close

Вход

Забыли?

вход по аккаунту

?

bk-2017-1251.ch012

код для вставкиСкачать
Nanocelluloses: Their Preparation, Properties, and Applications
Downloaded from pubs.acs.org by CITY UNIV OF HONG KONG on 10/25/17. For personal use only.
Chapter 12
Experiences with Scaling-Up Production of
TEMPO-Grade Cellulose Nanofibrils
Richard S. Reiner* and Alan W. Rudie
USDA Forest Products Laboratory,
Madison, Wisconsin 53726, United States
*E-mail: rreiner@fs.fed.us.
The USDA Forest Products Laboratory (FPL) has worked
toward production of cellulose nanofibrils (CNF) to a level
suitable for supplying for many preliminary research and
development projects. In particular, this chapter discusses the
process steps involved in scaling the production of CNF from
grams in the laboratory to pilot batch sizes up to 4 kg, namely
the reaction, washing, dispersion, separation and freeze drying.
The carboxylation of the pulp nanofibril surface was carried
out using the catalyst TEMPO with two reaction protocols
utilizing hypochlorite at pH 10 and sodium chlorite at pH
7. Various methods for mechanical dispersion, of oxidized
pulp fibers into individual nanofibrils are also compared.
Currently, aqueous suspension and freeze-dried TEMPO-grade
cellulose nanofibrils produced using hypochlorite are available
through the University of Maine’s Process Development Center
(along with other cellulose nanomaterials). However, end-use
applications will ultimately dictate the process optimization of
chemistry and dispersion that are engineered into producing
CNF with particular properties.
Not subject to U.S. Copyright. Published 2017 by American Chemical Society
Nanocelluloses: Their Preparation, Properties, and Applications
Downloaded from pubs.acs.org by CITY UNIV OF HONG KONG on 10/25/17. For personal use only.
Introduction
The act of modifying the bonding surface area of cellulosic fibers to change
their papermaking properties has existed since the ancient Egyptians began
hammering overlaid papyrus fibers into sheets or the 2nd century Chinese began
pounding rags, wood and grasses into individual fibers that were collected onto
mats and dried into sheets. In an over simplification, the cellulose fibers used
to make paper are individual plant cells. In the growing plant, these cells are
bound together with other natural polymers in a composite structure to form the
structure of the plant: trunk, stems, branches, etc. The cellulosic structure of the
plant cell wall might be compared that of rope, namely small individual threads
of a material are twisted and braided together to form larger and larger fiber-like
structures. Plant cell walls are composed of layers of microfibers, which in
turn are composed of bundles of nanofibers, which are composed of bundles of
cellulose chains that were produced by the plant during its growth. The process of
pulping is, through a combination of mechanical and chemical means, to separate
these individual fibers so as to be able recast them into thin sheets to make paper.
The process of beating and refining pulp fibers prior to sheet formation extends a
certain amount of microfibrils from the fiber surface and increases the cellulose
surface area available for bonding. Further refining and beating, begins to not
only break down the fiber structure, but also begins to extend nanofibrils from the
surface of the microfibrils. This acts to further alter the sheet strength, uniformity,
density, opacity and porosity. Taken to an extreme, such refining was used to
produce glassine, a strong, thin, glazed, semitransparent form of paper.
However, recently, more aggressive mechanical methods have been used to
completely disrupt the plant fiber cell walls and produce a very high surface area
fibrillar product. The first report is not recent, 1980, using multiple passes through
a homogenizer to get microfibrillated cellulose (MFC) as a gel-like cellulose
suspension (1). Addition of enzymes (2) and acids (3) have helped to reduce
processing energy and further extend the structural disruption toward nanofibrils.
The most complete nanofibrillar dispersions are achieved by chemical treatments
which act to add ionic functional groups to the surfaces of the nanofibril:
examples, TEMPO oxidation which oxidizes the C-6 carbon of cellulose to a
carboxylic acid (4–6), chloroacetic acid, which substitutes acetate groups for any
of the three hydroxyl groups on cellulose (7) or oxidation with periodate which
oxidizes C-2 and C-3 of cellulose to carboxylic acids (8). When suspended in
water, these ionic groups act to repel nanofibrils from one another and with the
aid of mechanical action can produce a colloidal suspension of highly dispersed
cellulose nanofibrils. Recently, cellulose filaments (CF), a long, thin ribbon
of cellulose that is mechanically peeled from from pulp fibers has also started
being produced (9). Table I lists cellulose nanofibril production or announced
production facilities around the world. By definition, nanocellulose must have at
least one dimension less than 100 nm in length. However, some of the materials
listed may not fit this definition but are included here as they are being developed
for many of the same types applications as nano-scale CNFs.
228
Table I. Cellulose nanofibril production facilities worldwide
Nanocelluloses: Their Preparation, Properties, and Applications
Downloaded from pubs.acs.org by CITY UNIV OF HONG KONG on 10/25/17. For personal use only.
Company
Country
Product notesa
Productivity
kg/day
Ref.
Forest Products
Laboratory
USA
4
TEMPO CNF
Paperlogic
USA
2000
CNF
(10,
11)
Univ. of Maine
USA
1000
CNF
(10)
American
Process
USA
500
Hydrophobic CNF
(10)
Kruger
Canada
5000
CF
(1, 12)
Performance
Biofilaments
Canada
Pre-Commercial
CF
(1)
Oji Paper
Japan
100
Phosphate CNF
(13)
Nippon Paper
Japan
1400
TEMPO CNFs
(14)
Daio Paper
Japan
275
CNF
(15)
Borregarrd
Norway
1000
MFC
(10,
16)
Betulium
Finland
Pilot
CNF
(17)
InTechFibers
France
100
CNF
(10)
Innventia
Sweden
100
CNF
(18)
Stora Enso
Finland
Commercial
MFC
(19)
UPM
Finland
Pre-Commercial
CNF
(10)
Norske Skog
Finland
Pilot
CNF
(10,
20)
VTT/Aaalto
Finland
Pilot
CNF
(10,
21)
Luleå Univ. of
Technology
Sweden
Lab
CNF
(10)
Imerys
UK
1500
MFC with minerals
(10)
Cellucomp
Scotland
1000
CNF with resins
(22)
SAPPI
Netherlands
20
CNF
(23)
a
Cellulose nanofibrils (CNF), microfibrillated cellulose (MFC), and cellulose filaments
(CF).
