P - SQ,2'J#'

l1111111111pu

k G

dD
, I~ ~~II

iF*nrrr.;*Si7Rlr*P

INSTRUMENTATION STUDY-
BLACK LIQUOR SOLIDS CONTENT
Project 3025

Report One
A Summary Report
to
MEMBERS OF GROUP PROJECT 3025

February 23, 1972

.- '-. ' i-.i rl- i-ri rr. li.-.r i - " . .L;r7LII.I.r t.,*a I Ibm(.lYI".IBP1._I-*-TiU' *W :- -ln'n .. .rr.l-.,r

J
 THE INSTITUTE OF PAPER CHEMISTRY
Appleton, Wisconsin

INSTRUMENTATION STUDY - BLACK LIQUOR SOLIDS CONTENT

Project 3025

IacDl>lzln Boedel Research Limited

LIBRARY
3350 East Broadway
manouver 12, B.Cq

Report One

A Summary Report

to

MEMBERS OF GROUP PROJECT 3025

February 23, 1972

MEMBERS OF GROUP PROJECT 3025

American Can Company

Bowaters Southern Paper Corporation
Container Corporation of America
Continental Can Company, Inc.
Crown Zellerbach Corporation
W. R. Grace & Co.
Green Bay Packaging Inc.
Hammermill Paper Company
Hoerner Waldorf Corporation
Inland Container Corporation
International Paper Company
Kimberly-Clark Corporation
Longview Fibre Company
MacMillan Bloedel Limited
The Mead Corporation
St. Regis Paper Company
Scott Paper Company
Weyerhaeuser Company
 TABLE OF CONTENTS

Page

SUMMARY 1

INTRODUCTION 2

APPARATUS AND PROCEDURES 5

Flow .Loop 5

Procedures 7

Solids Measurement 9

Liquor Analysis 10

EMC REFRACTOMETER 13

Description of Instrument 13

Instrument Checks 18

Instrument Calibration 21

Basic Instrument Response 23

Spent Liquors 31

Temperature Compensation 33

System Variables 36

Composition Effects 37

Refractive Index Measurement 44

On-Line Calibration 45

NUS VELOCIMETER 49

Principle of Operation 49

Specifications 50

Description of Instrument 51

Instrument Checks 58

Temperature Measurement 60
 iv

Operational Variables 62

Black Liquor Measurement 64

Temperature Compensation 71

Composition Effects 73

On-Line Calibration 82

DIRECT MEASUREMENT OF SOLIDS CONTENT 85

Residual Solids Methods 86

Distillation Method 88

Comparative Tests 92

CONCLUSIONS 95

ACKNOWLEDGMENTS 97

LITERATURE CITED 98
 THE INSTITUTE OF PAPER CHEMISTRY
Appleton, Wisconsin

INSTRUMENTATION STUDY - BLACK LIQUOR SOLIDS CONTENT

SUMMARY

A comparative evaluation of two commercial instruments for measuring

black liquor solids content (an Electron-Machine Corporation refractometer and

an NUS Corporation sonic velocimeter) was carried out. The study showed that both

instruments have the capability of measuring black liquor solids. The sonic

velocimeter is quite sensitive to air bubbles and would probably be limited to use

before the direct contact evaporator unless special provisions for deaerating the

liquor were made. Both instruments are sensitive to changes in chemical composi-

tion of the black liquor and would require frequent recalibration. Procedures and

limitations of on-line calibration are discussed in the report. The response of

the velocimeter increases as the inorganic-to-organic ratio in the liquor increases.

The response of the refractometer to such changes is less distinct, but tends to be

in the opposite direction. This leads to the possibility that use of the two

instruments in tandem could provide a means for detecting changes in liquor solids

content and in chemical composition simultaneously.

Page 2 Group Project 3025
Report One

INTRODUCTION

Reliable, continuous measurement of black liquor solids content is a

subject of considerable importance to the pulp industry. The solids content of

liquor introduced into a recovery furnace can have a pronounced influence on firing

behavior. Current trends toward better control of the recovery furnace for reasons

of safety and reduced air emissions require a greater degree of control over incoming

solids content and appropriate adjustment of operating conditions to handle varia-

tions in solids content. This, in turn, requires a reliable method for continuously

monitoring solids content. An important factor influencing the need for on-line

measurement of solids content is the recommendation by BLRBAC (Black Liquor Recovery

Boiler Advisory Committee) that a minimum solids content of 55% be maintained on

firing to avoid smelt-water explosion hazards. The need for solids measurement is

not restricted to high-solids liquors connected with furnace operation. There is

also a need for solids measurement on weak liquors to permit control of evaporator

and solids concentration entering the furnace envelope.

A number of devices are available for on-line measurement of black liquor

solids content. These include instruments based on refractive index, nuclear

attenuation, specific gravity, sonic velocity, and vibration amplitude. In addition,

the ammeter reading of the cascade evaporator motor drive is often used as an indica-

tion of liquor solids content. In the proposal for this project, it was indicated

that a general review of the state of the art of black liquor solids measurement

would be carried out. This was to include a review of measurement methods, instru-

ment manufacturers, and the extent of application in the industry. An extensive

survey of the use of black liquor solids monitoring devices in the kraft industry

has been recently carried out by the Instrumentation Subcommittee of BLRBAC (1).

