Black liquor soap separation

and acidulation: changing some

old habits
M e m t t Ketcham

Time- and labor-saving techniques as well as a new process design reduced

excessive maintenance requirements and gave the operator new tools
for improving crude tall oil quality.

Longview Fibre Co. operates its 3000-ton/day integrated in tall oil handling. These compounds are found in
kraft pulp and paper mill along the Columbia River in different parts of the living tree. The fatty acids are food
southwestern Washington. The mill produces container reserves, and the rosin acids provide a form of protection
and specialty grades from Douglas-fir and pine chips or wood preservative.'
purchased from outside suppliers. A small hardwood There are many tall oil end products. The paper
fraction is pulped using the neutral sulfite semichemical industry is very familiar with rosin, which can be
process for corrugated medium. processed into rosin size, a popular component for the
The company has operated a crude tall oil (CTO) sizing of paper. Other end products include detergents,
acidulation plant, which produces 60 tons/day, since vinyl floor coverings, protective coatings, ore flotation,
1962. The plant was designed around the batch process and oil drilling muds just to name a few.'
popular at the time. Acid supply is principally waste Kraft pulping of tall oil-containing softwood is the most
sulfuric acid from the bleach plant's R2 chlorine dioxide common source of tall oil. The cooking operation includes
generator. Supplemental waste and fresh acid is the addition of white and black liquors, the white liquor
available locally. Lignosulfonate is employed as a containing the bulk of the cooking alkali charge. Weak
separation aid. An operator is employed for the day shift, black liquor is added as an extender to provide sufficient
five days per week. He may be required to work liquor volume to cover the chips. It turns out that black
additional hours as soap supply dictates. liquor also provides a very important ingredient to the
Tall oil yields follow classic variables: season, source cook: saponifiable tall oil. This serves as a surfactant,
of chip (dead or dying, bug kill, etc.), and wood pile age promoting impregnation of the liquor into the chip.
and pile dimensions. The liquor is removed from the brownstock fibers in
The driving force behind the renovation of the CTO multistage washers. The method of washing is not as
plant a t Longview Fibre Co. was the increased effort important as the efficiency to the enhancement of tall
required to maintain the plant, the excessive operator oil soap recovery. Poor washing efficiency not only
hours required to process the soap, and the high moisture reduces tall oil tonnage but also reduces paper quality,
content of the final CTO. Also, the plant capacity was promotes foaming, increases refining requirements, and
requiring the operators t o frequently bypass the may contribute to a wet-end pitch problem. Hot wash-
skimmers, leaving the soap in the liquor headed to the water temperature is especially important in removing
recovery furnaces. The recovery furnaces do not function the low-solubility tall oil soaps. The black liquor removed
well with soap in the nozzle liquor, so it became a threat a t the front end of the washing process contains the tall
to production. oil soaps we plan to recover.
Tall oil-containing black liquor soaps are difficult to
Tall oil chemistry separate at low solids concentrations. Some may rise with
foams generated during early stages of pulp washing or
Tall oil is a mixture of organic acids. Rosin acids, fatty perhaps weak liquor oxidation. However, most soap
acids, and unsaponifiables are the compounds of interest separation systems operate best with a solid concentra-
tion of 25-30%. Figure 1 illustrates the solubility of tall
oil in black liquor.
'Drew, J. and Propst, M., Tall Oil, Pulp Chemicals Association, New York, Higher liquor solids are obtained by multi-effect
1981.
evaporators. Early in the development of liquor evapo-
Ketcham is the engineering section leader at Longview ration, it was found that black liquor soaps precipitate
Fibre Co., Box 639, Longview, Wash. 98632. on the tube walls, causing premature fouling. Early
February 1990 TappiJ o d 107
I 2. EvaporatorMmmer flow diagram. Only the first four of the seven effects have been shown for clarity.

B"
Skimmer
flash tank

To soap
accumulator skimmer
tank

~~

skimmers were employed to allow removal of the soaps

1. Solubility of tall oil inblack liquor
for this reason.

Skimmers
Soap skimmers are strategically placed in the middle of
multi-effect evaporators, where the liquor solids are a t
the 25% optimum. We place the skimmers between the
second and third effect of seven-effect evaporator sets.
At this point in the process, we insert a skimmer flash
tank to reduce the liquor temperature below a 1atm flash
point. Figure 2 illustrates this arrangement. Failure to
take this step may result in flashing in the skimmer. The
0
skimmer flash tank is vented to an evaporator dome
0 10 20 30 40 50 60 operating under a slight vacuum. Care must be taken
when installing a skimmer flash tank on an existing
TOTAL SOLIDS, %
evaporator set to avoid overloading an effect with flash
steam, which would upset the heat balance through the
set.
Skimmer design has changed considerably through the
years. John Drew and Marshall Propst in their book on
tall oil processes1 listed five design characteristics to
obtain maximum separation efficiency:
1. Minimize short-circuiting through proper baffling.
108 February 1990 TappiJournal
3. Crude tall oil plant

w
Soap accumulator

CTO ifinal
storage
I
Tostron black
liquor?ank 0 1 NeutalIEed spent brine
II X$ I1
1weak liquor
storage ap-liquor
interface

