EPA 600-A-94-181
Session and Paper Number: I3C
EVALUATION OF SUPERCRITICAL CO2 SPRAY TECHNOLOGY
AS A COST EFFECTIVE APPROACH TO REDUCTION
OF SOLVENTS IN WOOD FINISHING
A. B. Parsons and K. J. Heater, Battelle, Columbus, OH,
and P. M. Randall, U.S. EPA, Cincinnati, OH
Prepared for presentation at
AIChE 1994 Summer National Meeting/August 14-17, 1994
"Cost Effective Technology — Efficient Waste Minimization"
August, 1994
"UNPUBLISHED"
"AIChE shall not be responsible for statements or opinions contained in papers
or printed in its publications."
-------
ABSTRACT
Product quality, waste reduction, and economic issues were evaluated for a spray paint
application technology using supercritical carbon dioxide to replace some of the solvent in a
conventional solvent-borne coatings formulation. Product quality was evaluated by comparing
product finishes for a nitrocellulose coating applied by conventional spray with that of a similar
coating applied by supercritical carbon dioxide (CO2) technology. Waste reduction and
economics were documented from company records and interviews with key company personnel.
The technology was found to have good potential to reduce waste without affecting product
quality.
INTRODUCTION
This program was conducted by Battelle for the Pollution Prevention Research Branch
(PPRB) of the U.S. Environmental Protection Agency with the cooperation of Union Carbide
Corporation, Nordson Corporation, and Pennsylvania House Furniture Company. The PPRB
is evaluating and demonstrating new technologies for pollution prevention through the Pollution
Prevention Clean Tecfiholdgy Demonstration (CTD) Program.1
This report reviews the use of supercritical CO2 technology for paint spray application.
Pennsylvania House Furniture Company has used supercritical CO2 coating technology for more
than a year to apply a nitrocellulose lacquer finish to oak and cherry furniture on a chair-finish-
ing line. At current Pennsylvania House production rates, more than 250 furniture units per day
are coated with nitrocellulose lacquer by mis process.
-------
2
During the subject technology evaluation, three aspects of this technology were examined:
• Product Quality: To show that coating applied by this spray technology meets company
*
standards for a quality finish
• Pollution Prevention Potential: To demonstrate that use of this spray application
technology for solvent replacement in coatings reduces volatile organic compounds
(VOCs) released during finishing operations
• Economic Ramifications: To document the cost to install and operate this pollution
prevention technology on an existing spray coating finish line.
PROCESS DESCRIPTION
In the supercritical CO^ spray process, the solvent-like properties of supercritical CO2 are
exploited to replace a portion of the solvent in the conventional solvent-borne coating formula-
tion. Supercritical CO^ acts as a diluenlasolvent to thin die viscous coating just before
application, allowing the coating to be *tnm\-7t*\ and applied with a modified spray gun. Typi-
cally, most of the volatile, fast-drying solvents and some of the medium-drying solvents are
eliminated. Enough medium- and slow-evaporating solvents are retained to obtain proper level-
ing and film coalescence. The remaining solvent blend is adjusted to optimize performance
without changing the resin chemistry or pigment-loading levels. The actual amount of solvent-
-------
3
content reduction that can be achieved is affected by the type of coating, the desired film
thickness, the desired properties of the applied coating, and the environment in which the coating
is being applied. Union Carbide Corporation developed the use of supercritical CO2 for spray
coating applications, introducing this technology commercially in 1988 under the UNICARB™
tradename.2'3 Thermosetting, thermoplastic, air-dry, and two-component formulations, in clear,
pigmented, and metallic coating systems, have been developed successfully for use with the
UNICARB™ process.
Supercritical fluids are gases that exist at temperatures and pressures near or above the
critical point of the fluid as depicted on a phase diagram (Figure 1). At the critical point, the
properties of the liquid and the gas are similar or identical. The resulting single-phase fluid
exhibits solvent-like properties that can be altered by adjusting temperature and pressure. A
number of gases have been examined for use as supercritical fluids in applications such as
industrial and analytical separation processes, cleaning, chromatography, and coating. The
UNICARB™ process for coating uses nontoxic, nonflammable carbon dioxide as the supercritical
fluid for coating dilution. Carbon dioxide, readily available as a by-product of a variety of
industrial processes, has a critical temperature of 31 *C (88*F) and a critical pressure of 73 atm
(1070 psi), falling within the ranges already used for heated paint systems and airless spray
equipment.