The following conclusions were reached:

 ~~~~~~~
1~~~~~~~~

Group Project 3025 Page 3

Report One

1. Instruments are available for successfully measuring black liquor solids

concentration continuously. These instruments have sufficient dependability,

accuracy, and reliability to permit their use in promoting safe operation of

recovery boilers.

2. Refractive index-type instruments are more widely accepted and far more

successful with less maintenance requirements than other types of instruments

presently or previously used for this service.

5. Refractometer installations have proven sufficiently reliable and accurate to be

used for automatic alarming on low black liquor solids concentration.

4. There remains a considerable need for improvement of instruments presently

available and for the development of new techniques and/or instruments for

monitoring and control of black liquor solids concentration.

In view of the fact that a survey such as we had contemplated would

duplicate the work done by BLRBAC and the fact that most sponsors of this project

are members of BLRBAC, it was decided not to carry out a survey. Specific details

of the mill survey are on file with the BLRBAC subcommittee and are available to

any member of BLRBAC upon request. We are not free to divulge the findings of

that study.

All of the methods for continuous measurement of black liquor solids are

indirect methods in that they measure some property of the black liquor which is

dependent on solids content. Thus, they must ultimately be calibrated in terms of

some method for the direct determination of solids content. The accuracy of the

continuous methods is thus ultimately dependent on the accuracy of the direct

measurement and the calibration procedure employed. Direct measurement of black

liquor solids content is not completely straightforward. Black liquor is a complex

I
page 4 Group Project 5025
Report One

mixture of organic and.inorganic compounds., .and at the present time there exists no,

laboratory method for determination- of black liquor solids content which is accept-

able to all segments: of the kraft pulping industry. Parker et al. (2) have carried

out a comprehensive review of methods for off-line measurement of solids content and

have discussed the problems involved. The problems of accuracy in methods for

direct determination of black liquor solids are fundamental to the general problem

of measurement of this quantity and will be discussed later in this report.

The major emphasis of this project was directed toward a comparative

evaluation of two commercial instruments for black liquor solids measurement. The

two instruments which were chosen were as follows:

1. An Electron-Machine Corporation Model BL-168 Black Liquor Analyzer which is

based on measurement of the refractive index of black liquor.

2. An NUS Corporation Model 6161-139 Sonic Solution Monitor which employs a

measurement of sonic velocity of the liquor.

The evaluation was directed mainly toward the applicability of these instruments in

on-line measurement of black liquor solids, and investigation of potential sources

of errors in their use. A major objective was a study of the effect of liquor

composition on instrument response.

Group Project 3025 Page 5
Report One

APPARATUS AND PROCEDURES

FLOW LOOP

The two instruments were installed in a flow loop which was constructed

to permit simulation of on-line operation. A schematic diagram of the flow loop

is shown in Fig. 1. The instruments were mounted in a vertical leg of the flow

loop with the sonic velocimeter located immediately below the refractometer. The

4
sonic velocimeter was mounted in a standard -in. tee. The refractometer was

mounted on a special Electron-Machine elbow adapter which was furnished with the

instrument. The adapter was sized for a standard flanged connection to a 1-in.

process line.

The flow loop consisted of a circulating pump, a heat exchanger, the

instruments, a pressurized accumulator, and the necessary piping. A steam-jacketed

vessel with an integral stirrer was used for making up liquors prior to their

introduction into the flow loop. This vessel could be sealed and evacuated to

permit deaeration of the liquors before feeding them into the main loop. A Moyno

pump was used to pump the liquor from this storage vessel into the test loop.

In the initial configuration of the flow loop, a La Boor centrifugal pump

was used as the only circulating pump, with the Moyno pump used solely for feeding

liquor to the system. In order to avoid the possibility of introducing air into

the system at the centrifugal pump, the piping was changed to permit using the Moyno

pump as a circulating pump as well as a feed pump. In the final configuration,

either the centrifugal pump or the Moyno pump could be used as a circulating pump.

The accumulator, located at the high point in the loop, was simply a

cylindrical pressure vessel 10 in. i.d. and 30 in. long, with a side arm attached

halfway up for incoming black liquor. A sight gage was installed to permit
 Pr~qqI~rvr
vacuum connection P

0
CD

Accumulator
EMC
.Refrac-
tome~ter

NUS
Ve Locim-

Steam
Sample

Moy'no Centrifugal
pun
0

To drain
0
To drain

Figure 1. Schematic Diagram of Test Loop

, I I
Group Project 3025 Page 7
Report One

observation of the-liquor level. At the top of the accumulator, .provisions were

made for pressurizing the loop with nitrogen and for evacuating the loop. The

return leg to the circulating pump was taken from the bottom of the accumulator.

A bypass line ran directly from the heat exchanger to the accumulator to

permit control of flow through the instrument test zone. With the Moyno pump, an

additional mode of flow control was available through adjustment of the pump speed

with a Reeve's drive. A separate bypass line around the refractometer was installed

to permit reducing the velocity through the smaller refractometer adapter if desired.

With the exception of the piping in the immediate vicinity of the instruments, all

piping in the main loop containing the centrifugal pump was 2-in. diameter pipe.

The feed lines from the storage vessel and the piping around the Moyno pump were of

1-in. diameter pipe.

The capacity of the flow loop was about 15 gallons. The rated capacity

of the centrifugal pump was 100 g.p.m., while the maximum capacity of the Moyno

pump was about 10 g.p.m. The heat exchanger was rated at 20,000 B.t.u./hr. and

provided the capability to control temperature over the range from 60 to l40C.

The loop was capable of being pressurized up to 140 p.s.i.g. A thermocouple was

installed in the tee which held the velocimeter. Once it was established that the

temperature measurement in the NUS instrument was reliable, the NUS temperature

measurement was used for the bulk of the work. A sample tap was provided

immediately ahead of the velocimeter tee. This was used for taking samples for

off-line measurement of solids content.

PROCEDURES

The normal operating procedure was to start with high-solids liquors as

obtained from the mills. The semisolid liquor was scooped into the steam-jacketed

I
Page 8 Group Project 5025
Report One

storage vessel and heated .to boiling while stirring. When this point was reached,

all valves in the flow loop were opened, the accumulator vent was opened to the

atmosphere, and the liquor was then pumped slowly into the flow loop until the

desired level in the accumulator was reached. The storage vessel was then valved

off, the system pressurized with nitrogen to the desired level, bypass valves

adjusted, and circulation of liquor begun. With black liquor, the Moyno pump was

used almost exclusively as the circulatory pump. A run would then be carried out

varying temperature, velocity, and pressure as desired and observing the instrument

response. Liquor samples were taken for off-line solids measurement. At the

completion of a run, the liquor would be forced back to the storage vessel by the

pressurization of the loop. Water would be added at that point to reduce solids

concentration, the liquor heated to boiling, and the procedure repeated.

The flow loop was used for a series of runs with industrial black liquors

and some introductory runs with sucrose solutions. Much of the work on the effect

of composition on measurement response was made external to the flow loop. The

main reason for doing this was to cover a wider range of compositional variables

than was possible in the flow loop. The difficulty with using the flow loop itself

for studies of compositional effects was the relatively large volume of liquor

required (about 15 gallons) and the time needed to make up a sample, feed it into

the loop and carry out a run. To measure sonic velocity, the NUS probe was simply

inserted into a beaker containing the sample, with a water bath used for temperature

control. The EMC refractometer was not used in "out of loop" studies because the

geometry of the adapter section was not suitable for this mode of testing. Instead,

a laboratory refractometer was used to determine the effect of changes in composition

on the refractive index. The laboratory instrument was a Bausch and Lomb Type 334558

refractometer.
Group Project 3025 Page 9
Report One

SOLIDS MEASUREMENT

Direct measurement of solids content of liquor samples was used as a

reference in interpreting instrument response. Several methods were used for

these determinations including different ovendry procedures and a distillation

method. The most successful procedure was that taken from a final draft of a

proposed revision of TAPPI Standard T 625 ts- 6 4, "Solids Content of Black Liquor,"

as submitted to the Standards Editor of TAPPI in June, 1969. In this procedure,

which follows many of the recommendations of Parker et al. (2), liquor samples are

dried at 105C. for a minimum of six hours. An inert surface extender and a

controlled flow of dried air are used to increase drying rate and eliminate moisture

entrapment. Strong black liquors are diluted to allow volumetric handling and to

reduce scum formation.

Measurement of the solids content by the distillation method tended to

give slightly lower values of solids content than the ovendry method. In order to

keep a single base for evaluating the instruments, all of the data reported herein

related to the NUS velocimeter or the EMC refractometer are based on ovendry solids

values.

In the runs with black liquor in the flow loop, liquors were progressively

diluted to vary solids content. The amounts of dilution water used were recorded

and used to minimize uncertainty in measured solids values. The method of doing

this is described below. It was assumed that the accuracy of the solids determina-

tion improved as the liquor became more dilute, and so the most dilute liquor was

used as a base. The working equation for the solids smoothing was:

Yi = W + x .
1

I
Page 10 Group Project 3025
Report One

where

Y. = solids content of ith sample,

-1-

x = total solids in loop at end, lb.,

W = total water in loop at end, lb., and

A. = dilution water added from ith sample to end, lb.

Inverting Equation (1) gives:

Y.
-x= + 1 - x (2)

or

1 - 1 = w _ i
Y. x x

Equation (3) was used to minimize the uncertainty in the solids data. The solids

content ( Yi ) and the amount of dilution ( A. ) were measured quantities. A plot

of (l/Yi) - 1 vs. A. was then made. The best straight line consistent with a

reasonable estimate of the initial solids ( x , which relates to the slope) and

the greater reliability of low-solids samples was used to determine the set of

values of solids content in the runs. An example of this method of handling the

solids data is shown in Fig. 2. These particular data show an excellent agreement

between measured solids content and amounts of dilution water. The agreement was

generally poorer with higher solids content liquors.

LIQUOR ANALYSIS

Three different commercial liquors were used in the tests. Two liquors

(labeled A and B) were obtained from a southern kraft mill pulping southern pine to

high yield and mixtures of pine and hardwoods to low yields. Liquor B was obtained
 Group Project 3025 Page 11
Report One

6.0

5.5

5.

4.
0

o
H
I'

mI
P; 3.5

'H I>'

2.5

2.0

1.5

1.0
0 40 80 120 160 200 240 280 320
A. , DILUTION WATER ADDED, lb.

Figure 2. Interrelation Between Measured Solids Content and Dilution Water

Page 12 Group Project 3025-
Report One

from the same mill five weeks after Liquor A. A third liquor (Liquor C) was

obtained from a northern kraft mill pulping softwoods to conventional kraft yield

levels. The results of chemical analyses of these liquors are shown in Table I.

It should be noted that although Liquors A and B were obtained from the same mill

and have similar inorganic contents, there is a difference in the degree of oxida-

tion of the sulfur compounds. Liquor C contains significantly more sulfur and.also

has a distinctly higher inorganic content. In all cases, the liquors were obtained

at high solids concentration and were shipped to the Institute in sealed drums.

TABLE I .

CHEMICAL COMPOSITION OF BLACK LIQUORS

All results are expressed as weight percent on black liquor solids,

and are based on TAPPI Procedure T 625

A B C

Sulfated ash, % 34.0 32.6 36.5

Organic, % 66.0 67.4 63.5
Total sodium, % 19.4 19.3 21.9 ..
Total sulfur, % 3.4 3.4 5.0
Na2S, % 1.03 0.29 4.14
Na2 S2 03, % 35.93 3.16 4.10
Na2 SO4 , % 5.43 7.05 5.14
Inorganic sulfur, % 3.24 2.99 4.53

. . . . . . . . . : . .. . } . . . . .. . .~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Group Project 3025 Page 13
Report One

EMC REFRACTOMETER

DESCRIPTION OF INSTRUMENT

The EMC refractometer, model BL-168, uses the principle of critical angle

refractometry by which dissolved solids content is related to the refractive index

of the solution. The basic unit consists of a sensing head and control console,

Fig. 3. The sensing head is mounted on the adapter installed in the process line

and the process solution is viewed through a transparent window. The surface of

the solution which is in contact with the window is optically scanned by a motor-

driven octagonal prism. The refractive index is determined by a beam of light

from the prism which repeatedly sweeps at a varying angle at the surface of the

solution. At a point in the cycle, as the angle between the incident light beam

and the surface decreases, the light beam instead of refracting into the solution

is reflected back into the optical system. The point of change from refraction

to reflection depends on the refractive index of the solution and is known as the

critical angle. The prism scans the beam of light through the critical angle at

the rate of 480 Hz. A photodetector measures the alternating light and dark

periods and its electrical output is used to provide a readout of refractive index

in the required units (R. I., % concentration, degrees Brix., etc.). A principal

advantage of this method is that it measures the index of refraction at the surface

of the process solution. Since the light beam does not penetrate into the solution,

the instrument can be used for opaque as well as transparent materials.

The optical system is shown in Fig. 4. The light from the tungsten lamp

source passes through a collimating lens and an aperture to give a narrow beam.

The octagonal prism rotated at 60 r.p.s. causes the beam to oscillate as each face

of the prism traverses the beam; the light exiting from the prism is displaced by

an amount equal to the width of the faces. The cycle is repeated for each face
 U t U~~~~~~aaoopjUow oqs.9qq
4-)
0-
0
0-
wo~ ~
~ ~ ~~~~~~a~-p
c ~~~ ~ ~ ~ ~ ~ ~ ~ ~ ~~~~9wSO oqo
 > & a I

0
0
Lamp o

Collimating lens o

0
C-
C+
-Aperture

Direction of beam travel

Photo-
~/ Focusing- tube
/ /- lens

Rotating prism
(scanner)

Sapphire prism

Process solution
(D

Refracted angle
0o
('ow
(D 0
Figure 4. Schematic of Optical System of Refractometer.
Page 16 Group Project 3025
Report One

so that the frequency of oscillation is 8 X 60 = 480 c.p.s. This beam is focused

at the interface of the sapphire prism and the process solution. The effect of

focusing is to cause the light to enter the process solution at an angle which

varies as the light beam oscillates. At the higher angles, the light is refracted

into the solution. At the shallower angles, the light is reflected through the

sapphire, through a lens assembly, to the photodetector. The ratio of refracted

to reflected light varies with the refractive index of the solutions being measured.

The light and dark periods on the photodetector produce a 480-Hz a.c. output from

the device which is transmitted to the control console. The function of the

electronics in the console is to condition the signal so that the light and dark

periods can be compared. The resulting signal which is proportional to refractive

index is used to drive the front panel meter which reads directly in percent solids.

The signal passes through a preamplifier; the output of the preamplifier is squared;

the positive and negative portions averaged and compared by a differential amplifier,

the output of which drives the meter. The signal conditioning provides an output

which is dependent on light and dark periods (reflection and refraction periods),

and hence the output is not sensitive to changes in the light intensity.

Temperature compensation is applied at the input of the differential

amplifier. This compensation is necessary because the refractive index of the

solution changes with temperature. The temperature is sensed by a thermistor

located close to the sapphire prism on the sensing head.

An Electron-Machine elbow adapter (Fig. 3), supplied with the refractometer,

was used to mount the sensing head on the process line. The adapter not only

provides for physical attachment of the sensing head but also serves to direct the

black liquor flow against the window. Thus, fresh liquor is continuously brought

into intimate contact with the window. This is important because the index of
Group Project 3025 Page 17
Report One

refraction of the liquor is measured at the solution-window interface, and misleading

readings could be obtained with a relatively stagnant film.of liquor next to the

window. The turbulence associated with the jet of liquor impinging on the window

also helps to minimize the build-up of deposits.

It is necessary to provide cooling water for the scanning prism motor.

There is also provision in the head for a steam purge nozzle for cleaning deposits

from the window. In this evaluation, no use was made of any prism cleaning

apparatus, and the purge tap was simply plugged. There was never any indication of

deposits during the test period, and the window was found to be clean whenever the

head was removed from the loop.

The unit is factory calibrated :for each specific application and the

manufacturer states that it normally requires no adjustment when installed. The

unit which we evaluated was factory calibrated to read percent solids over the range

of 45 to 75%, and the temperature compensation was set for a particular liquor.

Manufacturer's specifications for the instrument are:

Range: 1.2 to 1.7 R.I. converted to indicate percent soluble solids

within this range.

Accuracy: 1% meter span - standard meter spans

5% soluble solids
10% soluble solids
20% soluble solids
40% soluble solids.

Response time: 1.0 second.

Solution
temperature
range: 0 to 400F. '

Temperature
compensation: Continuous over full range, adjusted to compensate within
0.1% over 20F. range specified by customer.
Page 18 Group Project 3025
Report One

Although the. instrument is. factory calibrated, there are several adjust-

ments provided in the control console. These are a ZERO and SPANto calibrate the

lower and upper ends of .the meter scale, TEMP to provide the proper temperature

compensation for the solution being measured, and AGC (automatic gain control) which

insures that the circuit is operating in the proper range. The operating and

service manual which is provided with the instrument provides simple instructions

for making these adjustments. If a major change in the range is desired, the

manufacturer recommends that the instrument be returned to the factory for recali-

bration since this requires a mechanical adjustment in the optical system.

No information is available as to the life expectancy of the various

components in the instrument. However, the manufacturer recommends that the

following spare parts be kept on hand: exciter lamp, lamp transformer, phototube,

prism motor, encapsulated preamp module, fuses, and replacement printed circuit

cards.

INSTRUMENT CHECKS

At the start of the program, several tests were made on the effect of

environmental conditions on the refractometer. These included checks on the

effects of line voltage variations, stray magnetic fields, and vibration. A

twenty-four-hour stability check was also run. All of these tests were run with

the loop filled with a sucrose solution but with no circulation.

In the test of the effect of line voltage on instrument response, the

instrument was plugged into a 15-amp. powerstat. The output of the powerstat was

monitored with a GE type P-3 voltmeter. The loop was filled with a sucrose

solution, and the refractometer was adjusted to read about 53% solids with the

temperature compensation disconnected. The solution was at a temperature of 355C.

Group Project 3025 Page 19
Report One

The line voltage was adjusted from 95 volts to 125 volts. The results of this test

are shown in Table II.

TABLE II

EFFECT OF LINE VOLTAGE ON REFRACTOMETER OUTPUT

Refractometer Output
Line Voltage, Temp. Compensation On, Temp. Compensation Off,
volts % solids % solids

115 53.1 56.8

105 53.1 56.8
95 53.1 56.8
125 53.1 56.8
115 53.1 56.8

It can be seen that variation in the line voltage between 95 and 125 volts had no

effect on the meter reading.

In order to check the effect of stray magnetic fields, a coil of wire

6
connected to a 0-cycle line was used to generate the field, and this was moved to

various locations around the probe and console. The only location of the field

that affected the instrument output was at the top and side of the probe (sensing

head). In this case, the reading was increased from 53.1 to 53.6%.. Placing the

field to the front of the sensing head (away from the flow loop) did not affect the

reading. Locating the field near the console unit and the interconnecting cable

did not affect the reading.

The vibration test was run by clamping a "Lightnin'" model F motor with

an off-centered shaft to a flange on the refractometer adapter in the loop.

Sufficient vibration was produced to be felt at both the refractometer sensing head

and at the NUS probe mounted below. The vibration did not affect the refractometer

reading.

I
Page 20 Group Project 3025
Report One

The twenty-four-hour stability check was run by allowing the system-to

stabilize for two hours and then taking readings periodically. The temperature

compensation on the refractometer was disconnected for this test. The reason for

this was that the temperature compensation pot was adjusted for a black liquor while

a sucrose solution was used for the test. The results are shown in Table III.

TABLE III

STABILITY CHECK ON REFRACTOMETER

Time, Refractometer Output, Temperature,

hr.:min. % solids C.

16:30 53.4 32.1

19:40 53.55 29.5
22:40 53.6+ 27.8
8:05 53.7 27.0
9:00 53.7 27.0
10:15 53.7 27.2
12:00 53.7 . 27.6
13:00 535.7 28.2
15:00 53.7 28.7

In general, the stability of the output over a twenty-four-hour period is

quite good. The slight increase in reading corresponds with a decrease in tempera-

ture as the sucrose solution responded to changes in ambient temperature. A

decrease in temperature of'the sucrose solution would increase its index of

refraction and thus cause a higher output reading. Allowing for the changing

temperature, it appears that the output did not change by more than 0.1% solids over

the twenty-four-hour period.

 1

Group Project 3025 Page 21

Report One

INSTRUMENT CALIBRATION

The major part of the program regarding the EMC refractometer was devoted

to an evaluation of the ability of this instrument to measure solids content of

black liquor, and to study those factors which could affect its usefulness under

industrial conditions. Since the instrument actually measures refractive index, it

must be calibrated to give liquor solids content as an output. Any variable affect-

ing the index of refraction (such as temperature or chemical composition) would

affect the response of the unit. For this reason, the calibration procedures

employed are important in interpreting the results of this study.

The output of the EMC refractometer is a reading on a meter having a

linear scale ranging from 45 to 75% solids. This is because this instrument is

intended for use in measuring high solids content black liquors and was factory

adjusted to the refractive index range of such liquors. There are three adjustable

potentiometers which are used to calibrate the instrument to read directly in

percent solids; these are ZERO, SPAN, and TEMP. The ZERO pot serves to set a base

line for the output meter. The SPAN pot adjusts the proportionality between the

output reading and refractive index. The TEMP pot determines the amount of

temperature compensation employed. These pots are twenty-five turn trim pots and

are not susceptible to ready determination of pot settings. Thus, if a given pot

is adjusted to a new value, it cannot be easily returned to its original setting.

In commercial operation, it is, of course, highly desirable to adjust the

instrument so that the output (meter reading) corresponds to the true solids content

of the liquor. In order to do this, a two-point calibration is used to set ZERO

and SPAN. This should be carried out at a fixed temperature in the midrange of

expected temperature variations. A known value sample in the low solids range is

put in and the ZERO pot adjusted until the meter reading corresponds to the known
 I

Page 22 Group Project 3025

Report One

value of the sample. Next, a known sample from the high end of the range is put in

and the SPAN pot adjusted until the meter reading corresponds to the known value.

Since adjustment of the SPAN can affect the offset, the procedure may have to be

repeated. Note that this calibration procedure assumes an essentially linear

relationship between the measurement (refractive index) and the percent solids.

The TEMP pot is adjusted to hold the meter reading.constant as the liquor is heated

or cooled. Adjustment of the TEMP pot normally requires a readjustment of the ZERO

also.

In this evaluation study, it was decided that more useful information

could be obtained by minimizing pot adjustments than would be obtained by setting

up the instrument to read solids directly for each liquor tested. The emphasis is

then on instrument response to a wide range of variables (including solids content)

rather than on deviations between meter reading and solids content. In particular,

attention was focused on the items listed below.

1. A determination of whether or not, for a given liquor at a given temperature, a

linear relationship exists between the meter output and the solids.content of

the liquor. If a linear relationship does exist, proper adjustment of the ZERO

and SPAN pots would put the meter reading into a one-to-one correspondence with

the liquor solids content.

2. A determination of the ability of the temperature compensation circuitry to

handle the effect of temperature changes on the refractive index. For tempera-

ture compensation to be useful, it should not be necessary to adjust the TEMP

pot at different solids levels.

3. A determination of the effect of liquor flow rate, system pressure, degree of

liquor aeration, and the presence of undissolved solids on instrument response.

Group Project 3025 Page 23
Report One

4. Examination of the relationship between chemical composition of the liquor and

the refractometer response. In order to cover a wider range of variables than

was possible in the flow loop, data from a laboratory refractometer were used to

complement data from the EMC instrument regarding these effects.

The instrument, as obtained, was factory adjusted for a particular high-

solids black liquor. The instrument checks discussed previously were made with the

loop filled with a sucrose solution. The ZERO was adjusted at that time to obtain

an on-scale reading. Initial runs were carried out with sucrose-Na 2 CO3 solutions

of various proportions, and with temperature compensation connected and disconnected.

At the start of these series of runs, the ZERO and SPAN pots were adjusted to keep

the meter on scale over the temperature range employed. The SPAN pot adjustment

was relatively small. At the completion of the sucrose-Na2 CO3 runs, actual black

liquor was used in the loop. Liquor C was the first liquor tested. The ZERO pot

was adjusted to give a meter reading corresponding to a preliminary estimate of

solids content. The TEMP pot was given one turn counterclockwise. The SPAN pot

was not adjusted. These pot settings were then held constant for all remaining

tests. The TEMP pot was given one turn clockwise at the completion of the program

to restore its original setting.

BASIC INSTRUMENT RESPONSE

The initial runs were carried out with specific mixtures of sucrose and

sodium carbonate. These were done to obtain some idea of the basic response of the

instrument with simple one- and two-component solutions of organics and inorganics.

The tests were carried out at several solids levels and over a wide range of

temperatures. Both uncompensated and temperature-compensated outputs were obtained.

The results for tests on sucrose solutions, solutions containing 3 parts of sucrose

I
Page 24 Group Project3.025
Report One

to 1 part of Na 2 COs, and solutions containing equal parts of sucrose and Na2 CO3 are

shown in Fig. 5, 6, and 7.

There are several interesting features of these data. The offset due to

temperature compensation is clearly shown. This is the reason a change in the TEMP

pot would normally require a readjustment of ZERO. It also shows why the tempera-

ture compensation should be connected and the temperature held at the midpoint of

the operating range when setting ZERO and SPAN. It is also clear that the instru-

ment is overcompensating for the effects of temperature in these runs. This is not

surprising since the TEMP pot was not set to provide the correct amount of compensa-

tion for these solutions. The uncompensated responses are nearly linear functions

of temperature over the range tested. There is, however, a tendency for the output

to drop more steeply at the high temperature end. The only exception to this

behavior is the set at 40% solids of the equal parts sucrose-Na2 CO3 solution. The

reason for this anomalous behavior is not known. The degree of curvature of the

uncompensated curves is not very extreme and a linear approximation over a 30C.

range would certainly appear adequate. In all cases, the compensated response has

an S-shaped characteristic.. The output rises more steeply over the midrange of

temperatures than it does at either end. The reasons for this behavior are not

known, but it should be noted that this occurs over a much wider temperature range

than the 40F. compensation range stated in the instrument specifications. The

basic response of the instrument (general form of the curves) was not significantly

affected by changes in organic-inorganic ratio.

The data in Fig. 5, 6, and 7 may be cross-plotted to show output versus

solids content at fixed temperature levels. These results are shown in Fig. 8, 9,

and 10. It can be seen that a linear relationship between output and percent

solids does exist for all cases shown. In addition, the lines for a given liquor
Group Project 3025 Page 25
Report One

76

Compensate d - 50% solids

72

68 40% solids

64
H

160
o30%
solids
Uncompensated -
50% solids

40% solids

I. o50% solids
48

44 I I oI I I I I I I I I I
20 40 60o 80 100 120 140
TEMPERATURE, C.

Figure 5. Refractometer Output on Sucrose Solution

0 0 0 0 0 0 Ni
Page 26 Group Project 3025
Report One

76

72

Compensated -
40% solids
68

35% solids

i 64
H

30% solids

';
9 60 _

u o 25% solids

56 -

52 Uncompensated -

48
30% solids
25% solids

44 .L I I I I I I I I I I I
20 40 60 80 100 120
T EMPERA T URE, C.

Figure 6. Refractometer Output on a 3:1 Sucrose-Na 2COs Solution

 Group Project 3025 Page 27
Report One

76

72I Compensated -
40% solids

68
35% solids

H 64

solids

60
Or

56

Uncompensated -
-", 40% solids

52

solids
30% solids

48

44 I I I I I I I I I I I I
20 40 60 80 100 120 14o
TEMPERATURE, C.

Figure 7. Refractoineter Output for a 1:1 Sucrose-Na2C0 3 Solution

L
* I I
Page 28 Group Project 3025
Report One

76 r-

72 m

68 -

Compensated

80C.
64

56

Uncompensated

52 m

48

I I I .I1 I I.
25 35 40 )- 5 50
SOLIDS CONTENT, %

Figure 8. Refractometer Output vs. Solids Content - Sucrose Solution

Group Project 3025 Page 29
Report-One

76,

72

68 1-O0oC.

Compensated
-,8o0 c.

..-
500c.

AR 64

6o
~

I-,'5oc.

8o~c.

100 0 C.

52

Uncompensated

4+8

44' I i I I
* 25 30 35 4+0 4'5 50
SOLTD S CONTENIT, lo

Figure 9. Refractometer Output vs. Solids Content - 3:1 Sucrose-Na2 CO3 Solution
 Page 30 Group Project 3025
Report One

4'

76 m

72 -

800c.
68
j 50oC.
Compensated

0
i 64
H
50c.
6o
Po

80oc.

556

Uncompensated

52

44 I I I . . I I .. .. I I
25 540
35 50
SOLIDS CONTENT, %

Figure 10. Refractometer Output vs. Solids Content - l:l Sucrose-Na2COs Solution
 Group Project 3025 Page 31
Report One

composition are all nearly parallel, regardless of temperature and whether the

compensated or uncompensated response is used. Establishing linearity of response

to output solids variation is basic to use of adjustable ZERO and SPAN pots for

direct readout of percent solids. The parallel nature of the curves shows that the

temperature effects can be separated out and handled by temperature compensation

circuitry. It also shows that the temperature compensation circuitry does not

introduce any nonlinearity into the response. In essence, these data confirm the

adequacy of the controls provided (ZERO, SPAN, and TEMP) for calibrating the instru-

ment to give a direct readout of solids content.

SPENT LIQUORS

Runs were carried out with the three different black liquors discussed

previously. The temperature compensation circuitry was connected for all of these

runs. The runs were carried out in the order Liquor C, then Liquor A, and finally

Liquor B. The run with Liquor C was initiated immediately after the completion of

the work with sucrose and Na2 CO3 solutions. The ZERO pot was adjusted so that the

meter reading corresponded to an initial estimate of the solids content in the loop.

The SPAN pot was not adjusted. The TEMP pot was given one turn counterclockwise.

No known changes were made in the pot settings during the remainder of the test

program. Since the pots on the refractometer did not bear markings which would

indicate what the settings were, it was not possible to verify that no changes were

made.

A plot of the refractometer output (meter reading) versus liquor solids

content is shown in Fig. 11 for Liquors A, B, and C. The line of direct corre-

spondence between the meter reading and the percent solids is also shown. The

indicated output is an average of the temperature-compensated output over the

temperature range covered at each point. It is evident that a reasonably linear

h I
Page 32 Group Project 3.025
Report One

76
I

i 72

68

Liquor B

664 Liquor A
H

Liquor C

601
p;

0o

56

/ Xi/ /~Line of
Direct
Correspondence
52

48 /

44
/ I lt
/
I~~~~~~~ I I I . I I I I I .0
40 44 .48 : 52 56 60 64
LIQUOR SOLIDS CONTENT,

Figure 11. Refractometer Output vs. Solids Content for Black Liquors
 Group Project 3025 Page 33.
Report One.

relationship between refractometer output and solids content exists for all three

liquors. This would indicate that proper adjustment of the ZERO and SPAN pots

could shift any of these response curves into direct correspondence. It must be

emphasized that the fact that no response curve is in direct correspondence is to be

expected since the necessary calibration procedure to bring this about was not

employed.

The fact that the response curves are not identical is primarily an indi-

cation of the effects of liquor composition, although some of the difference might

be due to long-term drifts in the instrument, since the runs were carried out over

a period of several weeks. A detailed discussion of composition effects will be

given later; however, some features of-these curves are of interest. The -lines for

Liquor A and Liquor B are nearly parallel. This is interesting because Liquors A

and B were obtained from the same mill, and according to the data in Table I have

similar chemical compositions. The differences between Liquors A and B could be

taken care of simply by adjustment of the ZERO. Liquor C, which has a significantly

inorganic content and much more sulfur, shows a different slope as well as an offset.

The fact that the line for Liquor C nearly parallels the line of direct correspond-

ence must be considered coincidental.

TEMPERATURE COMPENSATION

The ability of the instrument to provide temperature compensation was

examined in some detail. At each solids content, the temperature of the liquor was

varied over a range of about 40C. and the compensated output recorded. The

results of these tests are shown in Fig. 12 for Liquor C and in Fig. 13 for Liquor A.

The small initial adjustment of the TEMP pot was made for Liquor C and it is clear

that excellent compensation was obtained. There was no observable change in the

meter reading over the temperature range from 80 to 110C. for each of four

m
Page 34 Group Project..3025,.
Report One

76

721~-

I.68 1-

6o. i solids

58.4% solids
0

55.0% solids

52 [

