real-Time, Spectrum Digitizer and
a Rapid-Scanning, High-Resolution Mass Spectrometer", International
Mass Spectrometry Conference, Berlin, September 1967.

C. A. McDowell, Ed., Mass Spectrometry, McGraw-Hill, New York, 1963.

B. Halpern, J. W. Westley, E. C. Levinthal, and J. Lederberg, "The
Pasteur Probe: An Assay for Molecular Asymmetry", Life Sciences
and Space Research IV, M. Florkin and A. Dollfus, Eds., p. 239-249,
North-Holland, Amsterdam, 1967.

B, Halpern, J. W. Westley, I. von Wredenhagen, and J. Lederberg,
"Optical Resolution of DL Amino Acids by Gas Chromatography and
Mass Spectrometry", Biochem. Biophys. Res. Comm., 20, 710, 1965.

"Cytochemical Studies of Planetary Microorganisms - Explorations in
Exobiology", Status Report Covering Period April 1, 1967 to

October 1, 1967, IRL-1061, Instrumentation Research Laboratory,
Genetics Department, Stanford University.

"LINC - Programming and Use, 1", Vol. 16, Washington University,
Computer Systems Laboratory, 724 South Euclid Avenue, St. Louis,
Missouri.

H. L. Friedman, H. W. Goldstein and G. A. Griffith, "Mass
Spectrometric Thermal Analysis of Polymer Decomposition Products",
Conference Proceedings of Fifteenth Annual Conference on Mass
Spectrometry and Allied Topics, ASTM Committee E~-14, Denver,
Colorado, May 1967.

41