-
Microprocessor automation a UV-visible monochromator
Malcolm W. Warren, James P. Avery, Daniel W. Lovse, and Howard
V. MalmstadtUniversity ofIllinois, School of Chemical Sciences,
Department of Chemistry, Urbana, IL 61801, USA.
IntroductionIsolation of selected wavelengths of light with high
resolutiongrating monochromators is a common operation in
spectro-chemical methods of analysis. Commercial monochromatorsand
monochromator controllers are often unable to controlthe
monochromator’s operating parameters with the resolu-tion required
in automated systems. If a monochromator isto be used in a
completely automated system its controllershould provide a number
of important capabilities: initialis-ation of wavelength and
slitwidth operation; calibration ofthe wavelength and slitwidth;
the ability to select any wave-length available; the ability to set
any slitwidth available;provide operator interaction directly
(stand-alone mode);and permit interaction with other
microprocessors in themeasurement system (slave mode).
The controller should be extremely reliable for unattendeduse.
In addition, the controller should also be able to operatein high
noise environments without loss of accuracy andshould not be the
source of noise that will affect the operationof other parts of the
instrument. The EU-700 (GCA/McPher-son, 530 Main Street, Acton, MA,
01720, USA.)monochro-mator [1] was used in this study because it
can be readilymodified to provide these capabilities.
InstrumentationThe monochromator controller uses both
closed-loop andopen-loop control systems. In both the. closed-loop
and theopen-loop systems control signals are sent out by the
micro-processor, and are modified by the control board to
providesignals that can operate the electromechanical components
ofthe monochromator. In closed-loop systems the signals
originating in the monochromator are modified by thecontrol
board to provide feedback. The movements in theopen-loop control
systems used in this controller are assumedto be so accurate that
feedback is not required.
Wavelength driveThe wavelength drive is a sine-bar and precision
leadscrewassembly driven by a slew motor for large wavelength
changes(9 nm/sec) and a stepping motor for scanning and fine
wave-length adjustment (zero to 2 nm/sec). The stepping motor
isnormally engaged. The slew motor is engaged using an acsolenoid
when the slew motor is on.
Stepping motorThe stepping motor used in the wavelength drive
(No. 36D300 R7.2, Haydon Switch and Instrument, Inc. 1502
MeridenRoad, Waterbury, CN, 06705, USA.) provides
reproduciblewavelength increments of +/-0.01 nm/step at speeds up
to200 steps/sec. A four-phase signal is required by the
steppingmotor coils for forward and reverse movement. When
design-ing the driver circuits like those in Figure 1, the
designermust consider the inductive nature of the stepping
motorcoils. To eliminate the turn-off transients, diodes are usedto
shunt the current and protect the drive transistors.
Slew motorThe slew motor (Hi-torque motor, SPEC 1218, Multi
ProductsCompany, 2052 Grove Avenue, Racine, WI, 53405,
USA.)provides rapid movement of the wavelength drive
betweenwavelengths. An alternative to the slew motor used in
thismonochromator design would be a high-speed stepping
+5
IOKSTEP
LIMIT+51
IOK’STEP
LOWER LIMIT
Figure 1. Stepping motor drive circuit
+52v
15v COIL DRIVER-A
+5 IN914ioo
COIL DRIVER- B
CO’L DR’VER-C ICOIL DRIVER-D !_
STEPPINGMOTOR
76 Journal of Automatic Chemistry
-
Warren et al Automation of a UV visible monochromator.
motor. This would be less attractive because high-speedstepping
motors tend to be large, run hot, and require largepower supplies.
In general high-speed steppers are muchslower (3 nm/sec max speed)
than the slew motor (9 nm/sec)used in this monochromator. Major
modifications to themonochromator would have been required for
mounting themotor. Venting the heat produced is possible but
couldcreate problems with the optical system.
The slew motor is an ac induction motor with shadingcoils and a
squirrel-cage rotor [2]. The application of ac linevoltage to the
motor and the selection of the proper pair ofshading coils presents
a number of design problems. Ifmechanical relays are used to
control the slew motor power,the contacts arc, and noise spikes are
created on the logicpower supply lines. These noise spikes cause
errors in thecontrol logic leading to the loss of wavelength
information.A solid state relay, with photoisolation and
zero-crossingswitching (S-312 Crydom, 1521 Grand Avenue, E1
Segundo,CA, 90245, USA.) is used to provide power to the slewmotor,
eliminating this problem.