- ~~~~~~~~~47.2% solids
0 0 0 0~~~~~

4
44 I I I I I I I I I .1
70 50 90 100 110 120
TEMVIERATUPE Oc.

Figure 12. 'TemnperatUr~e-Compensated Response for Liquor C

Group Project 3025 Page 35
Report One

76 m

72 -

68 -

64 58.9% solids
H

~6o
54.9% solids
O 0

56

56
51. 7% solids
0~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

52

47.6% solids
0 0- 0 0_0-------~-0

48

II II II I I I I I g
44 I - - --

80 90 100 110 120

TEMPERATURE, C.

Figure 13. Temperature-Compensated Response for Liquor A

Page 36 Group Project 3025.:'
Report One

different solids levels. This would indicate that the amount of compensation

needed is not a strong function of solids content. The data in Fig. 13 for Liquor

A also show excellent temperature compensation. The largest difference in output

at any solids level is 0.4 of a unit on the meter over a 40C. temperature range.

There is a slight tendency toward a drop in indicated output at the high temperature

end. It should be noted that the TEMP pot was not adjusted for this particular

liquor. This would indicate that the effect of temperature on refractive index of

spent liquors is not a strong function of composition. This could be important

from the standpoint of industrial use of the instrument, because a simple setting of

the TEMP pot based on laboratory tests could .be adequate. One final point should

be noted in connection with the slight deviation present in Fig. 13. The tempera-

ture range over which the tests were carried out (40C.) is nearly twice as great as

the 40F. range for which temperature compensation is claimed in the instrument

specifications. - ;

SYSTEM VARIABLES

The effect of several other variables on refractometer response was also

examined. Flow rate through the refractometer adapter had no observable effect

over the range from dead stop to 50 gal./min. The pressure in the system was

varied from atmospheric to 60 p.s.i.g. with no observable effect on the reading.

The refractometer also did not appear to be affected by the presence of air bubbles

in the liquor. This was most apparent in initial runs with Liquor C when the

centrifugal pump was used. In these cases, liquor aeration was severe enough to

prevent sonic velocity readings, yet refractometer response was unaffected.

Stopping flow, which would permit bubbles to rise,-also did not affect the response.

An attempt was made to see how the refractometer response was affected by undis-

solved solids. This was not successful. The main problem was finding a method
Group Project 3025 Page 37
Report One

for characterizing the amount of undissolved solids in a circulating loop. A

second problem was a tendency for undissolved solids to plug the pump.