Cb

$1 liquor
Weakblack
To weak
system
black l i

Caustic storage I Lign %vat& tank

Acid
tank CIO, spent
Metering tank acid
Caustic
LignosuHonate

?.Minimize the distance the soap must rise to reach the skimmer supply pump. This may provide a cheap and
surface for collection. quick enhancement of the soap rise rate by lowering the
3. Minimize turbulence. This includes convective density of the soap. Begin with very small additions and
currents set up by poorly insulated tank walls. work up to 0.75 m3/h of air per 1000 L/min liquor input.
4. Allow no area in the tank to have a downward velocity The results are usually quick and sometimes dramatic
in excess of the soap particle rise rate. if too much air is inserted. It is possible to destroy the
soap pad in short order if the air flow is too high.
5. Provide sufficient retention time for the soap to rise
to the surface.
Soap accumulators
A simple experiment can discover how well your
skimmer is working. With the evaporator set operating In mills with multiple evaporator sets with skimmer tanks,
steadily (no surging), chemically examine the skimmer foam tanks, and skimmed heavy liquor tanks, the soaps
input and skimmer liquor discharge for CTO concentra- are generally collected in an accumulator. These accum-
tion. A properly designed and operated skimmer should ulators are usually decanted periodically, with the liquor
recover 90% or more of the tall oil-containing soaps. that settled out being recycled to the weak liquor tank.
Poor skimmer operation may be improved in most
cases with a s m a l investment. Insulation is important Soap washing
to minimize thermal currents caused by cold tank walls.
Adding baffles or velocity breakers to the liquor inlet The soap from the accumulator should be washed to
nozzle can minimize stirring. The liquor discharge nozzle separate the free liquor prior to acidulation. Failure to do
should be large and located near the tank bottom to a good job of washing and liquor removal increases the
reduce high suction velocities and vortexing. acid charge required for proper acidulation and reduces
Perhaps the most effective approach is the addition of the overall CTO quality.
a small amount of compressed air into the suction of the For several years, we washed the soap using medium-
February 1990 Tappi Journal 109
8 shear open mixers with neutralized brine from the
acidulation process as the medium. This was messy,
was lined with a proprietary membrane. An acid resistant
brick lining then was applied over the membrane. All pipe
3 required excessive energy, resulted in entrained air in the
soap, and seemed to require excessive operator attention.
penetrations and the agitator were constructed of
Carpenter 20 alloy. The roof was again mild steel with
This washing system needed extensive repairs, and we a lead interior lining.
decided to take the advice of some recently published This design worked well for 10 years but eventually the
literature on low-shear mixing. An in-line mixer of our membrane was breached, and the mild steel shell started
design was fabricated and installed. Pumping from the to dissolve. Patching and injection of various products
mill's two accumulators was changed from a continuous between the brick and the steel have prolonged the life
operation to a batch operation. During the batch pumping of the reactor. The reactor has been in service now for 27
operation, the neutralized brine is injected with the soap years and has been demanding more extensive repairs.
in the eye of the in-line mixer. The mixture is then Management authorized replacement of the reactor and
transported to the washed soap chest for separation. More the associated modernization of the facility.
recent trials have indicated that water is superior to the The new reactor was designed around the following
neutralized brine for soap washing and viscosity reduction objectives:
in our particular chemical balance and is now the washing
medium of choice. 1. Increase production by 30%.
2. Improve operator time efficiency.
Soap clarification 3. Improve the quality of the cook by more accurately
metering in the chemicals.
The conventionalsoap separation tank design employsone
4. Reduce maintenance to steam spargers, swing pipe, and
or two tanks with the soap piped in series. The last tank
brick lining.
in the train has a swing pipe or more likely a floating
decant pipe. The design of the soap storage tank at To meet the increased production objective, the'reactor
Longview Fibre used neither of these concepts. We used needed to be 30% larger. This was accomplished by
a series of suction valves at 1.75-m intervals up the side enlarging the diameter 0.7 m. The agitator size was
of the tank. All these designs require the operator's careful increased accordingly. The tank bottom was sloped 1-in-
attention to avoid pumping liquor into the reactor. Another 12 to aid in calcium deposit removal and drainage.
unfortunate feature of this soap storage tank design was Batch-type flow controls were added to the caustic, spent
the requirement for soap pumping to the reactor. Soap acid from the R2 generator, purchased acid, and ligno-
pumps are notorious maintenance hogs, limited to flows sulfonate (separation aid) addition systems to improve the
generally lower than 2000 L/min, and expensive. At the quality and repeatability of the cook.
time of modernization, it was taking fuur hours to transfer The steam spargers were redesigned for top entry. The
soap from storage to the 250,000-L reactor. The transfer steam flows down a sparger (4X) tucked behind the
time and risk of liquor contamination justified the agitation baffles. Holes in the side of the sparger distribute
construction of a redesigned soap storage tank. the steam more uniformly through the fluid. The vertical
The new design reused all of the old tank. The existing design eliminates the need for individual valves and allows
foundation was found adequate to extend the side wall of for sparger removal without entering the reactor.
the existing tank. The tank wall height was raised from The swing pipe used for decanting was eliminated. A
9.75 m to 16.8 m. A 40-cm gravityoutlet nozzle was installed decant box was installed a few feet above the normal CTO-
at slightly above the same elevation as the reactor roof. The to-lignin residue layer. Decanting of the CTO is accom-
modified tank now is operated with the lower portion plished by draining the CTO out the decant box nozzle
containing liquor which has settled out of the soap. The until equilibrium is obtained. To remove the remaining
soap floats on top of this liquor. The weight of this potential oil below the lip of the decant box, the operator pumps
10-m column of soap tends to increase the density of the in spent brine well below the lignin layer, displacing the
soap in the lower zones. The gravity outlet provides for 15- CTO out the decant box nozzle. Switching valves are then
min filling of the reactor with the highest-density soap toggled to allow the lignin layer to be similarly decanted
without a pump. The interface elevation between the liquor to a holding tank for disposal in the strong black liquor
and the soap can be detected quite accurately by observing streams and from there to the recovery furnaces. After
the abrupt temperature change of the tank shell at the the CTO and lignin residue is removed, the calcium
liquor-soap interfaces. For safe decanting, the operator deposits on the floor can be purged, again to the strong
adjusts the interface elevation to at least 1 m below the liquor system. A small portion of the remaining spent brine
outlet nozzle by pumping away lower level liquor. is returned to a holding tank for future use as displacement
liquid. The balance of the spent brine is neutralized in the
Audulator design acidulator with caustic. This product is used for washing
soap from the skimmers or pumped to the weak black
The old acidulator replaced during the project was liquor system for evaporation.
constructed of mild steel plate. The height-to-diameter The new reactor tank is constructed entirely of Alloy
ratio was 1.25. The tank bottom was essentially flat. Steam 20. The rationale behind the move away from brick and
injection nozzles (4X) were placed in the side wall near toward an alloy product was strictly economic. The Alloy
the bottom. A swing pipe also was installed. 20 reactor was 60% the cost of a brick lined reactor. The
Corrosion protection was provided for this highly brick also consumes needed volume and, as we learned,
aggressive environment. The mild steel tank interior first can mask an active corrosion problem behind the brick.
110 February 1990 TapplJournal
 Alloy 20 is a high-nickel alloy particularly good in hot
sulfuric acid environments. The agitator shaft and paddles
of Alloy 20 had performed very well in the 27-year-old
acidulator being replaced. The mills that use continuous
acidulation processes have used Alloy 20 successfully for
many years. The choice of Alloy 20 has been supported
in subsequent inspections after 24 months of operation. No
significant waste or cracks have appeared. It is likely that
constructing the lower portion of the reactor with Alloy
20 and the upper portion of a less costly austenitic stainless
steel may have been equally successful, but this option
wasn’t considered.