The shading coils are connected in pairs, the directionof
movement being determined by which pair of shadingcoils is shorted.
The shading coils present a special problembecause they operate at
9 volts ac peak to peak at 1.75 amps.A photoisolated solid state
relay, operable at these low volt-ages and high current, is not
presently available. However,the RCA SK3506 triac can conduct at
low ac voltages andhigh currents, and requires low gate currents.
It also providesgood isolation from high voltages when off. Thus,
while themonochromator controller is not completely isolated
fromthe shading coils, the low voltages involved and the
zero-crossing switching of the triac combine to prevent the
creationof noise spikes. The combination of a photoisolated
solidstate relay on the main coil and triacs on the shading coilsas
shown in Figure 2 has eliminated the noise spikes and re-sulting
errors in the wavelength counter due to slew motoroperation.
Slitwidth driveThe slits are straight knife edges adjustable
from 5 to 2000/am.The entrance and exit slits are connected to a
single controlshaft. A stepping motor of the same type used on the
wave-length drive is directly coupled to the control shaft using
apair of miter gears.
Filter selectionTo decrease the amount of stray-light passing
through themonochromator an optional filter module with a
ten-positionfilter wheel is available. The filter wheel is
positioned behindthe entrance slit. The proper filter is moved into
position
when the filter’s select line is shorted to the internal
commonusing one of several SK3506 triacs as shown in Figure 3.
Thestray-light filters can be selected in three different modeswith
the GCA/McPherson EU-700 filter module: a manualmode, an internal
automatic mode, and an external mode.The monochromator controller
can select the proper filterfor preselected wavelength regions
stored in a table in themonochromator program when the filter
module is operatedin the external mode. The filter selection
subroutine deter-mines the correct filter and outputs the encoded
filternumber to the filter selection circuit.
Encoder devicesIn this system several encoding devices are used.
These includea relative wavelength encoder, single-value absolute
encoderson the wavelength and slitwidth drives, limit switches on
thewavelength and slitwidth drives, and a
"filter-change-in-progress" signal on the filter mechanism.
Absolute VS relative encodersTo provide the required resolution
for the wavelength drive,an absolute encoder with at least one part
in ten thousandresolution (14 bits) across the wavelength range is
needed.This would be prohibitively expensive.
The high cost of absolute encoders can be avoided bymaintaining
a wavelength counter. This counter must beset each time the
instrument is turned on. A relative encoderwith the proper
resolution is used to update the wavelengthcounter. This method
would preclude the complete auto-mation of the monochromator
because it normally requiresthe operator to initialise the counter.
For this controllerthe relative encoder provided with the
monochromator isused across the wavelength range, and an automatic
absoluteencoder is used for the calibration sequence at a single
wave-length.
Single-value absolute encodersThe wavelength and slitwidth
absolute encoders are singlevalue encoders. These encoders allow
the automatic encod-ing of the wavelength to within 0.01 nm and
encoding of theslitwidth to within 0.25 /am. This is as accurately
as themechanical counters on the monochromator can be manuallyread
and set. The absolute encoders were constructed fromthe counters on
the top of the monochromator module. Thewavelength and slitwidth
counters are removed from themonochromator module. The wavelength
and slitwidthcounters are removed from the monochromator and
modifiedfollowing the design of Cordos and Malmstadt [3].
Themechanical counters are then replaced in their normal positionin
the monochromator.
AC HI
+5
IOK +5"RCA SK 5506UPPERLIMIT 452 7407 GMTI17408)
i00+5
Figure 2. Slew motor drive circuit
L_
--WAVELENGTH DRIveL/’SOLENOID-
Volume 3 No. 2 April 1981 77
-
Warren et al Automation of a UV visible monochromator
The encoder as diagramed in Figures 4a, b and c, functionsas
follows: the incandescent lamp is turned on during cali-bration. At
the calibration setting light can pass from theincandescent lamp to
the photocell through the hole drilledthrough the counter rings. An
interrupt transition is thengenerated to the microprocessor calling
the proper interruptsubroutines for the wavelength and slitwidth
counter. Theinterrupt subroutines set the wavelength and
slitwidthcounters for their respective encoders.