COMPOSITION EFFECTS

One of the main purposes of this investigation was to gain some insight

into the effects of liquor composition on refractometer response. It is clearly

shown in Fig. 11 that liquor composition can have a significant effect. At the

same actual solids content, Liquor B gives a reading which is 2.6 units higher than

Liquor A. This is of particular interest since both A and B were obtained from the

same mill, only five weeks apart, and the data in Table I indicate only minor

differences in composition. Based on the sulfated ash determination, the organic

content of Liquor B is 67.4%, while that of Liquor A is 66.0%. In addition, there

is a slightly higher degree of oxidation of the sulfur compounds in Liquor B. The

response line for Liquor C is generally below the other two curves and has a smaller

slope. Liquor C is from a different mill, has a smaller organic content and a

higher sulfur content than A and B. Thus, it is not too surprising that its

response curve has a different slope.

In order to gain greater insight into this behavior, the refractive index

was measured as a function of solids content for several solutions including single

components and mixtures. These data were taken on the laboratory refractometer

described previously at a constant temperature of 28C. Refractive index data were

obtained for Liquor C and for solutions of Na2 COs, sucrose, a 50-50 mixture of

Na2 CO 3 and sucrose, NaCl, Na2 SO4, oxalic acid, and phenol. These data are shown in

Fig. 14. It can be seen that Liquor C has the highest values of all the substances

tested. Refractive index'values fall off in the order Na2 CO3 , NaCl, sucrose,

Na2 S04 , and oxalic acid. Phenol has a high value for the solids content at which

the measurement was made. Examining these data, there is no obvious ranking of
Page 38 Group Project 3025
Report One

m
1.41.

Liquor C Sucrose

1.40

1.39 m

Na 2S04
1.38 -

;o
0

1. .37 -
O Liquor C
0 O NaC1
A Na2 CO0
F-1 * 50:50 Mixture of
Na2COs and sucrose

V Sucrose
d'
d
/
/Oxalic c/Acid
Acid
.
I+ Na2 SO4
X Oxalic Acid
^

1.35 m

/. ll Phenol
/
1.34 _

I I I I I I I I I I
1.33
0 10 20 30 40 50
PERCENT SOLIDS

Figure 14. Refractive Indices of Various Solutions

Group Project 3025 Page 39
Report One

organic and inorganic substances. Liquor C, which is mostly organic, has the

highest refractive indices, sucrose is significantly lower, while oxalic acid is

very low. Phenol, on the other hand, seems to have high values. The inorganic

compounds also vary considerably. Na2 CO3 is nearly as high as the black liquor,

while Na2 SO4 is much lower.

Further insight into composition effects can be gained by examining the

behavior of mixtures. This is most easily done with solutions of two pure compo-

nents, but it is also enlightening to look at the addition of various amounts of a

pure material to black liquor. Both approaches were used in this study.

Refractive index data for various mixtures of sucrose and Na2 CO3 at both 20 and 30%

solids are given in Fig. 15. It is seen that refractive indices of the mixtures

lie between those of the pure substances. However, the curves are not linear.

This would mean that a simple linear combination rule based on the addition of the

fraction of the refractive index of each pure substance would not predict the

refractive index of mixtures. Refractive index data for mixtures of black liquor

with sucrose and sodium chloride are given in Fig. 16. The data for liquor-sucrose

mixtures are at 60% solids. For solubility reasons, the data for liquor and salt

are at 25% solids. In both cases, the refractive index increases as the proportion

of liquor increases. This is not surprising since the data in Fig. 14 showed that

both sucrose and NaCl solutions had lower refractive indices than Liquor C at the

same solids content. With liquor and NaCl, a linear relationship is observed,

while a curved relation exists with liquor and sucrose. In general, it appears

that the refractive index of mixtures lies between the indices of the pure compounds,

but a linear combination of each component is the exception rather than the rule.

The tendency toward nonlinear combinations (indices of each component not additive)

seems to increase as the solids content increases.

 suoinnoS Soo?9Nq-asoao-nS go saoTpui aAV;OU.zjga '~i ~arLr9
(sani'os ao) asoaons Jimacuaai
OOT . ~L 0~.
g ~ 0
09Q'T
t19 Q T
SPTTOS %oa~ 9q1~'
9LQT-~A
tr
119 Q I
SPITOS 960:;r
gqq I
Q ~T
auo 4aodaHj
%og, cvoaCoaaj dnoao) 011 ~9s
 Group Project 3025 Page 41
Report One

solids)
1.46

1. 44'

1.38 solids)

1.36

1.34
0 20 40 60 100
PERCENT LIQUOR (OF SOLIDS)

Figure 16. Refractive Indices of Mixtures of Black Liquor and

Other Substances

m
mmmmommi
Page 42' Group Project 3025,
Report One

It is impossible to make any quantitative predictions about the effect of

liquor composition on refractive index. Black liquor is simply too complex to

permit this. The organic material originates from lignin and carbohydrates in the

wood and is at least partially substituted with sodium and sulfur. In addition,

there will remain a certain amount of cooking chemicals. Thus, the composition

(and refractive index) of the black'liquor would depend on wood species as well as

cooking conditions. The most simplistic treatment of liquor would have to consider

it made up of a ligneous fraction (aromatics), a carbohydrate fraction (sugars), and

the inorganics (NaOH, Na2CO3, Na2 S, etc.). If it is assumed that black liquor can

be described in this manner, then the data in Fig. 14, 15, and 16 would seem to

indicate that the high refractive indices of black liquor originate mainly with the

ligneous material. The inorganics tested and sugar (sucrose) all lie below the

black liquor line. Phenol, on the other hand, an aromatic compound, lies almost on

the black liquor curve. A means of correlating refractive index data for organic

compounds by means of an R factor is described in Table IV. Higher R values

give higher refractive indices. It should be noted that the presence of double or

triple bonds, aromatic structures, and rings tends to increase R values. It is

especially interesting to note the high R values associated with sulfur bonds.

The following generalizations, although by no means proved., are consistent

with all of the data examined. An increase in the inorganic content of a black

liquor will tend to lower the refractive index at a given solids content. An

increase in organically attached sulfur will tend to increase the refractive index.

The main contributor to the refractive index of black liquor is of ligneous origin.
Group Project 3025 Page 43
Report One

TABLE IV

REFRACTIVE INDEX OF ORGANIC COMPOUNDSa

The following equation can be used to correlate refractive indices of

organic compounds.

(n2 - ) M
R (n2 + 2) d

where:

n = index of refraction,
M = molecular weight, and
d = density, g./cc.

R , in units of cc./mole, is additive for bonds which are present.

A partial tabulation follows:

C 2.591 C6 H5 25.463 ethylenic bond 1.575

H 1.028 ClOH 43.00 acetylenic bond 1.977

=0 2.122 5S 7.729 4-member ring 0.317

>0 1.643 =S 7.921 3-member ring

OH 2.553

a
Taken from "Handbook of Chemistry and Physics," 4th ed.,
C. D. Hodgman, ed., Cleveland, 0., Chemical Rubber Pub-
lishing Co., 1958.
Page 44 Group Project 3025
Report One

Composition effects would certainly be an important factor in the

reliability of refractometers in the. field since changes in composition could be

read as changes in solids content. It appears very likely that a change in

organic/inorganic ratio would change the refractometer response. Wood species

and pulp yield would also play an important role. The effects of chemical

composition re-emphasize the need for proper calibration during use.

REFRACTIVE INDEX MEASUREMENT

An interesting observation was made during the refractive index measure-

ment with the laboratory refractometer. There are two methods for determining

refractive index with such a refractometer. In the first method, an incident beam

of light shines through a thin film of the solution, and the beam of light emerging

on the other side is observed through a lens system. This lens system is rotated

until the critical angle appears at a set of cross hairs in the eyepiece. What the

observer sees is a circle with the cross hairs at the center, and when set properly,

the lower half of the circle is dark and the upper half is light. The only differ-

ence in the second method is that the incident beam of light is on the same side of

the solution as the observation lens system. This permits the observation of

opaque solutions, and this is the same type of arrangement as is used in the EMC

refractometer. In observing the critical angle of black liquor by the second

method, it becomes immediately obvious that the line of separation between dark and

light is much less distinct than it is with transparent solutions using the first

method. In fact, when viewing through the eyepiece, it is sometimes difficult to

see the separation unless one sweeps back and forth through the critical angle. It

is difficult to quantitatively determine what effect this has on the EMC refractom-

eter precision; however, it does imply that the refractive index might not be

determined as precisely as is possible with transparent solutions.

Group Project 3025 Page 45
Report One

ON-LINE CALIBRATION

The basic response of the unit, showing a linear relationship between

refractometer output and solids content along with the demonstrated capability to

provide temperature compensation, indicates the potential of the refractometer for

reliable measurement of black liquor solids content in industrial use. The basic

requirement affecting reliability is the ability to maintain the proper calibration

of the instrument in the industrial environment. Accordingly, it is necessary to

consider the question of calibration in some detail.

There would appear to be three major factors which could cause significant

deviations between indicated output and actual solids content for an instrument

which is initially properly calibrated;

1. Build-up of deposits on the window,

2. Long-term drifts in the instrument, and

3. Changes in the chemical composition of the liquor.

Any solids which deposit on the window will interfere with proper measure-

ment because the optics will tend to see the deposits rather than the liquor. The

proper action to take in this case is to clean the. window and not adjust the

instrument calibration to compensate. Changing the calibration could cause erroneous

results. In the extreme case, solids deposits could completely isolate the liquor

from the optics. If solids deposits are found to cause trouble, a prism cleaner

should be installed and operated on a regular schedule. Provisions should also be

made for isolating the instrument from the flow line for inspection as needed.

Any long-term drifts in instrument response due to changes within the

instrument should be corrected by changing the calibration pot settings. No direct

evidence for such drifts was obtained in these tests. The instrument is stable for
Page 46 Group Project 3025
Report One

at least 24 hr., but stability tests over a period of weeks were not run. If the

drift is only in the zero, adjustment of the ZERO pot will correct for the drift.

This would require only a single determination of solids content. The problem is

more severe if there is drift in range or span, since a two-point calibration is

required to set ZERO and SPAN pots. If instrument drift is a more serious problem

than liquor composition variability (which is unlikely), it may be simplest to

periodically recalibrate the refractometer off-line with known solutions.

Changes in chemical composition of the liquor are probably the major

reason for refractometers getting out of calibration. The degree to which

composition changes cause difficulty depends on the detailed effect of composition

on response. It is possible to distinguish three cases:

1. A composition change that affects the actual value of refractive index at a

given solids level (hence the meter zero) but has little or no effect on the

change in refractive index per unit change in solids (slope of index vs. solids

curve or span) or the effect of temperature.

2. A composition change that affects the slope and intercept of the refractive

index-solids curve (hence zero and span) but which does not affect the

temperature coefficient.

3. A composition change that affects the entire refractive index vs. solids curve

and the temperature coefficient (hence zero, span, and temperature compensation).

The first case.can be handled by a simple adjustment of the ZERO pot. The second

involves adjustment of both ZERO and SPAN pots and hence a two-point calibration.

The third requires simultaneous adjustment of ZERO, SPAN, and TEMP.

Group Project 3025 Page 47
Report One

The first case is readily adapted to on-line calibration since it involves

single-point calibration. It is necessary only to take a sample of liquor while

recording the output reading, run an analysis of the liquor solids content, and then

adjust the ZERO pot to make up the difference between the indicated output and the

analytical solids content. If the effect of composition change is mainly to cause

a zero shift, this procedure would be adequate.

The second case is more complicated because adjustment of the ZERO and

SPAN pots requires a two-point calibration, and thus liquor analyses should be run

at both the low and high ends of the concentration range. This is not a very

suitable procedure for on-line calibration since it requires a significant change in

liquor solids content and freedom to vary liquor solids does not usually exist in a

mill situation. The time lag in analyzing solids content in liquor samples also

would cause problems in a two-point on-line calibration. There are two alternatives

which can be used here: Either the instrument would be calibrated off-line, adjust-

ing ZERO and SPAN, or it would be calibrated on-line, adjusting only the ZERO. In

the first case, two batches of liquor of known solids content would.be used in the

normal calibration procedure. The second approach could be used if the change in

response-curve slope was small, or if the solids content did not change much.

Under these conditions, only small errors result if the ZERO is kept in adjustment.

There is a hazard in this second approach. Over a period of time, the SPAN setting

may differ appreciably from the proper value, and large errors could occur if there

were large changes in solids content. For example, if a liquor is being monitored

at a normal solids content of 65%, and the calibration is maintained by ZERO adjust-

ment only, there could be considerable error at solids contents on either side of

65% (such as at 55% solids for automatic alarming purposes). If ZERO adjustment

calibration is used, the instrument should be taken off-line occasionally for ZERO

and SPAN adjustment.

Page 48 Group Project 3025
Report One

In the third case, where composition affects the temperature coefficient

also, the instrument must almost certainly be taken off-line for recalibration.

The interaction between ZERO and TEMP settings and the need to cycle temperature

practically requires off-line operation. Fortunately, the indications of this test

program were that composition changes had a relatively small effect on the amount of

temperature compensation needed. It would appear that the TEMP setting should be

checked occasionally when the instrument is pulled off-line for two-point ZERO and

SPAN adjustment.

The following summary statements regarding calibration of the refractometer

can be made.

1. The absolute accuracy of the refractometer in reading percent solids is deter-

mined primarily by the accuracy of the analytical procedure for solids measure-

ment employed in calibrating the refractometer.

2. The only suitable method for on-line calibration is adjustment of the ZERO pot

based on single-point analysis of solids content.

3. The instrument should periodically be removed from the line for two-point ZERO

and SPAN adjustment and to check on the ability to compensate for temperature

variations.

4. The calibration should not be changed to handle the effect of solids deposits on

the window. The window (prism) should be cleaned instead.