The pioneering work for the production of chemically-modified cellulose
nanofibrils uses a catalyst, (2,2,6,6-Tetramethylpiperidin-1-yl)oxyl (TEMPO)
radical, to oxidize accessible primary alcohols of the cellulose chains to carboxy
groups, which when combined with mechanical or ultrasonic energy, produced
an extremely dispersed grade of cellulose nanofibrils (4–6). A pilot plant at
229
the United States Department of Agriculture-Forest Service Forest-Products
Laboratory (USDA-FS-FPL) in Madison, Wisconsin has been working to scale
the production of TEMPO-grade cellulose nanofibrils (CNFs) in order to make
the material available for Research and Development purposes. This chapter will
summarize the scale-up efforts from gram-scale to the kilogram-scale production
in the pilot plant.
Nanocelluloses: Their Preparation, Properties, and Applications
Downloaded from pubs.acs.org by CITY UNIV OF HONG KONG on 10/25/17. For personal use only.
TEMPO Reaction with Cellulose
The TEMPO system oxidizes the C6 primary alcohol of cellulose chains.
The reaction appears to be physically restricted to nano-scale fibril surfaces
by the composition and organization of the cell wall leaving an unoxidized
core of the cellulose chains that make up the nanofibril. The limitation of the
TEMPO reaction to only the primary alcohols on the nanofibril surface keeps
the fundamental structural unit of the nanofibril intact, with the modified surface
chemistry.
There are two sets of reaction conditions, both using catalytic levels of
TEMPO to functionalize cellulose with carboxylate groups. The first uses
hypochlorite as the terminal oxidant along with sodium bromide. This reaction
is maintained at pH 10 and is carried out over the course of several hours at
room temperature (5) A second reaction system uses sodium chlorite as the
terminal oxidant. This reaction is maintained at pH 7 and is carried out over the
course of several days at 70°C (6). Generally speaking, the TEMPO/hypochlorite
system at pH 10 can achieve higher levels of carboxylate on the nanofibril
surface, up to 1.5mmol -COONa/g dry pulp, but also results in lower degree
of polymerization (DP) of the cellulose. The TEMPO/sodium chlorite system
at pH 7 will result in lower levels of carboxylate on the nanofibril surface,
up to 0.8 mmol -COONa/g dry pulp, while the cellulose retains a higher DP.
Furthermore, the higher carboxylate of the TEMPO/hypochlorite system at
pH 10 is likely to produce nanofibrillated suspensions with a higher level of
light transparency and films cast will also likely have a higher level of light
transparency than those produced by the TEMPO/sodium chlorite system at pH
7. The trade-off is with the DP of the cellulose and is effect on the strength
of cast films. The TEMPO/sodium chlorite system at pH 7 provides stronger
films than those produced from the TEMPO/hypochlorite system at pH 10 (6)
Another consequence of the oxidation level and DP is that aqueous suspensions
of the nanofibrils produced from TEMPO/sodium chlorite at pH 7 has a higher
viscosity than an equal concentration of cellulose nanofibrils produced from the
TEMPO/hypochlorite system.
There are two distinct stages to the mechanism of the TEMPO reaction.
The TEMPO catalyst, in its cationic oxidized state, first oxidizes the primary
alcohol on the cellulose surface to the aldehyde. Then the terminal oxidant, as
well as its various oxidation states, oxidizes the TEMPO catalyst back to the
active state, and oxidizes the aldehyde to a carboxylic acid (carboxylate). The
reaction produces one mole of acid (the carboxylic acid) on completion of the
oxidation cycle, resulting in a drop in pH. Maintaining the reaction rate requires
230
Nanocelluloses: Their Preparation, Properties, and Applications
Downloaded from pubs.acs.org by CITY UNIV OF HONG KONG on 10/25/17. For personal use only.
controlling the pH within a fairly tight window, pH 9-11. This is accomplished
with addition of alkali during the reaction, or use of a buffer. The reaction using
chlorite as primary oxidant has a particular problem in that TEMPO is most
active in the pH 9-11 range, but chlorite is inactive at that high of a pH. Chlorite
is generally unreactive, but under acid conditions reacts with hypochlorous acid
to make chlorine dioxide which is quite reactive. Chlorine dioxide returns to
chlorite when it oxidizes TEMPO. Cycling though the process multiple times is
required to utilize the majority of the 4-electron oxidizing capability of chlorite.
This cycle is more rapid at low pH. The result is a compromise with operation
at a pH around 7 and a much slower reaction. If the pH drops too far below
7, the TEMPO oxidation stops. If the pH rises too much above 7, the reaction
producing chlorine dioxide stops (6). Furthermore, the TEMPO/hypochlorite
system at pH 10 is facilitated with the presence of sodium bromide (5), while
the TEMPO/sodium chlorite system is dramatically improved if the reaction is
kick-started with a small dose of sodium hypochlorite (6).
Since the cellulose modification proceeds through the aldehyde as the alcohol
is oxidized to carboxylic acid, residual aldehydes may be present on the pulp. This
residual aldehyde level may be affected by the ratio of pulp to sodium hypochlorite
or sodium chlorite, the reaction time, the initial pulp (i.e., species, previous dry
state, hemicelluloses, etc.). Residual aldehydes may be oxidized to carboxylic
acid using a treatment with sodium chlorite and acetic acid (4).
In practice, the USDA Forest Products Laboratory has chiefly produce
cellulose nanofibrils using the TEMPO/hypochlorite reaction, but has explored
the reaction with TEMPO/sodium chlorite. The TEMPO/hypochlorite method
offered several advantages. First and foremost, the higher light transparency of the
suspension was a desirable characteristic for many of the researchers interested in
the materials. This is a unique characteristic relative to many other nanomaterials,
such as, carbon nanotubes and nanoclays. Other factors that favored use of
the sodium hypochlorite procedure included ambient reaction temperatures,
reaction times of hours instead of days, and maximizing yield (at least with
regard to maximizing light transparency by removing less well-dispersed residue
as discussed below). There is an expectation that as CNFs are developed into
potential products, it will be possible to optimize the properties of the CNFs
for the application either by changing the primary oxidant, or controlling final
carboxylic acid levels.