Re!ative wavelength encodersThe relative wavelength encoder is a
Model B 200X AVSoptical tachometer (H. H. Controls, 16 Frost
Street, Arlington,MD, 02174, USA.). The optical tachometer provides
quadra-ture output that determines the direction and magnitude
ofthe wavelength change. The circuitry that converts the
quad-rature outputs to increment or decrement pulses requiresthe
detection of an edge from each phase before producingthe pulse. The
circuit is shown in Figure 5. The resultingpulses generate
interrupts in the microprocessor that incre-ment or decrement the
wavelength counter.A relative wavelength encoder would not
necessarily be
required if movements were made only through use of thestepper
motor. Because of the slew motor’s inertia and thespeed at which
movement occurs, only large coarse wavelengthchanges can be made
with the slew motor. Even if the slewmotor were not used, however,
the relative encoder wouldprovide important assurance that the
wavelength was properlyset.
Limit switchesAnother set of absolute encoders is provided by
limit switcheson the wavelength and slitwidth drives as shown in
Figure 6a.The limit switches for both the slitwidth and the
wavelengthdrives are used as a point of reference for the
calibrationprocesures. The limit switches also provide important
pro-tection for the wavelength and slitwidth drives. Althoughthe
limit signals are returned and detected by the micro-processor, the
limit signals are also used on the control boardto prevent any
movement in the direction of the limit throughthe board’s control
logic. This prevents damage to eitherdrive.
When the sine-bar nears its limit of travel on the leadscrew,the
wavelength limit switches are actuated by the limitswitch actuator
bar attached to the sine-bar. The slitwidthlimit switches are
actuated when the slitwidth stops reachtheir mechanical limit.
a)
b)
I Ca r-INCANDESCENT
CdSe "b",l’’[’1’[ [)ORI TRRANSISTORI-TO TRANSISTOR AMP
+Sv
+5
LAMP 2N568 L IK 7414k"’x/
ENcODER ON
c)2N2222 ENCODER OUT
PHOTOCELL
Figure 4. (a) Cross section of a single value absoluteencoder
(b) Lamp drive circuit for single value absoluteencoder, (c) Enoder
pulse shaping circuit
Filter changeWhen the filter wheel is in a detent position, the
filter-changeline from the filter module is low. When a filter
change isoccurring, the filter-change line rises between filters.
Eachrising edge triggers a retriggerable monostable shown inFigure
6b, which sets the "filter-change-in-progress" signalhigh for one
second. When the filter wheel is moving, thefilter-change line
pulses more than once per second. Thiscauses the
"filter-change-in-progress" signal to remain highuntil the filter
wheel stops. The microprocessor inhibitsscanning by the
monochromator during the filter changeinterval.
The 8080 microprocessor systemThe microprocessor is the source
of the control signals forthe electromechanical system in the
monochromator, andreceives feedback signals from the monochromator,
as wellas instructions from the keyboard or the master
micro-processor. The present status of the monochromator is
FILTERNUMBER,
+5---
D 5c TB
_G2 I0
Figure 3. Filter selection circuit
7406MT2
FILTER SELECT LINEAUTOMATIC MODE
COMMONFILTER SELECT LINE-ONE
FILTER SELECT LINE-TWO
(SAME CIRCUIT REPEATSFOR ALL FILTERS)
78 Journal of Automatic Chemistry
-
Warren et al Automation of a UV visible monochromator
reported either to the operator through the display or to
themaster microprocessor. The microprocessor in the controlleris
constantly testing the status of the monochromator andchanging the
control signals sent to the electromechanicalsystem according to a
set of internal rules determined by themonochromator control
program.
The microprocessor system used in this study was designedby
Avery [4] and Lovse [51, using the Intel (3065 BowersAvenue, Santa
Clara, CA, 95051, USA.) 8080A micropro-cessor and components in the
Intel MCS-80 microcomputerfamily. Included are 3K bytes of
read-only memory, 256bytes of read/write memory, a keyboard and
alphanumericdisplay, an 8251 programmable communication interface,
an8253 programmable interval timer, an 8255 programmableperipheral
interface, and an 8259 programmable interruptinterface controller.
These microprocessor componentswere interfaced using standard Intel
methods. A blockdiagram for the microprocessor system is shown on
Figure 7.The control signals passed between the microprocessor
andthe control board are tabulated in Table 1.