Group Project 3025 Page 49
Report One

NUS VELOCIMETER

PRINCIPLE OF OPERATION

The NUS velocimeter provides an indirect determination of black liquor

solids content through measurement of the velocity of sound in the liquor. The

instrument operates on the "sing-around" principle for measuring sonic velocity.

A sound velocity transducer transmits a sound pulse through the solution under test

to a reflector which is separated from the transducer by a fixed distance. The

pulse is then reflected back to the transducer which also serves to detect the

pulse. After detection of the reflected pulse, a new pulse is transmitted and the

sequence repeated. The result is a train of pulses at a frequency ("sing-around"

frequency) which is dependent on the sonic velocity of the solution.

The "sing-around" period (time from the start of one pulse to the start of

the next) is determined by the sum of the pulse transit time (path length divided by

sonic velocity) and the electronic delay time (time between detection of one pulse

and transmission of the next). This results in the following equation relating the

sonic velocity to the sing-around frequency.

c _ Af(l + at) (4)

7 - Bf X 10-l

where

c = sonic velocity in meters per second,

A = sonic path length in meters,

B = electronic delay time in microseconds,

a = coefficient of thermal expansion of the velocimeter probe

material, oC.-1,

t = temperature, C., and

f = "sing-around" frequency, hertz.

I
Page 50 Group Project 3025
Report One

The factor of 7 appears in the denominator because the actual frequency is electron-

ically multiplied by 7 in the instrument.

SPECIFI CATIONS

The manufacturer gives the following specifications on the Model 6161-139

Sonic Solution Monitor which was tested in this program.

Sound Velocity Range: 900 to 2250 meters/sec.

Operating Temperature Range:

Probe: 20 to 120C.

Electronics: 0 to 50C.

Output:

Sound velocity: 4-20 milliamperes and 0-10 volts.

Temperature: 4-20 milliamperes and 0-10 volts.

Repeatability:

Sound velocity: 0.025%.

Temperature: 0.05C.

Temperature Compensation:

First-order correction of sound velocity for temperature variations to

within 0.1C. Compensation coefficient is continuously adjustable

from 0 to 10.00 meters/second/oC.

Probe Material: 316 SS, platinum, and teflon.

Group Project 3025 Page 51
Report One

DESCRIPTION OF INSTRUMENT

The instrument consists of a sensing probe (Fig. 17) which mounts on the

process line and a control console (Fig. 18). The probe contains the sonic

velocity transducer, while the control console contains the electronics and signal-

conditioning equipment. The instrument also measures the temperature of the solu-

tion under test. This measurement can be fed to appropriate circuits to provide a

sound velocity signal which is compensated for temperature over a given range.

Figure 19 is a simplified block diagram showing the basic operation of the

instrument. The probe is inserted into the process solution and the transducer

transmits and receives sonic pulses. In the control unit, the sound velocity

electronics operate to produce the "sing-around" frequency. The S-V electronics

also monitor the attenuation of the sound pulse in passing through the solution.

The frequency and attenuation are examined and the information passed to the NORMAL

and OUT OF RANGE lamps on the front panel. The outputs from the S-V electronics

are the "sing-around" frequency and the attenuation signal (essentially amplitude

ratio of detected and transmitted pulse). The attenuation signal is routed to the

front panel meter where it can be read. The "sing-around" frequency then passes

through the SOUND VELOCITY RANGE selector (apparently a frequency divider) to a

scaling counter. In the scaling counter, the sound velocity signal is offset by

the value set on the LOW END ADJUST switches. These latter are set up to operate

off an octal code. The digital output of the scaling counter (essentially the

digital overflow beyond the offset) is routed to a digital-to-analog converter.

The resulting analog signal is sent to the meter where it can be read as sonic

velocity deviation and is also routed to the compensation electronics.

 (DWM

0 (
CD

(D

'C"

Figure 17. Photograph Of NUiS Sensing Probe

Group Project 3025 Page 53
Report One

Figure 18. Photograph of NUS Control Console

 p
OUTPUT
Q(D
IMA ~CONTROL UNIT

SOOwD VaLOurf OF
-
CC~~~~~~~~~~~~~~DVIATION - - -rm
- -
1-~~~~~~~~~- LOW990 ADJUST SMTCKES L~~~~~O II5SVAC C

SONICOUPT
NTROVEE THERN05TAT
RA45 ATTENUATION
ME.TER)PIN

1jN
UNCOMPENSATEO
SOUND VCLOC.IY4-O

'VMPERATUR.E TEMPERAIURE
TEMPERAUI ULCMESTO

PR~~~~~bt ~ONTROL
~~~ CONTACT.

S
) CTO*C
E -IQ
OONt~~~~/,I'W

TEMPERNSTION SPOINT

ITNPOUTPU
11s4OvT
BUFFER CURRSELET
-~~~~OLTItTo
t~OYp TEMP
OTAITATTON DE.AP OUTPOUTPU
J D ATONV TJPT
MO
D pA - - - - nf
~~~~~~~~~~~~~~~~~~~~
- -

SOJN,.OU CIT
C

X* S OOCT5C
TTS.tZOOUI S..C

Q) STANOARO MOOSTLUS95 ONElI

WTTNELrewt SWINTYCH.
-.
0

CD

Figure 19. Block Diagram for NUS Velocimeter

0 1 I,
Group Project 3025 Page 55
Report One

At the pipeline, the temperature-sensing element generates a d.c. signal

which is proportional to temperature. A temperature reference is also provided.

These signals are sent to the TEMPERATURE NULL control, which is used to null out

the temperature signal. The NULL control is a vernier helipot calibrated in terms

of degrees centigrade. The nonnulled portion of the signal is sent to the meter

where it can be read as the temperature deviation. By reading the null setting and

the temperature deviation, the temperature at the probe can be monitored. The

temperature deviation signal, which is proportional to the difference between the

actual temperature and the null setting, is also sent to'the temperature COMPENSATION

ADJUST. This varies the gain on the temperature deviation signal to generate the

desired sonic-velocity-related signal for compensation. The adjusted signal is

then sent to compensation electronics where it is effectively added to the sonic

velocity signal coming from the digital-to-analog converter. A switch is provided

to allow compensation upward or downward as needed as well as to shut off the com-

pensation. The signal from the compensation electronics is sent through SPAN ADJUST,

another variable gain, and then to the output position on the meter. This last

gain adjustment allows the operator to increase readout resolution or to set up the

reading for a particular set of units (such as percent solids).

The output meter has a dial indicator with two ranges on the dial ( -10 to

+10 and 0 to +10 ). The signal which is being displayed on the meter is con-

trolled by a METER SELECTOR switch. The four functions which can be monitored are

as follows:

1. S/V DEVIATION. This reading, in conjunction with the SOUND VELOCITY

RANGE selector measures the deviation in the sound velocity from the

value set in the LOW END ADJUST switches. The value on the meter

( 10 full scale) should be multiplied by the setting on the SOUND

VELOCITY RANGE selector for a true sound velocity deviation.

Page 56 Group Project 3025
Report One

2. ATTEN. This displays the attenuation of the sonic pulse in the

solution in percent of full scale. If attenuation exceeds 90%,

measurements should not be performed.

3. TEMP DEVIATION. This displays the deviation in temperature (up to

10C.) of the solution from that set on the TEMPERATURE NULL control.

4. OUTPUT. This displays the sonic velocity deviation signal after

temperature compensation (if selected) and after SPAN ADJUST. This

would enable the operator to monitor the concentration of the solution

after the instrument has been adjusted.

The controls which are available to the operator include the following:

1. SOUND VELOCITY RANGE Selector. This brings the Control Unit into the

mode most compatible with the sound velocity range of the solution

under test. This control works directly with the S/V DEVIATION

display.

2. LOW END ADJUST Switches. These enable the operator to select the

operating range of the instrument by offsetting the sound velocity

span which is below the low range of the solution being monitored.

The numbers on the switches (which are in octal) may be used, together

with the S/V DEVIATION reading, to determine an actual sound velocity

value.

3. TEMPERATURE NULL Control. This enables the operator to select the

average temperature to which readings are corrected. This control,

in conjunction with the TEMP DEVIATION reading, permits determination

of the process temperature.

Group Project 3025 Page 57
Report One

4. TEMP COMP SELECT. 'This is a switch that connects or disconnects

-temperature compensation, and also the polarity of the compensation

( + for a positive temperature coefficient and - for a negative

temperature coefficient).

5. COMPENSATION ADJUST. This enables the operator to set the correct

amount of temperature compensation. After the setting has been made,

the control scale will indicate the temperature coefficient of the

solution.

6. SPAN ADJUST. This permits a change of gain on the output meter to

either increase readout resolution or indicate solids content changes

directly. The action is one of dividing the gross span by any number

between 0 and 10.00. If no adjustment is to be made, this control is

set at 1.00.

A convenient feature of this instrument is that all settings and adjust-

ments are made on calibrated switches or dials. Once a calibration procedure has

been performed, it can be precisely reset at'a later date for exactly the same

operating conditions. Even more significant is the fact that the numbers on the

dials (and switches) have actual physical meaning. For example, the TEMPERATURE

NULL dial markings are equivalent to degrees centigrade, COMPENSATION ADJUST

markings are equivalent to m./sec./C., and SPAN ADJUST markings give the divider

factor on gross span. The numbers on the LOW END ADJUST switches are an octal

representation of a number which is proportional to "sing-around" frequency and can

be used to compute the mean sonic velocity setting.

In summary; it is possible to read the following quantities with this

instrument:
Page 58 Group Project 3025
Report One

1. Sonic velocity of the solution in meters per second. This is

obtained by reading the sonic velocity (S/V) deviation on the meter,

multiplying by the setting on the S/V range switch, and adding the

result algebraically to the value of sonic velocity corresponding to

the setting of the low end adjust switches.

2. The percent attenuation of the sonic pulse sent through the solution.

This is read directly on the meter when the meter selector switch is

set on attenuation.

3. Temperature of the solution in degrees centigrade. This is obtained

by reading the temperature deviation on the meter, adding it algebra-

ically to the vernier reading on the temperaturenull control dial,

and adding 20C.

4. A temperature-compensated output proportional to some quantity which

in turn is proportional to sonic velocity (such as percent solids).

This would be read directly on the meter after proper calibration and

adjustment of the S/V range switch, low end adjust switches, tempera-

ture compensation adjust, and the span adjust.

In addition to the outputs and controls normally provided with the instru-

ment, a connection was made to tap the "sing-around" frequency signal and send it to

a Hewlett Packard 5512-A digital counter. This provided additional information and

a way to get sonic velocity data independent of the meters.

INSTRUMENT CHECKS

A series of checks on the effect of environmental conditions on the NUS

velocimeter were carried out simultaneously with the checks on the refractometer
 Group Project 3025 Page 59
Report One

described previously. These runs were made with sucrose solution in the loop and

included tests on the effects of line voltage variations, stray magnetic fields, and

mechanical vibrations. It was found that the sonic velocity deviation reading was

unaffected by line voltage variations from 95 to 125 volts. The reading was also

unaffected by the presence of a magnetic field near the probe, console, or connecting

cable. Mechanical vibration had no noticeable effect on the meter reading. It

was noticed, in a later test with the probe lying on a table, that the arcing from a

motor running nearby caused the meter to go off scale. In this particular situation,

the probe did not have the benefit of being shielded by the mounting flange and tee
section.

An overnight (24-hour) stability check was run on the velocimeter concur-

rently with the test on the refractometer. The results of that test are given in

Table V.

TABLE V

TWENTY-FOUR-HOUR STABILITY TEST ON NUS VELOCIMETER

Time of S/V Counter, NUS Temp.,

Day Deviation Hz oC.

1630 .0.0 134920 32.1

1940 -0.5 134750 29.5
2240 -0.7 134630 27.8
0805 -0.9 134540 27.0
0900 -0.9 134550 27.0
1015 -0.9 134570 27.2
1200 -0.8 134600 27.6
1300 -0.7 134630 28.2
1500 -0.6 134690 28.7

I
Page 60 Group Project 3025
Report One

The performance of the instrument was monitored by observing the readings

on the sonic velocity deviation meter as a function of time. Although the tempera-

ture compensation selector switch was turned off during this test, the sonic

velocity deviation reading would not be affected by compensation anyway. The

temperature of the sucrose solution in the loop cycled somewhat as the room cooled

overnight and then reheated. All of the observed changes in the meter reading are

essentially caused by the temperature changes. This is readily apparent in Fig. 20,

in which the sonic velocity deviation meter readings and the counter readings are

plotted versus the temperature of the solution. Since the counter reading is

essentially the "sing-around" frequency, the change in the meter reading with

temperature is reflecting actual changes in sonic velocity with temperature. Thus,

the instrument appears to be stable over at least a twenty-four-hour period.

TEMPERATURE MEASUREMENT

A check on the NUS temperature-measuring system was carried out early in

the program. An iron-constantan thermocouple (Conax) was installed in the flow

loop immediately upstream of the velocimeter probe. An ice bath was used as the

cold junction, and the thermocouple emf was read with a Cambridge Portable

Potentiometer No. L-316688. The NUS temperature was determined by adjusting the

temperature null control until the temperature deviation read zero, and then adding

20C. to the vernier readings on the null control dial. This check on the

temperature-measurement system was carried out with 50% sucrose solution circulating

in the loop. The results of this test are presented in Table VI. It can be seen

that the NUS instrument tended to read a slightly lower value over the whole range

from 30 to 120C. The deviation is largest at the extremes (about 1C.) and is

smaller in midrange (about 0.5C.). It should be noted that the instrument speci-

fications make no claim on accuracy of temperature measurement. What is specified

Group Project 3025 Page 61
Report One

0. 0 -

-0.2

r)

-0.8

-1.0 I I I I

135 5, 000 r-

1354, g00 F
w
134 ,8oo F

0
PI)
154,700 I-

134,600 F

134, 500 ".