Bleached kraft Eucalyptus machine dried pulp has been used as the starting
material for the TEMPO reaction. Four kilograms of drylap sheets were repulped
using a Voith 50L hydropulper (Appleton, WI) at about 8% solids. Repulping
was continued for about 30 min to assure the fibers were well dispersed and help
swell the fiber wall. The TEMPO reaction proceeds similarly with dried and
never dried pulps, but there are some differences in water retention values after
TEMPO treatment (5). The pulp was pretreated by soaking in water containing
0.2% sodium chlorite on pulp and adjusted to pH 2 with sulfuric acid. This was
stirred slowly overnight, then washed well with reverse osmosis (RO) water.
The pretreatment oxidizes any residual lignin and quinones that would otherwise
consume oxidant during the TEMPO reaction. The low pH removes trace metals
such as calcium, magnesium and iron. Once the pulp is carboxylated, multivalent
231
Nanocelluloses: Their Preparation, Properties, and Applications
Downloaded from pubs.acs.org by CITY UNIV OF HONG KONG on 10/25/17. For personal use only.
cations are tightly bound to the pulp, and can add color to the pulp and inhibit
dispersion of CNFs.
A bicarbonate/carbonate buffer is used to maintain the TEMPO/hypochlorite
reaction pH close to 10. Sodium carbonate, 1000 g is dissolved in 150 L of RO
water in a stirred, glass-lined reactor equipped with a baffle. Wet, pretreated pulp,
equivalent to 2 kg oven-dry, is added and stirred for 30 min. Then 200 g sodium
bicarbonate is added along with 225 g sodium bromide. Thirty-two grams of
TEMPO is added as a dry powder and mixture stirred for 30 to 60 min. While
TEMPO is slightly soluble in water, its dissolution is relatively slow. It was helpful
to minimize TEMPO crystal size by melting an appropriate amount, pouring it into
a ziplock bag, then cooling in ice. The bag could then be gently kneaded to produce
a relatively fine TEMPO powder for sprinkling into the reactor. The reaction is
diluted to 400 liters and sufficient sodium hypochlorite solution added to provide
5 mol NaHClO per kilogram pulp; this was typically about 7L of 12% solution.
(The concentration of stock sodium hypochlorite needs to be determined through
an appropriate analytical method.) The reaction was allowed to proceed overnight.
The carbonate/bicarbonate buffer provides an initial pH near 10.5 and the final pH
near 9.5. The contents of the reactor were collected by filtration on a 40-inch
diameter Nutsche filter, and washed in place with RO water. Typical, carboxylate
levels are 1.4 mmol -COONa/g with a TEMPO treated pulp yield between 90-95%.
Recently, FPL has assembled an auto titrating system that adds NaOH
solution as needed to control the reaction pH at 10 during the reaction. This
method provides a way for clearly determining the reaction endpoint as the pH
stabilizes and the control system no longer requires additional NaOH solution.
A small amount of sodium carbonate/bicarbonate is still used (50/10 grams
respectively in 400L) in order to improve pH control and prevent localized spikes
that could be detrimental to both the reaction and complicate the pH control.
With good control at pH 10, the reaction is complete after about 3 h. Using this
method, carboxylate levels on the pulp increased to about 1.5 mmol -COONa/g.
This increase was sufficient to further swell the treated pulp making it harder to
wash and necessitating some procedural changes discussed below. Pulp yields
have been about 88 to 92%.
Chemical Analysis of the Carboxylated Pulp
To analyze the carboxylate content of TEMPO reacted pulps, FPL has used
two methods. The first is a modification of TAPPI method T-237, “Carboxyl
Content of Pulp,” in which the reagent concentrations have been increased due
to the significantly higher carboxylate contents of these pulps. This method
only works using the reacted pulp before dispersion as nanofibrils. The pulp is
converted to the acid form and washed well with high quality deionized water. A
known test quantity is reacted with a standardized solution of 0.1N NaHCO3 and
0.25M NaCl, which bubbles off CO2 generated by the acid groups on the pulp.
The pulp is removed and an aliquot of the reagent solution is titrated for NaHCO3
consumption using 0.1N HCl to a methy red endpoint. Due to the buffering of
232
Nanocelluloses: Their Preparation, Properties, and Applications
Downloaded from pubs.acs.org by CITY UNIV OF HONG KONG on 10/25/17. For personal use only.
H2CO3/NaHCO3, the endpoint is transition is broad so the titration must be boiled
to drive off CO2 near the endpoint in order to sharpen the methyl red color change.
Another method used to determine the carboxylate content is through a
conductometric titration. The conductivity of the solution is monitored while
standardized acid or base added to a known quantity of CNF. The endpoints are
determined by changes in the slope of the titration data as the titrant is added.
This test can be performed on reacted pulp or dispersed CNF suspensions. Care
must be taken that the system has equilibrated with each addition of titrant.
Residual aldehydes may be oxidized to carboxylic acid by a follow-up
treatment with sodium chlorite and acetic acid (4). One can than measure the
difference in carboxylate levels before and after this post-treatment to determine
the residual aldehyde levels of the TEMPO reaction.
The carbohydrate analysis was performed using HPLC with an amperometric
detector (24). However, quantification of uronic acids is limited by lack of
standardization, which is reflected in the significantly lower total carbohydrates
detected.
Washing the Reagents from the Carboxylated Pulp
Due to the addition of carboxylic acid groups to the pulp, washing and
handling provides additional challenges during scale-up. At low ionic strength,
the pulp swells significantly in water slowing filtration. Further, the pulp is
also susceptible to fibrillating, even with just modest shear environments. Care
is taken to minimize handling until washing is completed. The 400 L reactor
gravity drains to the Nutsche filter eliminating the need for a pump to transfer
pulp for washing. A Nutsche filter is an agitated horizontal plate filter enclosed
in a vertical tank.
The pulp mat is sensitive to compacting which significantly slows the filtration
process. Appling pressure or pulling a vacuum creates an initial surge in filtration
flow rate, but this quickly subsided to a much slower rate. Increasing the pressure
differential across the mat again increases flow, but only for a short time until
additional fiber packing slows filtration even further. To prevent slow drainage
of the pulp, the lower chamber of the Nutsche filter is flooded to the filter prior
to transferring the TEMPO/hypochlorite reaction. The bottom valve is opened
slowly to begin draining and limit the, differential pressure to a small head.