Keyboard/displayWavelength and slitwidth readout and stand-along
controlare provided by a keyboard/display unit. The display
usesfive DL-1414 four-digit, 17-segment alphanumeric
intelligentdisplays (Litronix, Inc., 19000 Homestead Road,
Vallcopark!Cupertino, CA, 95014, USA.). The keyboard is composed
of20 encoded SPST keys (0-9 and A-J). The encoded signal isinput
through an 8212 8-bit input port by the keyboardsubroutine. The
encoded signal is decoded by the keyboardsubroutine and the proper
action is taken by the mono-chromator control program.
Monochromator control programThe monochromator control program
in the stand-alonemode enters commands through the keyboard. In the
slavemode commands enter through a serial I/0 port
whichcommunicates with the master microprocessor. The
mono-chromator control program can: calibrate and initialise
theoperation of the wavelength and slitwidth drivers, set
thegrating for any wavelength from 0.00 to 950.00 nm, set
anyslitwidth from 5 to 2000 /.tm, and select the proper stray-light
filter. The program also detects and reports any errorin the
electromechanical operation of the monochromator.
The controller program is organised into an
initialisationsequence, a main program loop, keyboard service
subroutines,and service subroutines for the main program loop. The
mainloop determines if the monochromator needs service and pro-
Table 1. Microprocessor input and output from the 8255and 8259
interfaces.
8255-Programmable peripheral interface
Port A (Output)
A0 Wavelength slew-upA Wavelength slew-downA2 Wavelength
step-upA3 Wavelength step-downA4 Slitwidth step-upA5 Slitwidth
step-downA6 Wavelength encoder lightA7 Slitwidth encoder light
Port B (Input)
B0 Wavelength forward limitB Wavelength reverse limitB2
Slitwidth forward limitB3 Slitwidth reverse limitB4
Filter-change-in-progress signalB5-B7 No used
Port C (Output)
C0-C3 Filter selectC4-C7 Not used
8259-Programmable interrupt interface controller
IR0 Wavelength encoder upIR Wavelength encoder downIR2 Slitwidth
calibrationIR3 Wavelength calibrationIR4 200HZ clockIR5-IR7 Not
used
CLOSED)LIMIT SWITCHES_it._,,/; (NORMALLY5K 1 tSK+5 +5
UPPER LIMITLOWER LIMIT
+5
"FILTER-CHANGE-IN lOl t5K-PROGRESS" SIGNALb) -’VV
FILTER CHANGE LINE
Figure 6. (a) Limit switch pull-up circuit (b) Monostablefor
filter-change-in-progress signal
74041 7404QUADRATUREB
_r-_
OUTPUT FRoMRELATIVEENCODER
7404 -7404I00 33K
+5
ENCODE UP
IK+5
Figure 5. Wavelength relative encoder pulse generator
Volume 3 No. 2 April 1981 79
-
Warren et al Automation of a UV visible monochromator
vides any needed service by executing one of the
servicesubroutines tabulated in Table 2. Input of commands
isperformed by the keyboard subroutine in the stand-alonemode and
by the slave subroutine in the slave mode. Thesesubroutines then
call the instruction subroutine. The in-struction subroutine
decodes the instruction and sets up andstarts any move required.
The commands and their functionsare tabulated in Table 3. The
program was written in assemblylanguage because of speed
requirements, since wavelengthinterrupts occur at 900 Hz when slew
motor is on. Themicroprocessor must update the wavelength counter
900times a second, and simultaneously must service the
display,while continually comparing the desired and actual
wave-length.
Power supplyThe control board requires 5 volts at amp and 32
volts at0.5 amps. The required voltages are supplied by an on-board
power supply. The microprocessor is supplied with+5, +15, and -15
volts from a Power-One CBB-75W powersupply (P.O. Box 1261, Canoga
Park, CA, 19304, USA.).
ResultsThe accuracy of the monochromator controller was
testedwith atomic lines from hollow cathode lamps. The measure-ment
system used is shown in Figure 8. The controller wastested in both
the stand-alone mode and in the slave mode.The master
microprocessor used programs written in BASICto control the slave
microprocessor.
The wavelength calibration procedure was tested by
firstcalibrating the wavelength drive, then moving to a
wavelengthbelow the 253.65 nm emission line of mercury hollow
cathodelamp and stepping through the emission line. The
measuredemission wavelength maximum was then compared to
thestandard value [6]. This procedure was repeated one
hundredtimes, and the measured peak maximum (253.55 nm)waswithin
the accuracy of the lead screw (0.1 nm) and wasreproducible to
within +/- 0.01 nm.