I I I I I i
26 27 28 29 50 31 52 55
TEMPERATURE, 0C.

Figure 20. Figur20.Data

from Stability Tes~t on NUS Velocimeter
Page 62 Group Project 3025
Report One

is a repeatability of 0.05C. and the ability to provide first-order correction of

sound velocity for temperature variations to within 0.1C. The instrument certainly

appears capable of doing this. Because of the reasonable agreement between the NUS

temperature measurement and the thermocouple and the convenience of the NUS method,

the NUS instrument was used for all temperature measurement in the flow loop.

TABLE VI

ACCURACY OF NUS TEMPERATURE MEASUREMENT

Thermocouple Thermocouple NUS Temp., NUS - T.C.,

emf, my. Temp., C. C. C.

1.75 34.2 33.3 -0.9

2.73 52.8 52.3 -0.5
3.75 71.8 71.4 -0.4
4.45 84.8 84.2 -0.6
5.27 100.0 99.1 -0.9
5.72 108.3 107.2 -1.1
6.42 121.2 119.9 -1.3

OPERATIONAL VARIABLES

The effects of several operational variables on instrument response were

examined. These included the flow rate of liquor past the probe, system pressure,

and the presence of entrained air in the liquor. It was found that the instrument

was unaffected by variations in liquor velocity past the probe over the range from

0 to 5 ft. per second. A small effect due to system pressure changes was found.

This was observed with Liquor A at 15% solids and a temperature of 27.9C.

Increasing the total pressure in the system from 0 to 60 p.s.i.g. caused an increase

in the "sing-around" frequency measured by the counter from 131,770 Hz to 131,800 Hz,

or an increase in sonic velocity of 0.023%. This would amount to a change in sonic

velocity of about 0.0004% per p.s.i. This observed effect seems to be quite small
Group Project 3025 Page 63
Report One

and would not seriously-affect the use of the instrument. If maximum.measurement

precision is desired, variations in total pressure should be avoided. It should be

noted that a pressure effect was not unexpected since sonic velocity depends on the

density and.bulk modulus of the solution, and these could be pressure dependent.

Entrapped air seriously affected the operation of the velocimeter, partic-

ularly at high solids contents. Aeration of the liquor resulted in an attenuation

of the sound pulse by more than 90%, thus causing the out-of-range light on the

front panel to come on. In all cases when air bubbles were present, the instrument

went out of range rather than simply giving an erroneous reading. Observation of

the attenuation signal provided a ready means of detecting aeration problems. Under

normal operating conditions, the attenuation of the sound pulse was low,.on the order

of 5% or less. When bubbles were present, the attenuation would rise to, over 90%.

Under some conditions, the attenuation reading might bounce from very high to very

low values. Pressurizing the flow loop to 60 p.s.i.g. did not overcome the air

bubble problem. It is not known if conditions could occur where aeration would

cause significant changes in sonic velocity without also causing sufficient attenua-

tion to trigger the out-of-range signal. In all of these tests, aeration did cause

the out-of-range light to come on.

In this program, the problems associated with aerated liquors were over-

come by preheating the liquor to a boiling condition in the steam-jacketed vessel

(see Fig. 1) which allowed the air bubbles to rise to the top, and by eliminating

the use of the centrifugal pump for recirculation. All of the liquors which were

used in this program were aerated when received from the mills, most probably because

of aeration during direct contact evaporation. The aerated nature of the liquor as

received could be seen by observing the alternate expansions and contractions in

volume of a liquor sample exposed to vacuum and atmospheric pressure. Thus, the
Page 64 Group Project 3025
Report One

problem with aeration in this program had to be solved by first deaerating the liquor

and then preventing reaeration. The deaeration was carried out by heating to a boil

while stirring in the steam-jacketed vessel. The loop was then filled from the

bottom by pumping the liquor from the vessel with the Moyno pump. When the loop

was filled, the valving was adjusted so that the Moyno pump served as the circulat-

ing pump also.

In nearly all cases where a high attenuation of the sonic pulse occurred,

the problem could be ascribed to air bubbles. In the flow loop, there were no

indications that black liquor itself became opaque to sound waves. No discernible

trend of increased attenuation with increasing solids content was observed. The

only evidence of liquor opacity was obtained in some beaker tests with Liquor C.

At temperatures of 40C. and solids contents above 50%, attenuation of the sound

pulse became significant. This is apparently associated with the high viscosities

of low-temperature, high-solids liquors. Otherwise, the cause of high attenuation

was air in the liquor. Zacharias (3) has stated, "Bubbles and solid particles

larger than 50 microns may cause erroneous readings. Also, emulsions containing

large liquid droplets may be opaque to sound at high concentration levels, and

slurries with large quantities of suspended particles may scatter the sound

excessively." This could indicate that undissolved salt cake could also cause

difficulties. The effects of the presence of undissolved salt cake were not tested

for the reasons described earlier in the section on the refractometer.

BLACK LIQUOR MEASUREMENT

There are, at a minimum, two requirements which must be met for the sonic

velocimeter to be useful for measuring solids content of black liquor. First, the

instrument must be capable of measuring sonic velocity in black liquor, and second,

there must be a reasonable functional relationship between the sonic velocity of the
Group Project.3025 Page 65
Report One

liquor and the solids content. The first question was alluded to in the previous

section. As long as air bubbles are not present in the liquor, the instrument does

measure sonic velocity and the black liquor is not opaque to the sound pulse. The

second question is best examined by determining the sonic'velocity response curves

of black liquor as functions of solids content and temperature.'

The most complete set of'data on the sonic velocity response was obtained

with Liquor A. Some additional data were obtained with Liquor B. When Liquor C

was used in the flow lobp, the'problem of deaerating the liquor had not yet been

solved. In these early runs, the velocime.ter was out of range nearly all of the

time and the only sonic velocity data on Liquor C were obtained in small beaker tests

at 40 C. The sonic velocity data on Liquor A are shown as a-function of temperature

at different solids contents in Fig. 21. A cross plot of these data to show the

dependence on solids content directly is presented in Fig. 22. Similar data for

Liquor B are given in Fig. 23 and 24. ,'

There are several features which are apparent in these data ~. The sonic

velocity decreases with temperature over the full temperature and solids range.

Curves of sonic velocity vs. temperature are steepest and essentially linear at the

highest solids contents and become less steep and.slightly curved at the low solids

end. No maximum in the sonic velocity-temperature curve was observed over the

range of variables tested.. These data indicate that the temperature dependence of

sonic velocity in black liquor is better behaved than might be expected. Zacharias

(5) says that aqueous solutions or emulsions generally exhibit nonlinear character-

istics in which velocity increases to some maximum value and then decreases. As

concentration increases, the value of sound velocity would change so that the

maximum would occur at lower temperatures. The data in Fig. 21 and 23 show that if

a maximum exists at all, it exists outside the range of normal interest. The kind
page 66 Group Projedt~3025
Report One

1900

1800 [-:

a)
ED

I-I

03

1700 F-

1600

I I I I I I I I I I
114o 160 18o 200 220 240 260
TEMPERATURE, OF.

Figure 21. Sonic Velocity Data for Liquor A

Group Project 53025 Page 67
Report One

1900 I 800F. - .
r

2000 F.

22 00F.

, 240 0 F.

1800 I-

Ea
a,
Q)

100

i1600

I I I I I I I I I I I I
10~ 20 30 40 50 6o 70
SOLID~S CON~TENT, %
Figure 22. Sonic Velocity vs. Solids Content for Liquor A
Page 68 Group Project. 3025
Report One

1900 s

56.1% solids 4

4r

a)
cH)17 0 0

O0

o-
45.5%

1600
..

16o 180 200 220 240 260 280

TEMPERATURE, OF.

Figure 23. Sonic Velocity Data for Liquor B

Group Project 3025 Page 69
Report One

1900,-

180 0 F.

, 2000OF.

,2200 F.

18001-~

w
.- ,.
240 0 F.

.r-4 ,..26o 0F.

0
'-I

C-,

0
cQ 1700 . 28O0'F.

16oo l-

I I I I I I I I I I
20 30 40 50 60 70
LIQUOR SOLIDS, %

Figure 24. Figur 24.Sonic

Velocity vs. Solids Content for Liquor B
Page 70 Group Project 3025
Report One

of temperature dependence shown in these data should make it easy to provide

temperature compensation because little error would be introduced by linearizing

the curves over a 20-40F. range.

The relationship between sonic velocity and solids content is shown in

Fig. 22 and 24. The curves tend to be steepest and most linear at the low solids

end and then bend over and become curved and less steep at high solids. Both

linearity and steepness increase at lower liquor temperatures. Slopes of the

response curves for Liquor A span the range from 6.7 to 3.0 m./sec./% solids. At

the low solids end, linearization of the curves over a wide range of solids contents

would be adequate. At high solids, linearization would be valid only over a small

(say, 2.5% solids) range of solids contents. It is evident that the sonic

velocity response of black liquor is sufficiently straightforward to permit its use

for monitoring solids content.

A comparison of Fig. 22 with Fig. 24 shows that composition of the black

liquor affects sonic velocity response also. The curves for Liquor B fall below

and to the right of Liquor A. This means that at the same solids content and

temperature Liquor A gives a higher sonic velocity than Liquor B. This effect of

composition has several implications. First of all, it is fruitless to attempt to

discover an algorithm for describing the concentration and temperature dependence of

sonic velocity in black liquor because it would change with each change in liquor

composition. Secondly, since liquor composition would be expected to vary, it

would be necessary to continuously maintain calibration during on-line operation.

Composition effects will be discussed in more detail a little later, as will the

problems of on-line calibration.

Group Project 3025 Page 71
Report One

TEMPERATURE COMPENSATION

The effectiveness of the temperature compensation circuitry which is

provided in the velocimeter was checked over the specified temperature range of

10C. using Liquor C and a midrange temperature of 90C. The adjustments of the

amount of compensation were made on a high solids content liquor which was then

progressively diluted. Unfortunately, no direct determinations of liquor solids

content were made for this series, and so all of the data are on a relative basis

ranging from about 60% solids down to 30% solids.

The temperature compensation is not a factory calibration and it is

necessary to set it up according to instructions provided in the instrument manual.

The liquor is brought to the desired midrange temperature, and the TEMPERATURE NULL

set so that the temperature deviation reads zero. The LOW END ADJUST switches are

then set so that S/V deviation reads zero. The product temperature is then changed

to an extreme point and the temperature COMPENSATION ADJUST turned to bring the

output meter back to zero. This inserts the correct amount of temperature compen-

sation.

The results of the compensation test are shown in Fig. 25. It is evident

that for the sample for which the compensation was set up [number (1)] excellent

compensation is obtained over the 10C. range. There were no deviations over the

compensated range. With the progressively diluted liquors, the instrument was

overcompensating for temperature. This simply reflects the fact that the sonic

velocity vs. temperature curves are not parallel but rather are steeper at high

solids. Thus, the compensation settings are valid only for a narrow range of

solids content. A comparison of the compensated and uncompensated response shows

that the compensation circuitry is indeed important.

Page 72 Group Project 3025
Report One

1910

High
1890
solids

>110

18t6

1850
I K
4

0~ i3
a)

U)

0 1810

0
X Uncompensated \
1790 0 Compensated
Compensated
Range

1770

1750 Low solids

I
173 0 L. I _ I - L1_ I I fI I I

70 80 90 100 110
TEMPERATURE, oc.

Figure 25. Temperature-Compensated Response of Velocimeter

Group Project 3025 Page 73
Report One

COMPOSITION EFFECTS

The investigation of the effects of chemical composition on velocimeter

response paralleled the study on the refractometer. Figure 26 clearly shows that

composition effects are present. This curve shows the 180F. lines for Liquors A

and B taken from Fig. 22 and 24. The response curve for Liquor B lies below and

to the right of the curve for Liquor A. This same behavior occurs at other

temperatures. Thus, at the same temperature and solids contents, Liquor A gives a

higher sonic velocity. Sonic velocity data for Liquor C are also shown in Fig. 26.

Unfortunately, these data were taken at 40C. (104F.) so that a direct comparison

with Liquors A and B cannot be made. In order to make such a comparison, the data

in Fig. 21 were used to estimate the change in sonic velocity with temperature from

180 to 104F. at various solids contents. These estimated values were then sub-

tracted from the Liquor C curve to give an estimated curve for Liquor C at 180F.

This latter curve is shown as the dashed line in Fig. 26. This line indicates that

Liquor C would give a higher value of sonic velocity than either A or B at the same

solids content and temperature. Thus, the data indicate that sonic velocity

decreases (at a given solids level and temperature) from C to A to B. It is of

interest that this is essentially the same order in which refractive index increases

(see Fig. 11).

In order to gain some insight into why these changes occurred, sonic

velocity data were obtained on some simple solutions and mixtures. These data were

obtained outside of the flow loop at 40C. Sonic velocity data were obtained for

Liquor C, various mixtures of sucrose and Na2 CO3 ,.Na 2 SO4, NaCl, oxalic acid, and

phenol. These data are shown in Fig. 27. Several trends are apparent in these

data. The general form of the curves is somewhat nonlinear, with an increasing

slope at higher solids content. This may be contrasted with the data on black
Page 74 Group Project 3025
Report One

1900
Liquor C

1800 -
4,

u~
C-
O
k
a)
0)

1700

Liquors A and B at 180F.

Liquor C at 104C.

Dashed line is for Liquor C

extrapolated to 180F.