There is a significant enough difference in ionic strength between the reaction
solution and the RO water used for washing such that the carboxylated cellulose
fibers swell during washing. This increases the sensitivity to collapse as washing
progresses. Once the bulk of the reaction has been drained above the fiber mat but
before air is drawn through the mat, RO water is sprayed over the top of the fiber
mat and allowed to percolate through. This is repeated several times with a total
of about 200L of wash water until conductivity exiting the fiber mat is sufficiently
low. After the final wash water has been drained through the mat, modest air
pressure (~1 bar) is applied to dewater the mat. In this way, two kilograms of
carboxylated pulp can be washed with a 40-inch diameter Nutsche filter in about
an hour.
233
Nanocelluloses: Their Preparation, Properties, and Applications
Downloaded from pubs.acs.org by CITY UNIV OF HONG KONG on 10/25/17. For personal use only.
This displacement washing method worked well in early reactions when
carboxylate buffer was used to control the reaction pH. However, using the
autotitrator, during the TEMPO hypochlorite reaction to hold the pH at 10 has
resulted in a small increase in the final carboxylate level of the pulp. This has
resulted in an even higher level of swelling of the pulp mat as RO water displaced
the reagents. In fact, the swelling was significant enough to cause rupturing of
the pulp mat and creating a fractal pattern of cracks that allowing wash water to
channel rather than flowing evenly through the pulp mat; to avoid this, a small
change to washing procedure was required. The completed TEMPO/hypochlorite
reaction was transferred to the Nutsche filter and drained as before, but the spent
reagents were then drained from the lower chamber of the Nutsche filter. Then,
while under vacuum, wash water is introduced from the bottom. This bottom
flow releases the pulp mat more cleanly from the filter cloth and refills the lower
chamber of the Nutsche filter once again. Once the water level was above the
Nutsche filter mixer, the pulp was mixed gently for a short period to swell the pulp
while suspended in the initial wash water. The water was drained and, the pulp
mat washed further using the displacement procedure. Dewatering was complete
with air pressure as described above. A small amount of yield is lost through the
filter cloth before the pulp fibers form a mat thick and tight enough to fully retain
cellulose particles. Even with this procedure, the TEMPO treated pulp requires
longer wash times.
This washing approach is not suitable for larger scale processing and another
approach beyond increasing filtration area per kg of pulp is needed. A likely
approach might be washing with a series of decanters; one will need to take some
care with balancing the g-force affecting the settling rate of the swollen, oxidized
pulp with the compaction of the solids and the intensity of mixing that is required
between stages so as to avoid significantly increasing the water retention value of
the pulp or yield losses due to dispersion of CNFs.
Mechanical Dispersion of the Carboxylated Pulp
Once the carboxylation of the pulp has been achieved, a certain amount of
shear energy must be added in order to disintegrate and disperse the material as
nanofibrils in water. Various methods have been explored, including stirring,
pumping, blending, ultrasound, refiners, homogenizers, etc., with success to
various degrees. At FPL, the goal is to produce a material that can be used
for research and development over the broadest array of potential applications
balanced with issues such as capacity, yield and process simplicity. This goal has,
channeled FPL CNF production toward the TEMPO/hypochlorite method and
maximizing the carboxylate content of the pulp in order to produce a dispersed
material with high levels of suspension light transparency. As applications are
developed, one can return to the variables of CNF production in order to optimize
a material with the appropriate balance of carboxylate content, cellulose DP,
suspension transparency with chemistry, energy, and other process considerations.
The initial scale-up fibrillation of CNF was performed by circulating
a tank of carboxylated pulp (1.3 mmol -COONa/g) at 0.1 % solids using a
234
Nanocelluloses: Their Preparation, Properties, and Applications
Downloaded from pubs.acs.org by CITY UNIV OF HONG KONG on 10/25/17. For personal use only.
multistage centrifugal pump (Grundfos CR2) for five days. Periodic aliquots of
the suspension were centrifuged in the laboratory at 4500G for 10 min and the
supernatant analyzed for solids content. A centrifuge was used to separate smaller
CNF particles that remain suspended at high gravity-force from larger particles.
This suspension was passed twice through a Sharples cylindrical-bowl-type
centrifuge operating at 12,500G with about a one minute residence time for each
pass. The supernatant was concentrated to nearly 1 wt% solids by circulating
through a tubular ultrafiltration membrane (4-ft B1 module with polyvinylidene
fluoride (PVDF) 200,000 molecular weight cut off (MWCO) tubes; Membrane
Solutions). This suspension was clarified by passing once through a homogenizer
(Microfluidics M-110EH-30) equipped with an 87 micron orifice. Solids analysis
showed about 57% of the oxidized pulp was converted to a clear, colloidal
suspension of CNF product in this first fibrillation cycle. The solid residue from
the centrifuge was suspended in 200L of water and circulated with a multistage
centrifugal pump for another five days. This yielded an additional 18% of the
oxidized pulp as a colloidal suspension. While, this initial procedure was able to
generate nanofibrils and confirmed that laboratory methods work at larger scale;
it was an impractical solution for pilot production. These results are summarized
in Table II.
In order to increase the rate at which pulp fibrillation could be processed, FPL
also evaluated use of a disk refiner. A Sprout-Walden stainless-steel disk refiner
was outfitted with plate pattern D2B503. A variable speed Moyno® pump was
used to continuously recirculate the CNF suspension through the disk refiner. An
inline heat exchanger was added to the circulation line in order to prevent the CNF
suspension from overheating during processing (the system typically reached a
steady state temperature near 50°C while processing CNF suspension). Warming
of the suspension and disk refiner during the processing caused the refiner gap to
decrease, so the gap was adjusted on-the-fly such that the processing power was
maintained at 20 kW; the gap was estimated to be 150-200 microns. A valve for an
air bleed was installed near the intake of the disk refiner to prevent the disk refiner
from flooding, which causes a pumping suction at the inlet and a surge in power
consumption.
Disk refiner tests were done using 60L of 0.1 wt% carboxylated pulp (0.76
mmol -COONa/g), which had been produced using the TEMPO/sodium chlorite
method. This suspension was processed in the disk refiner for 45 min while
circulating at a rate of 60 L/min. As before, samples of the suspension were
centrifuged at 4500G for 10 min to remove material that was not fully suspended.