Slitwidth calibration was tested by moving to 100, 50,and 20
/.tm, after calibrating the slitwidth. The intensity oflight passed
through the monochromator to the detector wasmeasured 100 times at
each slitwidth. The measured intensitynever varied more than one
per cent for any slitwidth setting.This corresponds to an error of
0.2 ktm in the slitwidthsetting.
Accuracy during unattended operation was tested for aslong as 8
hours while the monochromator controller was inthe slave mode. The
monochromator wavelength drive was incontinuous movement during the
test period. At the end ofthe test period the monochromator
controller wavelengtherror was less than 0.01 nm. This is within
the resolution ofthe monochromator. The error for the slitwidth
movementwas less than the step resolution or less than 0.25/am.
Because of the high precision of this controller the wave-length
accuracy error caused by the leadscrew run-out errorfor a portion
of the wavelength ,range could be measured.The measured peak
maximums were compared to the knownwavelengths for specific lines,
and the error was calculated asshown in Table 4. Although some of
the measured errors are
SYSTEM BUS
|| 2,048MHZ SEE TEXT FOR DESCRIPTIONTO MASTER OF INPUT/OUTPUT
FORMICROPCESSOR 8255 AND 8259
Figure 7. Microprocessor system
Table 2. Main program loop service subroutines
KEYBD
SLAVE
SMOVIN
MOVING
DISPLA
MCSLIT
MCWAV
ERROR
FILTER
Determines if a valid input has occured from the keyboard and
takes any required action when enabled.
Determines if a valid input has occured from the slaveinput port
when in the slave mode.
Supervises any change in the slitwidth.
Supervises any change in the wavelength.
Displays the current values for the slitwidth and
thewavelength.
Supervises the slitwidth calibration procedure.
Supervises the wavelength calibration procedure.
Determines if an error in electromechanical operationhas
occured.
If enabled selects the proper stray-light filter.
Table 3. Microprocessor commands
COPY (B)
CSLIT (B)CWAVE (B)DISLW (B)ENSLW (B)FILDIS (B)FILENA (B)FILSEL
(M)
GOTO (B)
Sets wavelength counter equal to value in the inputbufferStarts
slitwidth calibration procedureStarts wavelength calibration
procedureDisables wavelength slewEnables wavelength slewDiables
filter selectionEnables filter selectionUses filter number stored
in the input buffer to selectthe correct filterStarts wavelength
change to wavelength in the inputbuffer
OUTFIL (M) Outputs the current filter number to the
mastermicroprocessor
OUTSLI (M) Outputs the current slitwidth to the master
micro-processor
OUTWAV (M) Outputs the current wavelength to the master
micro-processor
SCOPY (B) Sets slitwidth equal to value in the input bufferSGOTO
(B) Starts slitwidth change to value stored in the input
bufferSHOFF (B) Turns keyboard shift offSHON (B) Turns keyboard
shift onSLAVOF (M) Returns microprocessor to the stand-alone
modeSLAVON (M) Places microprocessor in slave modeSTATUS (M)
Outputs current status to the master microprocessorSSTEPD (B)SSTEPU
(B)STEPDN (B)STEDUP (B)STOP (B)WAITW (B)
WAITS (B)
Steps the slitwidth down 0.25/JanSteps the slitwidth up
0.25/.tmSteps the wavelength drive down 0.01 nmSteps the wavelength
drive up 0.01 nmStops any monochromator movementSets wavelength
scanning speed using value in theinput bufferSets slitwidth
scanning speed using value in the inputbuffer
B Commands available to both the keyboard and the master
mircro-processor
M Commands available only to the master microprocessor
SLAVE 8080 MASTER 8080
CONTROL BOARD
MONOCHRoMATOR
Figure 8. Experimental measurement system
80 Journal of Automatic Chemistry
-
Warren et al Automation of a UV visible monochromator
greater than the monochromator’s specification, it should
benoted that the monochromator used in this study is over tenyears
old. Because this error is a rather smooth and stablefunction of
wavelength, it is possible to calculate the wave-length of an
unknown emission line in a complex sample toan accuracy better than
the monochromator specification.The measured wavelength is slightly
dependent on temper-ature, as shown in Table 4. Thus, for the best
accuracy, themonochromator should be calibrated at the t.emperature
atwhich it is to be used. Also, data can be collected for a
fewpoints in the region of interest, and the wavelength shift canbe
calculated for the operating temperature by comparingthese data to
the calibration values at a standard temperature.The microprocessor
controller enables the overall performanceof the monochromator to
be improved beyond its basicspecifications, as well as providing
the automation features.