1600 _
9'
I I I I I I I I I I I
0 10 20 30 40 50 60
SOLIDS CONTENT, %

Figure 26. Effect of Liquor Composition on Sonic Velocity Response

Group Project 3025 Page 75
Report one

'0:50
Na2003

1900

~..-Liquor C

75:25

1800

C)

a)

Sucrose

0.
0 .
3.700

0)
Cl

i6oo o Mixture s of' sucrose ana

Na2 CO3
ALiquor C
o3 NaCl
+ Na2 SO4
oxalic X Oxalic acid
M Phenol

I I I I I I I I I I I I
1500
0 10 20 30 )40 50 6o
SOLIDS CONTENT, %

Figure 27. Sonic Velocity Data for Various Solutions

Page 76 Group Project 3025
Report One

liquor in Fig. 22 and 24 taken.at higher temperatures and solids levels showing a

decrease in the slope at higher solids values. Another feature in Fig. 27 is that

all of the curves seem to be converging to about a value of 1525 m./sec. at zero

solids content, which is the approximate sonic velocity of water at this temperature.

Perhaps the most striking feature of these data is the distinct separation between

organic and inorganic compounds. Solutions of Na2CO3 , Na2 SO4, and NaCl all have

high values of sonic velocity, while the organics - sucrose, phenol, and oxalic acid -

all have relatively low values of sonic velocity. The curve for Liquor C, containing

organics and inorganics, lies somewhere in between. This behavior may be contrasted

with the index of refraction data on these same materials (Fig. 14) which did not

show trends of this nature.

The data in Fig. 28 showing sonic velocities at various proportions of

sucrose and Na2 CO3 and those in Fig. 29 for mixtures of black liquor and sucrose and

NaCl all show the same basic response. The sonic velocity of these organic-inorganic

mixtures increases as the inorganic content increases in a linear manner. The sonic

velocity of the mixtures is a linear combination of the sonic velocities of the pure

components in all four curves. This may again be contrasted with the data on

refractive index which tended to give nonlinear results. It should not be concluded

that a linear combination law is generally valid for the sonic velocity of mixtures.

However, it does appear that sonic velocity is more susceptible to linear addition

of each component than refractive index.

On theoretical grounds, the sonic velocity, c , in any liquid is related

to density, p , and bulk modulus, B , the reciprocal of compressibility, by the

expression c = (B/p)y2 . In order to gain some understanding of what changes

occur in solutions, the sonic velocity data were plotted versus specific gravity as

shown in Fig. 30. These data show that the sonic velocity increases as the specific
 Group Project 3025 Page 77
Report One

1900

25% solids

1800
o
U

CQ

20% solids

1700
U
?-1

1600

20 40 60 80 ' 100
Na2 COs PERCENT SUGAR (ON SOLIDS) Sucrose

Figure 28. Sonic Velocities for Sucrose-Na 2 CO s Solutions

Figure 28.

h
Page 78 Group Project 3025
Report One

2000 Sucrose-liquor
60% solids

1900
H
0O

co
CD

1800

NaCl-Liquor
25% solids

I
1700
0 16.7 33.3 50 66.7 83.3 100
PERCENT LIQUOR IN SOLUTE

Figure 29. Sonic Velocities of Mixtures of Black Liquor

and Other Substances
 Group Project 3025 Page 79
-Report One

Na2 003

1800l-
0 NaCl

0 Na2 sO4

1700 I-

U)

a)
4-p Liquor C
a)

03 sucrose

Cl)
16o0
I

I
i

0
Phenol

I
I

I
1500 I I -I
1.000 1.1000 1.2000
SPE CI FI C GRAVI TY

Figure 50. Sonic Velocities as Functions of Specific Gravity

Page 80 Group Project 5025
Report One

gravity of the solutions increases, but along different curves for different sub-

stances. This is very interesting because the equation indicates that if the bulk

modulus, B , remained constant, then the sonic velocity would decrease as specific

gravity increased. The fact that the sonic velocity of solutions increases as

specific gravity increases (both due to an increase in solids content) means that

the bulk modulus increases more rapidly than density ( B/p increases) as solids

content increases. Hence, the measurement is responding more to changes in bulk

modulus than to changes in density. In particular, this indicates that sonic

velocity is a unique property of a given solution (like refractive index or density)

and is not simply another way.of measuring density. A sonic velocimeter and a

densitometer are not equivalent.

The fact that the various methods of measuring solids content of liquor

are susceptible to errors due to changes in liquor composition is a consequence of

their employment of indirect methods for determining solids. The problem arises

because the sonic velocity (or refractive index) at a given solids content and

temperature of a solution is a function of its composition. This would be expected

to be true at least to some degree for all indirect methods of determining black

liquor solids content. In order to illustrate this point, some handbook data on

specific gravity as a function of solids content are shown in Fig. 51. The data on

black liquor are taken from the TAPPI data sheets. It is obvious that the use of

solution density to determine solids content is fraught with the same problem as is

the use of sonic velocity or refractive index.

One advantage to the sonic velocimeter may be that the interpretation of

composition effects appears to be much more straightforward than it is for the

refractometer. There seems to be a definite correlation between high sonic

velocities for inorganic solutions and lower velocities for organic solutions.
 Group Project 3025 Page 81
Report One

1.300r-

Black
liquor
Na2 SO4

Sucrose
1.2 00F

0r,
Na2 CO3
U
m
m

1.100 k

1.000o L I I I I I I I I I I I I
0 10 203 4ll
0 50 60
SOLIDS CONTENT, %

Figure 51. Specific Gravity Data for Various Solutions

k I I mmmi
Page 82 Group Project 3025
Report One

It seems quite reasonable to assume that the major variable affecting sonic velocity

in black liquor is the organic/inorganic ratio. Although refractive index also

responds to changes in organic/inorganic ratio, the dependence is not as clear-cut

or direct.

ON-LINE CALIBRATION

Since changes in liquor composition affect the sonic velocity response,

and it is essential to avoid interpreting compositional changes as changes in solids

content, it would be necessary to maintain instrument calibration during on-line

operation. As with the refractometer, the major problem would seem to lie with

changes in the functional dependence of sonic velocity on solids content and temper-

ature rather than with drifts or other problems in the instrument itself. Thus, a

program for on-line calibration is a major factor in obtaining reliable solids

monitoring of black liquor.

The NUS velocimeter is a versatile instrument which can give several

measurements on the solution under test. These include sonic velocity, temperature,

and sound pulse attenuation in the solution. It can also be set up to read

directly some other quantity (such as percent solids) which is a function of sonic

velocity. This latter signal can be compensated for temperature variations. It

is, of course, such an output which would be used in monitoring solids content of

black liquor. This instrument is quite easy to use for measurement of sonic

velocity and temperature. It is also easy to set up for reading solids content if

the sonic velocity of the solution as a function of solids and temperature is known.

Because of the effects which liquor composition may have on sonic velocity response,

it is inadvisable to rely on a single set of curves (such as Fig. 21 or 22) in inter-

preting instrument response. It would be necessary to take periodic samples of the

Group Project 3025 Page 85
Report One

liquor and run a direct determination of solids content, and then use this informa-

tion to adjust the calibration of the velocimeter.

The basic problem in on-line calibration is making the necessary adjust-

ments on minimal information. Obviously, if a complete set of sonic velocity

response data was obtained for each sample, the instrument could be easily readjusted.

The calibrated dial settings on the velocimeter pots aid such an adjustment. How-

ever, it is not possible to run a full set of calibration curves on each sample, and

less data would be available for changing the calibration. Normally, only the actual

solids content of the liquor sample and the velocimeter readings would be known.

The adjustments needed to set a calibration are a zero, a span, and the

amount of temperature compensation employed. The zero amounts to selecting a

particular meter reading to correspond to a particular solids content (usually the

lowest expected) and is basically determined by the LOW END ADJUST switch. The

span is the change in output (meter reading) which corresponds to a given change in

solids content (usually selected so that the meter covers the full range of solids

expected) and is controlled primarily by the SPAN ADJUST pot. The amount of

temperature compensation is determined by the gain applied to the temperature

deviation signal which is controlled by the COMPENSATION ADJUST pot. The setting

of the S/V RANGE switch affects both the zero and span and would not normally be

changed during a minor adjustment of the calibration.

The adjustments required to maintain calibration depend on the degree to.

which the sonic velocity response function has changed. As with the refractometer,

this could take on three forms: adjustment of zero only, adjustment of zero and

span, and adjustment of zero, span, and temperature compensation. The first would

require only a single-point calibration, and the second a two-point calibration.

The third would normally require only two different concentrations to set zero and
Page 84 Group Project 3025
Report One

span plus the freedom to vary temperature; however, the manufacturer recommends a

three-point calibration since the temperature compensation should be set at the mid-

range concentration.

Examination of the data in Fig. 26 showing sonic velocity vs. solids data

for the three different black liquors indicates that the major effect of composition

changes will be a zero shift. The slope of the sonic velocity vs. solids curves

(the span) appears to be less affected. If the instrument is set up to monitor

solids over a reasonably narrow range (say, 50-60% solids), then the main effect of

composition changes would be a zero offset. The proper corrective action in such

a situation is a change in the LOW END ADJUST switches, and not a change of the SPAN

ADJUST pot. When the liquor sample is taken for a laboratory measurement of the

o.d. solids, the S/V deviation, temperature, attenuation and output readings should

be recorded along with the settings on the pots and control switches. When the

true solids content value becomes available, it should be compared with that obtained

from the meter output. The sonic velocity difference associated with any discrep-

ancy between the indicated and true solids content should be calculated and used to

make the necessary change in the LOW END ADJUST switches. This would not be the

simplest procedure because the switches are based on an octal code and a base

conversion would be necessary. Over a period of time, this procedure may lead to

significant error in the span, and the instrument should be taken off-line for full

recalibration of zero, span, and temperature compensation according to the manu-

facturer's recommended procedure.

6
 Group Project 3025 Page 85
Report One

DIRECT MEASUREMENT OF SOLIDS CONTENT

A common theme running through much of the previous discussion is that the

instrumental methods for determining solids content are indirect methods based on

the measurement of some property dependent on solids content, and that these instru-

ments must be calibrated by means of a direct measurement of solids content.

Regardless of the stability and precision of these instruments, their ultimate

accuracy is no greater than the accuracy of the solids determination employed in

calibration. Since all methods for monitoring solids content are ultimately based

on some direct method, some attention must be paid to this subject. This report on

the refractometer and velocimeter would be incomplete without some discussion of

direct measurement of liquor solids.

The use of the term solids content or concentration implies that black

liquor can be considered to consist of two substances: something called liquor

solids, and water. It is, of course, possible to make this distinction, and the

solids are then identified with the total nonaqueous constituents of the liquor.

However, it is important to make a distinction between the nonaqueous constituents

and the residual solids remaining after evaporation of the moisture from the liquor.

These are not normally the same since volatile matter other than water may be removed

during evaporation. Because of the loss of volatiles, residual solids should always

be less than the total nonaqueous constituents. It is a somewhat philosophical

question to decide which quantity is truly desired, since black liquor is evaporated

as part of the firing procedure. No attempt will be made to resolve that question

in this report.

Parker et al. (2) discuss the measurement of black liquor solids in some

detail. They list two basic approaches to the problem. The first and most common

is to measure the material in the liquor which is nonvolatile under specific test

I
Page 86 Group Project 3025
Report One

conditions. This might be called the residual solids approach. The second

approach is to measure the water content of the liquor and calculate the solids by

difference. Residual solids methods include those based on the use of drying ovens

(e.g., TAPPI procedures) as well as the use of infrared lamps and a moisture balance

(e.g., Cenco moisture balance). The amount of water in black liquor can be deter-

mined by titration using a modified Karl Fischer method (4). Another approach to

determining the amount of water present is an azeotropic distillation with an

organic solvent. Various procedures employing distillation for black liquor solids

measurement have been developed.

RESIDUAL SOLIDS METHODS

The main problem in evaporating liquors to a solid residue for gravimetric

determinations of solids contents are as follows:

1. The liquor may form a scum which greatly retards the rate of water removal.

This results in excessive evaporation times and leads to erroneous high values

for solids content due to unevaporated moisture.

2. There may be thermal decomposition of the liquor and. subsequent volatilization

of organic materials, especially if long drying times or excessive temperatures

are employed. Volatilization would tend to cause measured values of solids

content lower than the true values.

3. In the presence of air, oxidation of various components of the liquor may take

place. Since the oxidation would not normally occur to the extent to which

volatile combustion products are formed, the effect of the oxygen uptake would

be to cause erroneously high values for solids contents.

Group Project 3025 Page 87
Report One

Oven-drying procedures have been the basisfor TAPPI methods for black

liquor solids determinations. These procedures have evolved to try to-overcome

the problems mentioned above. Parker et al. discuss the latest Proposed Method in

their paper. In this revised procedure, black liquor samples are dried at 105C.

for a minimum of six hours. An inert surface extender.(such as sand or alundum)

and a controlled flow of dried air are used to increase drying rate and eliminate

moisture entrapment. Strong black liquors are diluted to allow volumetric handling

and to reduce scum formation. The

.. procedure can be summarized as follows.

1. Dry and weigh 25-30 g. of inert material in a wide, shallow glass container.

Add sufficient liquor sample to give 1-3 g. of dry solids and weigh. Add

10-20 ml. of water to dilute and distribute samples in excess of 30% solids.

2. Dry the liquor for a minimum of 6 hours in a gravity convection oven controlled

at 105C. 3C. with an air space of one cubic foot or less, adapted to pass

dried, preheated air at a replacement rate of about twice per minute. After

the initial:drying and weighing, drying is repeated, for one-hour intervals until

the weight loss is less than 0.1% solids per hr.