Just 13% of the CNF successfully suspended. The process was repeated using a
2 wt% suspension. The suspension became quite viscous and it was necessary
to manually stir the tank to prevent channeling of processed CNF through the
tank. Periodic samples were collected, diluted to 0.1 wt% and centrifuged
as before. At the higher percent solids, 45 % of the CNF was colloidally
suspended. At pilot scale, the bulk suspension was diluted to 0.1 wt% passed
twice through the Sharples centrifuge, concentrated to about 0.5 wt% solids using
tubular ultrafiltration, and clarified by passing once through the Microfluidizer
homogenizer.
235
Increasing the disk refiner processing time to 2 h increased the colloidal yield
marginally to 50%. The Sharples centrifuge, operates as a batch process with
respect to retained solids. It needs to be stopped periodically to empty the solids
from the bowl. The run time could be increased with a crude initial separation of
the largest particles using a side-hill screen prior to centrifugation.
Nanocelluloses: Their Preparation, Properties, and Applications
Downloaded from pubs.acs.org by CITY UNIV OF HONG KONG on 10/25/17. For personal use only.
Table II. Percent of CNF dispersed using various dispersion techniques
Disk refiner
TEMPO/sodium chlorite pH 7
0.76 mmol COONa/g pulp, 0.1 wt%
Multistage centrifugal pump
TEMPO/hypochlorite pH 10
1.3 mmol COONa/g pulp, 0.1 wt%
0.125 days
6 %
5 min
4.4 %
1 day
21 %
10
7.8 %
2 days
31 %
15
8.6 %
5 days
57 %
20
9.2 %
30
10.7 %
45
13.0 %
Continued processing of rejects
5 days
18 %
Disk refiner
TEMPO/sodium chlorite pH 7
0.76 mmol COONa/g pulp, 2 wt%
Disk refiner
TEMPO/hypochlorite pH 10
1.4 mmol COONa/g pulp, 2 wt%
0 min
5.8 %
0 min
3.7 %
5
24.7 %
5
9.5 %
10
47.9 %
10
16.4 %
15
65.1 %
15
21.3 %
20
76.3 %
20
23.9 %
30
89.2 %
30
31.7 %
45
91.3 %
45
45.2 %
The refined solids were diluted from 2 to 0.35 % and circulated across a 100mesh, stainless-steel screen tilted at a 45 degree angle. Additional water was added
to the circulating rejects until the filtrate passing the screen ran relatively clear,
Dilution was intentionally limited to not exceed 20 times the refining volume.
Removing the coarsest material allowed the Sharples centrifuge to be operated
with limited stoppages for emptying collected solids.
Testing the refiner method on TEMPO oxidized pulp using the hypochlorite
method reduced the amount of time needed recirculating in the refiner while
also increasing the yield to 90%. Whereas the chlorite method CNF could only
be concentrated to 0.5% in the membrane system, the combination of higher
carboxylate levels, up to 1.5 mmol -COONa versus 0.8 mmol -COONa, combined
with the lower degree of polymerization of the cellulose, allowed the tubular UF
236
Nanocelluloses: Their Preparation, Properties, and Applications
Downloaded from pubs.acs.org by CITY UNIV OF HONG KONG on 10/25/17. For personal use only.
system to increase the concentration of hypochlorite oxidized CNF to about 1
wt% solids before the viscosity became too high to maintain flow in the membrane
tubes. The comparison of CNF dispersion using a disk refiner is shown in Table
II with sugar analysis of treated pulp as well as refined CNF and rejects in Table
III for the disk refiner at 2% solids produced using the TEMPO/sodium chlorite
method.
Table III. Sugar analysis of carboxylated eucalyptus pulp and refined (2%
solids) CNF produced using the TEMPO/sodium chlorite method buffered
at pH 7 with phosphate followed by disk refining. (Arabinose, galactose and
mannose were not detected.)
Glucose
Xylose
Total
Carbohydrates
Eucalyptus
79.7 %
15.1 %
94.9 %
TEMPO reaction,
0.76 mmol/g
61.9 %
12.5 %
74.4 %
Colloidal CNF
60.4 %
11.1 %
71.6 %
CNF rejects
65.2 %
7.8 %
73.0 %
Sample
Figure 1 shows rheology data for samples of TEMPO-CNF prepared with
the different oxidants using a Bohlin Rheometer operating with a C25 cup and
a 2000g-cm torsion element. The viscosity of a 0.38% solids CNF suspension
using the TEMPO/sodium chlorite method is comparable to the 0.95% solids CNF
suspension using the TEMPO/hypochlorite method. Additionally, Figure 1 shows
both CNF suspensions are thixotropic, meaning it is a non-Newtonian fluid, with a
viscosity changing with shear rate, and with shear history. Thixotropy is a common
characteristic of many gels.
Having determined that other equipment was not doing an adequate job of
dispersing the TEMPO treated CNF to a clear suspension, FPL switched to using
homogenizers for the complete dispersion process rather than just the clarification
step. Using the TEMPO/hypochlorite reaction provides a near 100 % yield during
the dispersion without the need for separation or concentration steps.
The Forest Products Laboratory has had access to three different
homogenizers over the years for dispersing CNFs: a Microfluidizer Model
M-110EH (MicroFluidics, Inc., Westwood, Massachusetts), a mini-DeBee (Bee
International, South Easton, Massachusetts) and a GEA Niro Soavi 3015EH (GEA
North America, Columbia, Maryland). All three use one or multiple plungers or
pistons, to develop high fluid pressures. The fluid is then dropped rapidly across a
reaction zone to create massive amounts of turbulence that act to disrupt the fiber
structure into individual nanofibrils. The Microfluidizer combines a small orifice
with impinging flow: a “Y-cell” which divides the high-pressure stream into two
which are each passed through small orifices that are directed at each other to
increase turbulence further. The Mini DeBee also passes the high-pressure fluid
through a small orifice followed by a secondary turbulent zone. There is some
237
Nanocelluloses: Their Preparation, Properties, and Applications
Downloaded from pubs.acs.org by CITY UNIV OF HONG KONG on 10/25/17. For personal use only.
flexibility as to how this secondary zone is set-up; FPL only used a series of
varying diameter orifices that created numerous zones of shear and turbulence in
series after the primary homogenizing orifice. The GEA Niro Soavi system uses
an opposing valve where the high-pressure fluid is passed through an annular gap
formed by a static “passage head” and free-floating “impact head.” The size of
the gap is controlled by the pressure applied to the “impact head;” the gap size is
also influenced by fluid properties and flow rates.
Figure 1. Viscosity characteristics of CNF suspensions as shear rates are
increased then decreased.