Special equipment is needed to introduce the CC^ into the reduced-solvent formulations and
to heat and pressurize the resultant mixture before spraying. Typically 10% to 50% by weight
COj may be introduced. Usually, the coating is heated to 40* to 70*C and applied with spray
pressures of 82 to 109 atm (1200 to 1600 psi). Coatings used at Pennsylvania House are applied
-------
4
with spray guns similar to those used for airless applications. However, the spray-gun nozzle
design was s'ightly modified to optimize the spray pattern because the decompression of
supercritical CO2 results in finer atomization of the sprayed coating and smaller particles than
are common with the use of airless spray equipment.
Pennsylvania House uses supercritical CO2 spray technology on the chair-finishing line to
apply nitrocellulose lacquer finishes while reducing the VOC emissions from their finishing
operation. To bring this technology to production-line use, Pennsylvania House worked closely
with manufacturers and suppliers to optimize supercritical CO2 technology including process,
equipment, and coatings formulations. In the conventional finishing process, two coats of nitro-
cellulose lacquer (21 to 23% solids) are applied manually with airless spray equipment. The
supercritical CO2 finishing process uses only one coat of nitrocellulose lacquer to achieve the
desired film build and finish quality. The nitrocellulose lacquer formulation, optimized for the
supercritical CO2 spray system, has approximately a 41 % solids content.
PRODUCT QUALITY EVALUATION
ma : The specific objective of the product quality evaluation was to determine whediernitrocellu-
lose lacquer, applied by the supercritical CO2 spray process, provided a wood finish of equal
or better quality man that applied by the conventional nitrocellulose formulation and spray
technique previously used by Pennsylvania House. At Pennsylvania House, the appearance and
quality of the final finish are judged through visual examination by inspectors on me coating
-------
5
line. Special attention is given to gloss, smoothness, and the lack of surface defects such as
blisters or pinholes.
Product quality was evaluated through subjective evaluations performed by Pennsylvania
House staff members and a panel of Battelle coatings personnel. Test substrates were finished
on the production lines during an on-site visit. A set of samples was finished using the one-coat
UNICARB™ process, and two other sets were finished using one and two coats of the "standard"
nitrocellulose formulation and the airless spray equipment still in place on the chair-finishing
line. All panels were finished by the same production methods that typically are used on the
chair line at Pennsylvania House.
The subjective product quality evaluations demonstrated that a coating applied by the
supercritical CO2 spray process yielded a product with a finish quality equal to or better than
the finish quality obtained by conventional materials and methods. Samples finished by the
supercritical CO2 process and by the two-coat conventional process were rated as "acceptable"
(Table 1). Samples sprayed with one coat of the nitrocellulose by the conventional process were
not acceptable.
In addition to the subjective visual inspections of the test samples, Battelle staff made mea-
surements of gloss (Table 2) using ASTM D529. These measurements provide some quantitative
insight into the physical attributes of the finish of each of the coating processes.
The test procedure outlined in ASTM DS29 recommends averaging six gloss measurements
for a 3-inch x 6-inch sample area, which correlates to 49 measurements on the 7-inch x 21-inch
test substance used here. The mean and standard deviation of the 49 data points represent the
overall gloss appearance of each sample, alleviating subjective biases of the person performing
-------
6
the measurements while still incorporating the assessment of any nonuniformity in the gloss
across the sample surface. The breadth in standard deviation of the data can be used as a gauge
of the uniformity of the sample finish across the complex geometry of the test panels. Gloss
9
test results for each of the nine panels are reported as the mean of 49 determinations and then
averaged for each of the sample sets for easy comparison among each of the finishing processes
hi Table 2. The averaged gloss data for the UNICARB™ samples are statistically the same as
those for the conventional two-coat process. The gloss data of both of these sets show that they
are substantially glossier than the one-coat conventional finish sample set.
These results are supported by Pennsylvania House records for consumer acceptance.
Internal quality control audits on chair-line products show a decrease in finish defects using the
supercritical CO2 system. Based on the number of furniture units requiring rework because of
finish defects, production efficiency has improved since the UNICARB™ process was
implemented.
POLLUTION PREVENTION POTENTIAL
The pollution prevention! potential of mis technology is based on reducing the emissions of
organic solvents without adding to other wastestreams. The nitrocellulose lacquer finishing
process used on the chair line can contribute to pollution in two ways: VOC emissions from the
coating formulation during the finishing operation, and spray-booth wastes, including solvent-
laden filters and nitrocellulose "dust". A blend of fast-evaporating solvents, medium-evaporating
solvents, and slow-evaporating solvents is used hi conventional spray coatings. In the supeicriti-
-------
7
cal CO2 spray process, most of the fast- and medium-drying solvents are replaced by supercri-
tical CO2 and the slow-drying solvents are adjusted slightly for better film formation. Although
reducing VOC emissions is important, it is equally important to demonstrate that the supercritical
CO2 process does not add pollutants to other wastesireams.