Software for the control programs, artwork for the PC boardsand
complete documentation are available from the authors.
REFERENCES[1] GCA Corporation, "Scanning Monochromator EU-700
and
EUE-700 Series", Acton, MA, USA, 1968.[2] Fractional and
Subfractional Horsepower Electric Motors.
C.G. Veinott, McGraw-Hill Book Company, New York, 1970.[3]
Cordos, E., and Malmstadt, H.V., (1975) Anal. Chem. 45,
425(2), "Programmable Monochromator for Accurate HighSpeed
Wavelength Isolation".
Table 4. Wavelength accuracy reproducibility and temperaturedata
for the. EU-700 monochromator
Temperature(C)24.0
25.1
27.1
Averagevalue
253.54365.06404.73435.90546.10253.55365.09404.77435.91546.12253.57365.10404.77435.92546.13
Literaturevalue (6)
253.65365.01404.66435.84546.08253.65365.01404.66435.84546.08253.65365.01404.66435.84546.08
Errornm
-0.11+0.05+0.07+0.06+0.02-0.10+0.08+0.11+0.07+0.04-0.08+0.09+0.11+0.08+0.05
[4] Avery, J., Ph D. Thesis, University of Illinois, School of
ChemicalSciences, Urbana, IL, 1978.
[5 Lovse, D., Ph D. Thesis, University of Illinois, School of
ChemicalSciences, Urbana, IL, 1977.
[6] The Chemical Rubber Company, Handbook of Chemistry
andPhysics, Cleveland, OH, 1970, p E-220.
Novel apparatus for the automation ofsolvent extraction
John G. Williams, Peter B.. Stockwell*, Michael Holmes, Derrick
G. Porter.Laboratory of the Government Chemist, Stamford St.,
London SE1 9NQ, UK.
IntroductionChemical analyses usually require pretreatment of
the sampleprior to measurement. When solutions are employed as
trans-port media for the analytes, convenient minimisation
ofphysical and chemical interferences may be brought aboutby
solvent extraction. Previously, automatic solvent extrac-tion
systems have been based on manual techniques, wherebythe eye of the
human operator is replaced by some form ofphase boundary detector
and the tap of the separating funnelis replaced by an
electromechanical valve. Most phase bound-ary sensors have broadly
similar characteristics. Apart fromoperating problems they all rely
on the differential trans-mission of electromagnetic radiation on
either side of theboundary. Trowell has described phase boundary
detectorsbased on differential conductivity [1] and differential
capa-citance [2]. Associated with the latter, relying on a
dielectricchange, are methods using a change of refractive index
[3] tocontrol a bistable valve as an interface flows past a
fixedpoint in the system. Recent, unpublished work performedat this
Laboratory has shown that ultrasonic transducers arealso capable of
phase boundary detection.
Other well tried forms of solvent extraction (other
thanchromatographic) involve the migration of the species
ofinterest across a semi-permeable membrane under the influ-
*Present address: Plasma Therm Ltd., 6 Station Road, Penge,
UK.
Volume 3 No. 2 April 1981
ence of either a concentration gradient or a potential
gradi-ent, or a combination of the two. Methods relying on
thegravity separation of two completely immiscible phases
aresometimes employed in continuous-flow air-segmentedanalytical
systems; when well designed they are relativelytrouble free. Vallis
[4] designed a somewhat differentapproach to automated solvent
extraction based on a rotat-able cup-shaped-vessel with a porous
lid attached to the lip.The cup is placed inside a collecting
vessel; if the porouslid is made from hydrophilic material such as
sintered glass,water will pass into the collecting vessel at low
rotationspeeds leaving the organic phase in the cup. An
increasedrotation speed then ejects the organic phase. The use of
ahydrophobic material such as sintered PTFE enables thepreferential
rejection of the organic phase.A prototype separator which .has
been designed and
built at this Laboratory using a completely new approach,
iscurrently the subject of a patent application [5]. In
principle,separation is effected by absorption of both phases into
aporous nickel-chrome alloy disc mounted on a motor-drivenshaft.
Controlled angular acceleration and centripetal forceon the
droplets within the pores enables one phase to beseparated from the
other. The speed of rotation of theporous disc is coupled
microelectronically to the verticalcomponent of its motion so that
separated droplets leavingthe disc tangentially are trapped by
hitting the walls of
81