The three systems have comparable performance. Operating procedures
generally consisted of three passes, each with increasing operating pressure or
decreasing orifice diameter or gap size. With the microfluidizer, the initial pass
was used a 200-micron Y-cell operating at about 600 bar followed by two passes
using an 87-micron Y-cell operating at 1400 bar. The Mini DeBee also used a
200-micron orifice operating at about 600 bar for the first pass. This initial pass
was conducted at near 5% solids followed by dilution to 1% solids. Then the
CNF suspension was then passed through a 125-micron orifice operating at about
1000 bar followed by a pass through a 100-micron orifice operating at about 1400
bar. Both the microfluidics and Mini DeBee suffer plugging of the orifice with
dirt and debris that presumably came in with the pulp. This problem was reduced
by pumping the suspension through a 150 mesh screen after the first pass through
the homogenizer using the 200 micron orifice
Both the Microfluidizer and Mini DeBee were limited by capacity, operating
at rates near 200-250 mL/min. For the FPL pilot plant producing TEMPO treated
238
Nanocelluloses: Their Preparation, Properties, and Applications
Downloaded from pubs.acs.org by CITY UNIV OF HONG KONG on 10/25/17. For personal use only.
pulp in 2 kg batches, five full working days were required to process 2 kg of CNF.
FPL has recently purchased a GEA Niro Soavi 3015EH. With a similar methodogy,
CNF is dispersed by an initial pass at 600 bar at 1.5 to 2 wt % solids. After dilution
to 1 wt% solids, a second pass is performed at 900 bar followed by a third at 1200
bar. This system is capable of operating at up to 240 L/hour so processing 2 kg
of CNF suspension (200L at 1 wt% solids) over three passes can be produced in
about 3 h. This system has not had problems with debris plugging the opposing
valve. The limited amount of debris likely accumulates at the valve until the end
of a pass, when the valve is opened. However, with this opposing valve system,
great care must be taken with plumbing design and operation to ensure that no
significantly sized air bubbles are fed into the system as this causes catastrophic
failures of the valves. Maintenance appears to be a normal process for machines
operating at such high pressures. All three systems have been hampered by routine
maintenance and repair needs.
The use of a homogenizer, or any aggressive processing method for that
matter, for the dispersion of CNF seems to be a balance between improving
the dispersion of CNF as seen through increasing light transparency versus
degradation of the CNF as seen through decreasing viscosity and mechanical
properties. Figure 2 shows that the first passes of CNF through the homogenizer
have a significant effect on CNF suspension light transparency, subsequent
passes show diminishing improvements with each additional pass. Furthermore,
the initial passes results in a significant increase in suspension viscosity, but
beyond the first a couple passes at the most extreme homogenizer conditions, the
suspension viscosity is observed to plateau then decline. This drop in viscosity
with extensive processing probably indicates damage to the CNF and will likely
result in diminished mechanical strengths of films or other applications.
Figure 2. Light transmittance of CNF suspension with increased levels of
processing using a Microfluidizer.
239
Nanocelluloses: Their Preparation, Properties, and Applications
Downloaded from pubs.acs.org by CITY UNIV OF HONG KONG on 10/25/17. For personal use only.
Freeze Drying Cellulose Nanomaterials
Cellulose, is hydrophilic and when the surface has been modified with
carboxylate groups, as with cellulose nanofibrils (CNFs), or sulfate groups, as
with cellulose nanocrystals (CNCs), they are even more hydrophilic. There are
obvious economic issues associated transporting bulk CNFs around the country,
or the world, as 1 wt % suspensions. Furthermore, many of research efforts are
aimed at incorporating cellulose nanomaterials into hydrophobic matrixes and
require a dry starting material. The critical need is for methods that provide a dry
cellulose nanomaterial that can be redispersed easily at the point of use. Ideally,
the method doesn’t have additives which need to be removed before use or which
are detrimental to certain applications. For some applications, a material that
is highly porous such that it can be impregnated with other matrixes is also of
value. Attempts to redisperse air-dried cellulose nanomaterials demonstrated this
was inadequate for many potential applications. FPL focused on freeze drying
as a means to provide a dry form of cellulose nanomaterials that best fit most
application needs. It should be noted that these techniques were developed at
FPL using cellulose nanocrystals but were generally assumed to be applicable to
cellulose nanofibrils as well.
Figure 3. Cellulose nanocrystals freeze dried and resuspended (6 wt%) with
only stirring after different freezing methods compared to the initial (ini) CNC
suspension (6 wt%). Sample A: simple, slow freezing of dish placed in freezer.
Sample B: Rapid freezing of dish placed in liquid nitrogen. Sample C: Frozen
layer that developed in aqueous suspension in small ice cream maker. Sample D:
Frozen in small ice cream maker with 10 vol% t-butanol added.
240
Nanocelluloses: Their Preparation, Properties, and Applications
Downloaded from pubs.acs.org by CITY UNIV OF HONG KONG on 10/25/17. For personal use only.
When a suspension of aqueous cellulose nanomaterials is frozen and the
water removed under high vacuum, e.g., freeze drying, the resulting dry material
is a crunchy, crumbly mass containing large flakes and gaping pores. When
mixed with water, the dried material produces a hazy suspension with some
particles large enough to be seen with the naked eye (Figure 3A). Some of the ice
crystals that formed during freezing, grew very large, and glacially concentrated
the cellulose nanocrystals to such an extent that CNC films were literally cast
around the ice crystals. The laboratory solution was to freeze the CNC and CNF
suspensions as rapidly as possible by adding the suspension dropwise into liquid
nitrogen (LN2). While the ability to redisperse is improved (Figure 3B), the
method is not practicable at larger scale.