Pennsylvania House has been able to reduce the number of coats of nitrocellulose lacquer
from two to one. The volume of nitrocellulose lacquer used in each finishing operation was
determined during the initial phases of implementing the supercritical CO2 process at
Pennsylvania House. Metering devices were placed in-line on the airless spray guns used to
apply the conventional nitrocellulose formulation, and on the coating inlet line to the
supercritical fluid supply unit used to feed the coating mixture concentrate and the supercritical
CO2 to the modified spray guns used with the UNICARB™ process.
Pennsylvania House records indicate that it takes approximately 473 ml of the conventional
formulation to apply the two coats needed to achieve the desired quality in the finished product.
The UNICARB™ process required about 207 ml of the reduced-solvent formulation per furniture
unit to achieve the same quality.
Table 3 compares the volatile solvent content (% by weight) of the two formulations used
by Pennsylvania House. The UNICARB™ coating is formulated using 17.5% less solvents (on
an absolute basis) than the conventional formulation. Only 9.67% of the UNICARB™ formu-
lation is comprised of Hazardous Air Pollutant (HAP) materials, compared to 35.78% for the
conventional formulation. On a per-gallon-of-coating-sprayed basis, this difference would result
in a relative decrease in VOC emissions of 22.81 %, with a 72.97% decrease in HAPs using the
-------
8
UNICARB™ formulation. VOC contents are reported as 563 gms/f for the UNICARB™
formulation and 707 gms/f for the conventional system.
Assuming an average yearly production of 50,000 units and the use of 207 ml for the one-
»•
coat UNICARB™ process and 474 ml for the two-coat conventional formulation, the UNICARB™
formulation uses 10,220 I and the conventional formulation uses 24,604 i to finish the units.
Based on the reported VOC contents, this system change corresponds to an annual reduction in
VOC emissions of 67.5% when the newer process is used.
Supercritical C02 is used hi the UNICARB'* process to decrease VOC emissions. Carbon
dioxide is not being produced through use of the UNICARB™ process. The CO2 used hi this
technology is supplied by various distributors of CO2 which obtain CO2 as a by-product of other
chemical processes. Thus, the supercritical CO2 method of applying coatings does not actually
contribute to the emission of additional CO2 into the atmosphere.
Coating overspray at Pennsylvania House is collected on dry filters that are compressed and
stored hi 208-Ł (55-gal) drums for disposal by landfill. Waste products include dry and solvent-
laden filters and nitrocellulose "dust", both loose and trapped in the filters. The solid waste was
not increased or decreased by implementing the supercritical COj technology.
ECONOMIC ANALYSIS
The objective of the economic analysis was to determine the payback period for the switch
to the supercritical CO2 process from the previously used conventional system. The initial
investment in capital equipment and installation costs were considered along with operating costs
-------
9
(materials, waste disposal, labor, and utilities). A return-on-investment (ROI) was calculated,
based on the costs associated with capital expenditures, including equipment and installation, and
the return generated through lower personnel, operating, and materials costs. Details on this
ROI calculated using worksheets provided in the Waste Minimization Opportunity Manual (U.S.
EPA, 1988) are included in the full report.1
Implementing the UNICARB™ finishing process on the chair line at Pennsylvania House
resulted in substantial annual savings in both utilities and labor as shown in Table 4. The annual
operating costs were based on the production of 50,000 chairs per year. The UNICARB™
process costs of $46,000 include $37,000 for the coatings formulation and CO2 concentrate and
$9,000 for the CO2 equipment rental. The conventional formulation costs for the same number
of furniture units would be $47,000. By converting from a two-coat process to the one-coat
process, Pennsylvania House was able to decrease its utility costs by $11,000 because there was
one less booth to operate. Labor costs were reduced by $46,000 because one less finisher and
one less sander were needed. Waste handling and disposal costs and finishing line maintenance
remained the same for both processes.
Cost savings, realized from a decrease in raw materials costs, were offset by the leasing
fees for the COj tank and pump at Pennsylvania House. Additional savings could be realized
by decreasing the size of the existing ovens to reflect the change to a one-coat system.