3. Duplicate determinations on the same sample should differ by no more than 0.3%

solids.

It is important, when using this procedure, to keep the sample size within

the recommended range. An excessively large sample will lower the surface/volume

ratio of the liquor and lead to scum formation.. All tests should be run in dupli-

cate. This procedure is not intended as a rapid or routine control method. It

may be used to calibrate such procedures.

One of the big drawbacks to oven-drying procedures is the length of time

needed to carry out the test. TAPPI Method T 625 ts- 64 called for a 24-hour drying
Page 88 Group Project-5025
Report One

procedure. 'The revised procedure discussed above would still require from 7 to

10 hours to obtain results. In order to overcome this problem, moisture balances

combined with an infrared lamp for heating the sample are used for rapid drying

and simultaneous weighing. The weight loss can be followed continuously and the

end point of very slow loss or constant weight easily determined. The scale on

these devices can be calibrated to read solids content directly provided one starts

with a fixed initial weight of sample. With such a balance, a solids determination

can be made in a fraction of an hour, as opposed to the many hours required with the

TAPPI procedure. It is thus suitable for use in the mill as a routine method to

enable operators to run a recovery furnace efficiently and safely. However, the

moisture balance should be considered a derived method requiring calibration and

not as a primary standard method. The reason is that the high liquor surface

temperatures which are associated with the higher drying rate may lead to thermal

decomposition, formation of an impervious skin, or oxidation of the liquor. Parker

et al. state that "use of the so-called moisture balances for rapid drying and

simultaneous weighing have been evaluated with ambiguous results." It would appear a

that the moisture balance is suitable for monitoring the more indirect methods

(refractometers, velocimeters, etc.) in order to detect drifts and as a check on

sudden changes. However, actual changes in the calibration should be based on

more accurate procedures.

DISTILLATION METHOD

Distillation methods for moisture analysis in pulp, paper, board, and

other materials are well known, and have been applied to black liquor analysis.

Basically, the approach consists of distilling off the water from the liquor in

the presence of an organic solvent and collecting it in a trap. From the amount

..
 Group Project 3025 Page 89-
Report One..

of liquor originally present and the amount of water collected, the solids content

can be calculated.

A schematic diagram of the apparatus for determining solids content by the

distillation method is shown in Fig. 32. A weighed amount of black liquor is

introduced into the flask along with the solvent. The distilling trap is initially

filled with solvent. Upon heating, both solvent and water are vaporized in the

flask and recondensed up in the condenser. As the condensate drops down, the

denser water collects in the trap and the solvent is refluxed to the flask. The

procedure is continued until all of the water is distilled over. The connection

between the flask and the trap should be insulated to prevent condensation of water

at that point so as to facilitate capturing all of the water in the collecting trap.

The temperature of the contents in the flask will remain close to the boiling point

of water until essentially all of the moisture has been driven from the liquor.

Then it will rise to the boiling point of the solvent. The end point of the

distillation is signalled by the rise in temperature and when the volume of the

water in the trap no longer increases with time. The volume of water in the trap

is then read and the percent solids calculated. The precision of the test is such

that duplicate determinations should agree within 0.5% solids.

Parker et al. (2) make the following comments with regard to distillation

methods. "Allowable sample sizes are normally small and vary directly with the

sample solids content. Thus weak liquors with high moisture contents require

smaller sample sizes to keep the amount of-water distilled within the capacity of

the collecting trap. Consequently, test accuracy varies widely as.solids content

changes. Calibration and cleanliness of the apparatus are of utmost importance in

the distillation methods, rendering agreement between different laboratories some-

what difficult."

I
Page 90 Group Project 3025
Report One

/- Open

Odor trap --

Condenser ---

"-- Cooling water

Distilling trap

- Solvent layer
a

Insulation Water layer

(distilled from the black
liquor sample)

Round-bottom flask -\

Solvent
Weighed black liquw
and
boiling chips
Heater

Figure 32. Distillation Apparatus for Solids Determination

on Black Liquor
Group Project 3025 Page 91
Report.One

Proponents of the distillation method believe that it offers several

advantages over residual solids methods. It is a faster method than oven-drying

methods, since a distillation determination can be completed in about an hour.

It is also more precise than a moisture balance approach. Among the advantages

claimed for the method are the following:

1. With concentrated liquors, the relatively large-size sample of liquor used

increases accuracy by permitting a more representative sample.

2. It is not necessary to weigh out liquor samples to a specific weight as with

the moisture balance method.

3. Distillation.is a direct means for determining water content. Solids can then

be defined in an unambiguous manner as the total nonaqueous part of the liquor.

4. The black liquor is heated at a controlled temperature. The temperature

remains close to the boiling point of water until evaporation is essentially

complete. The maximum temperature is limited to the boiling point of the

solvent.

5. Absorption of oxygen from the air is eliminated because the sample is covered

by solvent. Since solids are determined from the amount of water present,

oxidation of the liquor after the initial weighing would not cause error unless

water were produced.

6. Results are independent of the time of distillation once the end point is

reached.

7. The time for carrying out an analysis, about one hour, compares favorably with

the long times needed for oven methods.



Page 92 Group Project 3025

Report One

8. The distillation method is reproducible. Measured amounts of water added to

black liquor can be recovered quantitatively.

The distillation method would seem to be sound in principle. It may be

somewhat less precise than the TAPPI method because weighing the liquor and deter-

mining the amount of water are done only to within 0.1 gram. However, it should

give correct results if carried out properly.

COMPARATIVE TESTS

In this work, both the ovendry and the distillation methods were employed

for the direct determination of solids content. Although all of the data presented

on the refractometer and velocimeter are in terms of ovendry solids values, the

experience with the distillation method was quite good. There was nothing in the

course of making the tests that directly indicated one method was inherently more

accurate than the other.

It is interesting to compare values obtained by the two methods. Such

data are shown in Table VII. One set of data was taken with Liquor C for a delib-

erate comparison of the two methods. The other set was taken with Liquor A during

flow loop tests and includes a best estimate of the solids content taking into

account the amounts of dilution water added. It can be seen that with Liquor C

the distillation method tends to give consistently lower values of percent solids

than does the ovendry method. The disagreement between the two methods is less in

the data taken with Liquor A, although the distillation method does seem to give

slightly lower values at high solids contents. It should be emphasized that these

tests were not intended to be a conclusive comparison between the ovendry and

distillation methods. No great pains to achieve maximum accuracy were taken.

 eai; qqTh pae;~oosss Sa.xo.zTi eLlq JO 4sow -paqoa-Too mem, aae;m jo ;tnouis aasa
le suvTfl WGUOD sPTTOs JO ~flTBA J~hOT IG 'IsnJ1 -P~40TTO0 J~4sh JO 4ufouip aq
e
uIOaJ PeuTB~qO sl anlTSA T4 4Uqq. puUIl UT euaxoq eq 4sntII 4T 'Pa'-r~l' uOT4PTTT~sTP eWIq
tflTm peuTB;qo senT'eA .IetOT k[4uaasdd-e et{; JO SUTUL-9m DtJ4 SUT;Xada;uT uj
0-91 g,T
6,oc 6,o~
01 Sc g
F1*
,o~ Ero6
F"~
L, 911
0,9iT
Is
9, I o* 61 L. Ei11
T9 N 0, 81
C
6 9 SPTIOS %
'V .zonbTrq tuo uq-,e
T* CT 61-
g,~E C;. 9F
T.~C 6,9~
L 011 g, F11
TI.
z~ 6g QT9
~3* 89 6 69 sPTTOS %
'D .ionbrl.uo 'e4,~
e;~muT;4SE[ potql;ew
POT14aTST X~,zpt~AQ
egia
S COHLIHK KOiiwrliLSIaI CKV DIC~aA0 W0O1J SafLrIVA sciaros do mosT-~dWIOD
IIA arMVlI
1
w.Tsa eq o;~aveddi s~bp SbflTeA .xehoT 4qft~eUIos eATS 0 poq4etu UaT!;B
-TT4sTP eq IJ.Rtete;e;'iee-o-j e [~dnp UT Unax ;ou eaxen s;se; elq4 jo RuwW
f
euo ;.iodeu
Z6 evcI gF~~4oeCoa.dI dnoaxo
Page 94 Group Project 3025
Report One

distillation method (stopping short of the end point, incomplete ' condensing, or

trapping water droplets on the side arm) would tend to reduce the amount of water

collected and hence cause an overestimation of solids content. The two errors

which could lead to low solids values are either the solubility of some solvent in

the water layer or a consistent underestimation of the initial weight of liquor.

It may well be that consistently lower values by the distillation method are actually

indicative of a systematic error in the oven-drying procedures. Such an error

could be due to oxidation of the liquor during drying or to an inability to drive

off the final amounts of moisture. These errors would tend to be greatest with

heavy black liquors. Thus, it is possible that the distillation method is more

accurate for high-solids liquors, while the ovendry method is more accurate for

weaker liquors.
Group Project 3025 Page 95
Report One

CONCLUSIONS

The results of this test program clearly show that both the refractometer

and the sonic velocimeter are capable of effectively monitoring black liquor solids

content on a continuous basis. In both cases, the response is a monotonically

increasing function of solids content which can be linearized over reasonable

concentration ranges. The effect of temperature on the response is sufficiently

straightforward so that temperature compensation can be provided. Thus, with

proper calibration, either instrument can be used to measure black liquor solids.

No major operational limitations were found regarding the refractometer.

The response was not affected by changes in liquor flow or system pressure.

Aeration of the liquor did not affect the response. No problems with deposits on

the window were encountered; however, the loop was drained and washed with water

frequently so the possibility of build-ups over a longer time period remains.

The only significant limitation to the use of the velocimeter was the

effect of air bubbles in the liquor. It was practically impossible to obtain

meaningful data with the velocimeter until the liquors (which were obtained already

aerated) were deaerated. This would not hinder the use of the velocimeter within

or after the multiple-effect evaporators, or a high-solids indirect concentrator,

since the evaporation process would keep the liquor free of air. However, it is

expected that a direct contact evaporator would aerate the liquor. Thus, the

velocimeter would not be very applicable beyond the contact evaporator unless some

means for on-line deaeration of the heavy liquor were employed.

Both the refractometer and the velocimeter are sensitive to changes in

chemical composition of the liquor. . The relationships of sonic velocity or

refractive index to solids content and temperature are functions of composition.

Page 96 Group Project 3025
Report One

Thus changes in chemical composition may be' interpreted as changes in solids

content of the liquor. This can easily cause an error of several percent in the

solids content reading. The velocimeter tends to be quite sensitive to the

inorganic/organic ratio of the liquor. The sonic velocity increases (which would

be interpreted as an increase in solids content) as the inorganic content increases.

The relationship appears straightforward and may eventually be susceptible to

quantitative treatment of an empirical nature. The effect of composition changes

on refractometer response was more difficult to interpret. It does appear that

the refractive index tends to decrease (which would be interpreted as a decrease in

solids content) as the inorganic content increases. This is the opposite direction

from the velocimeter.

Because of the composition dependence of the response of both instruments,

it is necessary to recalibrate them periodically to account for changes in liquor

composition. The frequency of calibration would be dependent on the experience of

a given mill; however, it would appear that the instrument should be checked at

least once each shift. The procedures to be followed in changing the calibration

"on-line" are discussed in the body of the report and need not be repeated here.

If experience indicates that frequent adjustments of the calibration are necessary

to maintain agreement between indicated solids and laboratory test values, the

instruments should be taken off-line occasionally for a full recalibration.

It would appear that these two instruments could be effectively operated

in tandem so as to make use of the fact that they tend to respond in opposite direc-

tions to changes in liquor composition. This would certainly permit a distinction

to be made between'changes in solids content and changes in composition. If both

instruments responded together, it would'indicate a real change in solids content.

On the other hand, if they moved in opposite directions, it would indicate a

composition change. It would appear possible that after a certain period of data
Group Project 3025 Page 97
Report One

logging, use of a refractometer and velocimeter in tandem would permit simultaneous

monitoring of solids content and inorganic/organic ratio based on an empirical

algorithm;

ACKNOWLEDGMENTS

The authors would like to acknowledge the following people who contributed

significantly to this work. Mr. 0. C. Kuehl and Mr. A. O. Johnson erected the flow

loop and helped keep it running. Messrs. H. J. Grady, D. G. Sachs, and C. Piette

assisted in taking data and running the liquor analyses. Mrs. E. A. Cary did her

usual fine job in typing the manuscript. We would also like to thank the paper and

instrument companies who supplied us with the instruments and the liquors which were

used.
Page 98 Group Project 3025.
Report :One

*. -r , -,. LITERATURE CITED,

1. Black Liquor Recovery Boiler Advisory Committee (BLRBAC) Instrumentation

Subcommittee. Report on use and application of black liquor solids monitoring
devices. BLRBAC, Oct. 1, 1969.

2. Parker, J. L., Hensel, R. P., and Wagoner, C. L., Tappi 53, no. 5:874-7 (May,
1970).
3. Zacharias, E. M., Instruments and Control Systems 43, no. 9:112-113 (Sept.,
1970).

4. Phillips, ;J. H., and Rubright, M. M., Tappi 3 6, no. 9:392 (Sept., 1953).

THE INSTITUTE OF PAPER CHEMISTRY

T. M. Grace, Research Associate

Division of Materials Engineering
and Processes

tLCNv.L z4 ^
Robert N. Larsen, Research Fellow
Division of Materials Engineering
and Processes

A.
 '.1
L Vt'Lqot'90L0 090
- J
-