Among applications where water needs to be frozen and ice crystal growth
minimized includes ice cream and some types of frozen drinks. FPL purchased
a small one-pint kitchen ice cream maker. The initial test with CNC suspension
did not go well as a thick layer of ice quickly formed on the bowl of the mixer
because the blade was unable to scrape the ice from the sides of the bowl. Freeze
drying of this frozen layer, however, produced a CNC powder that resuspended
easily and looked similar to material dried after flash freezing with LN2 (Figure
3C). In ice cream, the dissolved sugar and suspended fats help control how the
ice cream suspension freezes. What was needed was an additive that would
provide a freezing point depression like the sugar in ice cream. More specifically
a freezing point depression results in freezing over a larger temperature range
than is obtained with pure compounds. In addition, the additive needed to have a
number of other properties to make it easier to work with and protect the freeze
drying equipment: a freezing point similar to water so it would be captured in the
freeze drier condenser and not damage the vacuum pump, and a vapor pressure
higher than water so it would sublimate from the frozen CNC suspension and not
contaminate the product. Tertiary-butyl alcohol, was selected. It has very high
water solubility, a freezing point of 25°C and a boiling point of 83°C. Ten percent
by volume was added to a rapidly stirred CNC suspension and in which was then
transferred to the mini ice cream maker. Twenty minutes later a small bowl of
CNC similar to a soft serve ice cream was transferred to a freezer to complete
the freezing. After freeze drying the starting dry mass of dry CNCs was attained,
confirming that the t-butanol had been removed and much of the t-butanol was
captured in the condenser of the freeze drier. Upon resuspending in water with
minimal mixing, the CNCs looked similar to the original, never-dried CNC
suspension (Figure 3D). Optimizing the t-butanol addition, a 5% concentration
did not prevent the freezing problems with the mixer seizing, and addition levels
above 10% did not appear to provide additional benefits. The 10% addition level
was selected to provide a margin of error for scale-up. FPL was able to obtain an
unused convenience store frozen drink machine to produce the 150L of cellulose
nanomaterial as a frozen slush needed to load a Vertis GPFD-XL35 freeze drier.
With this equipment, a 150 L batch requires about three weeks to complete the
drying process. The freeze drier has limited condenser capacity and needs to
be operated at reduced vacuums to avoid collecting water in the vacuum pump
oil. The t-butanol is recovered from the condenser and distilled to the 87 vol%
t-butanol azeotrope for reuse.
241
The CNCs are typically freeze dried from a 10 wt% suspension in water and
t-butanol and collected as dry sheets that crumble to a fine powder. TEMPO CNFs
are freeze dried from a 1 wt% suspension in water and 10% t-butanol. The resulting
sheets are similar to a Styrofoam board.
Nanocelluloses: Their Preparation, Properties, and Applications
Downloaded from pubs.acs.org by CITY UNIV OF HONG KONG on 10/25/17. For personal use only.
Conclusion
FPL has been working toward producing cellulose nanomaterials at a scale
suitable for preliminary research and product development. A simple process
schematic that FPL is currently using for the conversion of bleached Kraft pulp to
CNFs is shown in Figure 4. Cellulose nanocrystals have been distributed through
the University of Maine’s Process Development Center since 2012 while cellulose
nanofibrils produced using the TEMPO/hypochlorite method at pH 10 have been
available since late 2015. Both are available as never-dried aqueous suspension
and freeze-dried solids. FPL has scaled established laboratory methods for
production of TEMPO-grade cellulose nanofibrils to several kilograms per batch.
First, commercial bleached eucalyptus drylap is repulped and pretreated with
sulfuric acid and with sodium chlorite at pH 2 to remove non-process elements
and oxidize any residual lignin or chromophores that might consume reagents
during the TEMPO treatment. The pulp is oxidized using TEMPO and sodium
hypochlorite at pH 10 over the course of several hours under ambient conditions
resulting in carboxylation of the surface of the elementary nanofibrils within the
pulp fiber structure to levels near 1.5 mmol/g pulp. The washed, carboxylated
pulp fiber, now swollen but intact, is subjected to a high shear environment
in order to disrupt the fiber structure and disperse the carboxylated nanofibrils
as a clear, viscous, colloidal suspension of approximately 1 wt% solids for
distribution. FPL has also developed a method producing freeze-dried CNF. The
aqueous suspension is mixed with 10% by volume t-butanol and partially frozen
with a commercial soft-serve machine. This is spread into trays, frozen solid
overnight and freeze-dried over the next several weeks. (The FPL freeze dryer is
limited by its condenser capacity.)
There are multiple aspects of the CNF chemistry and processing that must
be balanced with each other to optimize once potentially successful uses begin
to emerge. Does the application utilize the carboxylate surface chemistry of the
TEMPO-grade CNF? If so, what level is required? Questions of surface chemistry
must also be balanced with their effect on physical properties. Generally
speaking, the higher the carboxylate level of the CNF, the more easily the CNF
can be dispersed into a colloidal suspension. High levels of suspension and
film transparency can ultimately be achieved with higher carboxylate levels on
the CNF. However, higher levels of carboxylation generally results in a lower
degree of polymerization of the cellulose chains within the nanofibril. This can
potentially limit ultimate strength of the nanofibril and films or matrices into
which it is incorporated.
242
Nanocelluloses: Their Preparation, Properties, and Applications
Downloaded from pubs.acs.org by CITY UNIV OF HONG KONG on 10/25/17. For personal use only.
Figure 4. Process schematic for producing CNF from commercial bleached Kraft
pulp. CNF is produced as aqueous suspension following mechanical dispersion
of pulp carboxylated using TEMPO and sodium hypochlorite. Freeze drying is
carried out as needed to produce dry CNF.
Cellulose nanofibrils, and cellulose nanomaterials more generally, are a
novel form of cellulose where the potential applications are just beginning to
be explored. As with any new material, there are likely to be opportunities
where cellulose nanomaterials provide just the right combination of chemical
and physical properties to create new and unique applications that couldn’t be
achieved, or even conceived, prior to their availability. But just as likely, cellulose
nanomaterials will also be quietly incorporated into many products we are already
using. Since cellulose nanomaterials are derived from renewable, bio-based
feedstock, any application utilizing them will be considered to be “greener” than
it was before, either from the standpoint of the ease of recyclability or from
limited environmental impact caused by disposal. We are even likely to find some
products, after use, might be disposed of by tossing it in the backyard compost
pile. But “greener” also includes improving durability, increasing longevity and
reducing material consumption, so perhaps we’ll also have things in front of us
for decades with cellulose nanomaterials inside.
References
1.
2.
3.
Turbak, A.; Snyder, F.; Sandberg, K. Microfibrillated cellulose, a new
cellulose product: Properties, uses and commercial potential. J. Applied
Polymer Sci. 1983, 37, 815–827.
Pääkkö, M.; Ankerfors, M.; Kosenen, H.; Nykänen, A.; Ahola, S.;
Österberg, M.; Ruokolainen, J.; Laine, J.; Larsson, P. T.; Ikkala, O.;
Linström, T. Enzymatic hydrolysis combined with mechanical shearing
and high-pressure homogenization for nanoscale fibrils and strong gels.