Pennsylvania House has not downsized the production ovens to gain gas utility savings.
The annual operating cost of the supercritical CO2 finish line is approximately $58,000 less
per year than that of the conventional line. The initial capital investment for the UNICARB™
-------
10
process was $58,000, of which $46,000 was for equipment purchase and $12,000 for installation
of the equipment.
The more detailed economic evaluation found in the full report demonstrates a positive
9
return on investment after the first year, with a total payback period within three years if gas
utility savings are included, and five years if gas utilities are not included.
CONCLUSIONS
This technology evaluation shows that supercritical CO2 spray technology has potential as
a pollution prevention option in the application of solvent-borne coatings. This supercritical CO2
technology is not limited to one coating type, but could be used to reduce the solvent level
required to spray apply a variety of solvent-borne coatings. The wood furniture facility, where
this evaluation was conducted, maintained product quality with a nitrocellulose lacquer finish and
reduced VOC emissions from the coating process. No additional wastes entered the waste-
stream. Immediate operating savings of $58,000/year were realized. A 100% ROI should be
achieved within five years after implementation.
LITERATURE CITED
1. Heater, K. J., Parsons, A. B., Olfenbuttel, R. F., Randall, P., "Evaluation of Supercritical
Carbon Dioxide Technology to Reduce Solvent in Spray Coating Applications," U.S.
Environmental Protection Agency report hi press.
2. "Carbide Licenses Coatings Technology". Chan. Engr. News, July 30, p. 9 (1990).
-------
11
3. Hoy, K., "Unicarb System for Spray Coatings — A Contribution to Pollution Prevention."
Eur. Polym. Paint Colour J. 181: 438, 440-2 (1991).
70
-62 -40 -18 4 27 49
Figure 1. Phase diagram for carbon dioxide.*
* Nielsen, K.A., Busby, D.C., Glancy, C.W., Hoy, K.L., Kuo, A.C., and Lee, C.,
"Supercritical Fluid Spray Application Technology: A Pollution Prevention Technology for
the Future," Union Carbide Chemicals and Plastics Company, Inc. Presented at the 17th
Water-Borne & Higher-Solids Coatings Symposium, February 21-23, 1990, New Orleans,
LA.
-------
12
Table 1. Sample Panel Finish Quality
Results for Subjection Gloss
Process
Conventional One-Coat Process
Conventional Two-Coat Process
Supercritical CO2 One-Coat Process
Acceptable Gloss
(% panels)
33a (33)b
100 (100)
100 (100)
Unacceptable Gloss
(% Panels)
67 (67)
0(0)
0(0)
a Evaluations by furniture company staff are stated first.
b Evaluations by Battelle Coating staff are enclosed with parentheses.
Table 2. Gloss Data on Sample Panels
Finishing Process
Conventional
One Coat
Conventional
Two Coat
UNICARB™
Sample
Number
46482-9-1
46482-9-2
46482-9-3
46482-10-1
46482-10-2
46482-10-3
46482-11-1
46482-11-2
46482-11-3
Average
Gloss Data/Panel
20.3
±4.3
20.4
±3.1
20.3
±3.1
33.2
±1.6
35.0
LK ± 2.2
28.7
±2.8
35.3
±3.2
30.5
±3.1
28.7
±2.9
Average Gloss
Data/Set
20.3
32.3
31.5
-------
13
Table 3. Comparison of Volatile Solvent Content of Conventional
and UNICARB" Coating Formulations as Percent Weight
Materials Description
MEK-heptanone
methoxypropylacetate
xylene
isopropanol
toluene
N-butyl acetate
isobutyl acetate
2-butoxyethanol
MIBK
isopropyl acetate
Other
HAP(Y/N)
No
No
Yes
No
Yes
No
No
Yes
Yes
No
Total VOC (% by weight)
Conventional
(% by weight)
7.36
16.80
11.20
10.39
11.83
6.89
3.27
5.32
1.46
2.37
76.88
UNICARB™
(% by weight)
37.25
6.55
9.67
5.87
59.34
Table 4. Summary of Annual Operating Cost Comparing
Conventional Finish Line with Supercritical CO2 Finish Line
Item
Coating Materials
CO2 Storage Equipment
Spray Booth #2
Finish Labor
Sanding Labor
Electricity
.TOTAL
Conventional
($/year)
47,000
—
23,000
23,000
11,000
104,000
Supercritical
C02 ($/year)
37,000
9,000
—
46,000
SAVINGS = 58,000
------- |