Recooking the lignin residue

The lignin residue that forms between the spent brine and
the floating CTO contains large quantities (up to 70%) of
trapped CTO. Much of this can be liberated by simply
acidulating again. The lignin and fiber will tend to
concentrate each time and will require periodic purging
to the black liquor cycle. Recooking the lignin layer
recovers sufficient additional CTO to make it worthwhile.

mocirvine
The “wet))CTO decanted from the top of the acidulator
contains 6-10% water. Contracts of sale to processors
generally require water to be less than 1.5%by volume.
To reduce the water volume left from the decanting
process, we elected to realign the two existing wet CTO
tanks from series to parallel operation. An existing “dry)’
CTO storage tank was converted to a “wet’)tank and placed
in series after the twin parallel tanks. A new shipping tank
was constructed adjacent to a major mill roadway, which
allows easy shipping by truck of the finished product.
The twin parallel/single series drying tank concept has
solved what used to be a nagging quality issue between
the mill and the CTO buyer.

Conclusions
The tall oil-containing soap separation systems and the
soap acidulation and drying systems had been requiring
excessive maintenance dollars, operator overtime, and had
been producing marginally acceptable CTO in quantities
below what improved processing methods could provide.
A new process design (Fig. 3) reduced excessive mainte-
nance requirements and gave the operator new tools for
improving CTO quality. Production rates were improved
by the required 30% while reducing operator overtime.
Skimmer bypassing has been reduced but not entirely
eliminated.
There are many ways to improve even tried and true
processes. I hope that we have shown even small expen-
ditures can reap large gains in overall pr0ductivity.U
The author gratefully acknowledges the process development assistance
of Gary Dietel of the Longview Fibre engineering staff and Dennis
McCarthy, CTO plant operator.
Received for review June 15,1989.
Accepted Aug. 25, 1989.
Presented at the TAPPI 1989 Engineering Conferenceand the TAPPI 1989
Pulping Conference.

February 1990 TappiJournal 111