Biomacromolecules 2007, 7, 1687–1691.
Zimmerman, T.; Pöhler, E.; Geiger, T. Cellulose fibrils for polymer
reinforcement. Adv. Eng. Mater. 2004, 6, 754–761.
243
4.
5.
Nanocelluloses: Their Preparation, Properties, and Applications
Downloaded from pubs.acs.org by CITY UNIV OF HONG KONG on 10/25/17. For personal use only.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
Saito, T.; Isogai, A. TEMPO-Mediated Oxidation of Native Cellulose. The
Effect of Oxidation Conditions on Chemical and Crystal Structures of the
Water-Insoluble Fractions. Biomacromolecules 2004, 5, 1983–1989.
Saito, T.; Kimura, S.; Nishiyama, Y.; Isogai, A. Cellulose Nanofibers
Prepared by TEMPO-Mediated Oxidation of Native Cellulose.
Biomacromolecules 2007, 8, 2485–2491.
Saito, T.; Hirota, M.; Tamura, N.; Kimura, S.; Fukuzumi, H.; Heux, L.;
Isogai, A. Individualization of Nano-Sized Plant Cellulose Fibrils by Direct
Surface Carboxylation Using TEMPO Catalyst under Neutral Conditions.
Biomacromolecules 2009, 10, 1992–1996.
Wågberg, L.; Decher, G.; Norgren, M.; Lindström, T.; Ankerfors, M.;
Axnäs, K. The Build-Up of Polyelectrolyte Multilayers of Microfibrillated
Cellulose and Cationic Polyelectrolytes. Langmuir 2008, 24, 784–795.
Liimatainen, H.; Visanko, M.; Sirviö, J. A.; Hormi, O. E. O.; Niinimaki, J.
Enhancement of the Nanofibrillation of Wood Cellulose through Sequential
Periodate–Chlorite Oxidation. Biomacromolecules 2012, 13, 1592–1597.
Cellulose Filaments: An Exciting Opportunity for Canada’s Forest Sector.
FPInnovations Fact Sheet, June, 2013. https://fpinnovations.ca/media/
factsheets/Documents/cellulose-filaments.pdf (accessed July 7, 2017).
Miller, J. Nanocellulose State of the Industry, December, 2015.
http://tappinano.org/media/1114/cellulose-nanomaterials-production-stateof-the-industry-dec-2015.pdf (accessed July 6, 2017).
Kinney, J. Southworth Co.’s Paperlogic Hopes Innovative Nanocellulose
Papers Keep Turners Falls Mill Busy. Mass Live. December 19,
2014.
http://www.masslive.com/business-news/index.ssf/2014/12/
southworth_co_s_paperlogic_hoping_to_dev.html (accessed July 6, 2017).
FPInnovations and Kruger Celebrae Dedication of Cellulose Filament
Demostration Plant.
http://www.kruger.com/en/news/celebratinginnovation/ (accessed July 7, 2017).
Announcement of the Installation of a New Manufacturing
Demonstration
Facility
OJI
Holdings
Has
Developed
a
Ground-Breaking “Phosphate Esterification” Production Process
for Cellulose Nanofibers.
https://www.ojiholdings.co.jp/english/
news/?itemid=971&dispmid=1282&TabModule1281=0 (accessed July 6,
2017).
Nippon
Paper
Industries
Operates
the
World’s
Largest
Cellulose Nanofiber Production Facility at the Ishinomaki
Mill.
http://www.nipponpapergroup.com/english/news/year/2017/
news170425003763.html (accessed August 1, 2017).
Announcement of Operation of Pilot Plant of Cellulose Nanofibers.
https://translate.google.com/translate?hl=en&sl=ja&u=https://www.daiopaper.co.jp/news/2016/pdf/n280509b.pdf (accessed August 1, 2017 using
Google Translate).
The World’s First Commercial Production Facility of Microfibrillated
Cellulose.
http://www.exilva.com/The-world-s-first-commercialproduction-facility-of-Microfibrillated-cellulose (accessed July 6, 2017).
244
Nanocelluloses: Their Preparation, Properties, and Applications
Downloaded from pubs.acs.org by CITY UNIV OF HONG KONG on 10/25/17. For personal use only.
17. Betulium: A Novel Chemical Derived from Agricultural by-Products.
https://www.tekes.fi/en/tekes/results-and-impact/cases1/2016/betulium-anovel-chemical-derived-from-agricultural-by-products (accessed July 6,
2017).
18. Pilot Plant for Nanocellulose. http://www.innventia.com/en/Our-Ways-ofWorking/Demonstration-and-pilot/Pilot-plant-for-nanocellulose/ (accessed
July 6, 2017).
19. A Future for MFC. http://www.storaenso.com/newsandmedia/a-future-formfc (accessed July 7, 2017).
20. Norske Skog Saugbrugs Develops New Technology to Produce
Better Paper and New Products.
http://www.norskeskog.com/Shownews.aspx?PID=3264&M=NewsV2&Action=1&NewsId=287
(accessed
July 7, 2016).
21. Nanocellulose Film: Thorough Understanding of the Suitability of
Nanocellulose for Different Applications. http://www.vttresearch.com/
services/bioeconomy/light-weight-bio-based-structures/novel-webproduction-concepts/nanocellulose-film (accessed August 1, 2017).
22. Sugar Beet Waste Product Could Be Billion Dollar ‘Wonder Material.’
Reuters. September 16, 2015. http://www.reuters.com/article/us-suger-beetidUSKCN0RG29I20150916 (accessed July 7, 2017).
23. Sappi to Locate Ground-Breaking Nanocellulose Pilot Plant at Brightlands
Chemelot Campus in the Netherlands. https://www.sappi.com/sappi-locateground-breaking-nanocellulose-pilot-plant-brightlands-chemelot-campusnetherlands (accessed July 7, 2016).
24. Davis, M. W. A rapid method for compositional carbohydrate analysis of
lignocellulosics by high pH anion-exchange chromatography with pulse
amperometric detection (HPAE/PAD). J. Wood Chem. Technol. 1998, 18,
235–252.
245
Документ
Категория
Без категории
Просмотров
8
Размер файла
1 011 Кб
Теги
1251, 2017, ch012
1/--страниц
Пожаловаться на содержимое документа