4 Microwave-Assisted Paint Removal (Task 4)
The microwave-assisted process for removing lead-based paint from wood and composites was developed and patented by A. Kumar and has been assigned to the U.S. Army (U.S. Patent No. 5,268,548, 1993). In the microwave-assisted paint removal process, microwave coupling compounds called susceptors are applied as a waterborne slurry or as a polymer binder paste to the painted surface. Microwaves have the unique ability to rapidly and selectively heat the coated surface. Compounds such as graphite or iron oxide can reach temperatures up to 1000 °C in less than a minute when exposed to microwaves (800 watts). The susceptor material increases both the efficiency of the system as well as the uniformity of the heating process. The microwave applicator uses the same standard 2.45 gigahertz magnetron tubes that are used in household microwave ovens. The applicator horn is designed to focus microwave energy onto a coupling material (the susceptor), where it is used effectively. The paint is debonded from the substrate by heat and is removed easily by scraping. A microwave shield is provided for worker protection. Since the airborne lead levels from this process are below the EPA and OSHA threshold requirements, containment structures, environmental monitoring, and worker health monitoring are not required. The microwave-assisted LBP removal system can make the worksite safer and reduce negative environmental impacts.
Two prototypes of the microwave device were tested: a CERL-developed unit designed to apply 300 W of energy over an area of 15 x 15 cm, and a unit developed by HVS Technologies (under contract to CERL) designed to apply 1000 W over an area of 2.54 x 2.54 cm. These design differences have implications for dwell time, surface temperature, ease of application, and overall paint-removal time.
Development of CERL Microwave Applicator and Paint-Removal Process
A number of laboratory experiments were necessary in order to develop the microwave-assisted paint removal equipment and procedure. Initial testing of the procedure was conducted on painted wood substrates in a conventional microwave oven. This initial testing later gave way to the development of a custom- designed microwave applicator system for removing paint from the surface of a wall. Various parameters such as microwave power, exposure time, and susceptor material were important factors in the development of the custom-designed applicator and procedure.
Experiment 1: Microwave Generator Power Output Capabilities
Purpose
The purpose was to determine the fixed distance from a microwave waveguide output that would provide maximum energy output. A custom-built Micro Dry variable-power microwave oven was used for the experiment, with a consumer-grade Sharp microwave used for purposes of comparison to determine the approximate amount of energy being absorbed at the hot spot by a specified amount of water.
Setup
Figure 1 illustrates the physical layout of the experiment.
Figure 1. Experimental setup within the custom-built, variable-power microwave oven.
Procedure
The following is a copy of the instructions for the experimental procedure:
Part A: Identify the Hot Spot in Custom-Built Oven
1. Set oven power level. Start with moderate power of 600 W; test series will be repeated at 900 W and 1200W.
2. Fill paper cups with 275 ml tap water.
3. Place cup at measured distance from waveguide opening, heat for 1 minute, record temperature using standard lab-grade pyrometer.
4. Repeat Step 3 at a different distance from the waveguide opening with another water sample; record temperature.
Part B: Measure Maximum Energy Absorbed at the Hot Spot
1. Heat water in consumer-grade microwave unit for 1 minute; record temperature.
2. Heat water in the hot spot of the custom-built microwave unit set at 1000 W for 1 minute; record temperature.
Data
Table 1 presents the data recorded for Part A of this experiment.
Table 1. Temperatures recorded at different distances from waveguide opening in custom-built variable-power microwave oven.
Distance (cm) |
Temp (°C at 600 W) |
Temp (°C at 900 W) |
Temp (°C at 1200 W) |
4 |
34 |
40 |
42 |
8 |
28 |
29 |
30 |
12 |
28 |
29 |
30 |
16 |
26 |
26 |
28 |
20 |
25 |
27 |
26 |
24 |
24 |
26 |
25 |
28 |
24 |
25 |
25 |
32 |
24 |
25 |
25 |
The data for Part B of this experiment are summarized below:
Temperature after 1 min. in Sharp oven from 24 °C: 77 °C
Temperature after 1 min. in MicroDry oven from 24 °C: 75 °C
Power calculations for Sharp oven:

Power calculations for MicroDry oven:

Power calculation notes: specific heat of water = 4.2159 J/gK at ~50 °C; density of water found to be 0.9584 g/cm3.
Results for Experiment 1
Experiment 1 indicated that there is a great deal of dispersion a short distance from the waveguide opening. With a calculated wavelength of approximately 15 cm, there were no noticeable high or low points that a standing wave would be expected to demonstrate at distances of twice its wavelength. At the 1000 W setting with the cup of water less than 1 cm from the waveguide, the MicroDry unit was able to heat the water almost as well as the conventional Sharp oven, and high power (around 944 W) is attained, but only near the waveguides opening.
Experiment 2: Effect of Resistance Levels on Heating Time
Purpose
The purpose of Experiment 2 was to determine what effect the resistance of the susceptor material has on heating characteristics, including time and burning point.
Setup
A 20 dB gain horn antenna was fitted over the opening of the waveguide for this and later experiments. Figures 2 and 3 illustrate the physical setup.
Figure 2. Laboratory test setup for testing effects of susceptor resistances on heating time.
Figure 3. Resistance measurement setup.
Procedure
The following is a copy of the instructions for the experimental procedure:
1. Measure and record resistivities of all susceptor samples, taking three measurements in each of the three regions, and then taking an average.
2. Determine a good array of highest, lowest, and intermediate resistance levels.
3. Heat each susceptor sample at full power until smoke is produced. Record exposure time and temperature in each of the three regions.
Data
Tables 2 and 3 present the data recorded for Experiment 2.
Sample |
R1 (ohms) |
R2 (ohms) |
R3 (ohms) |
R avg. (ohms) |
Sample |
R1 (ohms) |
R2 (ohms) |
R3 (ohms) |
R avg. (ohms) |
H 01 |
160 |
100 |
200 |
153 |
H 15 |
145 |
130 |
200 |
158 |
H 02 |
160 |
100 |
200 |
153 |
H 16 |
190 |
85 |
115 |
130 |
H 03 |
160 |
100 |
230 |
163 |
H 17 |
115 |
85 |
125 |
108 |
H 04 |
230 |
90 |
200 |
173 |
H 18 |
125 |
80 |
110 |
105 |
H 05 |
195 |
85 |
160 |
146 |
H 19 |
160 |
85 |
135 |
127 |
H 06 |
115 |
85 |
125 |
108 |
H 20 |
145 |
100 |
155 |
133 |
H 07 |
240 |
85 |
120 |
148 |
H 21 |
170 |
120 |
185 |
158 |
H 08 |
270 |
105 |
220 |
198 |
H 22 |
145 |
145 |
210 |
166 |
H 09 |
160 |
75 |
115 |
116 |
H 23 |
215 |
85 |
105 |
135 |
H 10 |
130 |
80 |
135 |
115 |
H 24 |
125 |
90 |
270 |
161 |
H 11 |
280 |
90 |
105 |
158 |
H 25 |
110 |
75 |
250 |
145 |
H 12 |
170 |
120 |
160 |
150 |
H 26 |
95 |
80 |
275 |
150 |
H 13 |
125 |
90 |
130 |
115 |
H 27 |
80 |
75 |
230 |
128 |
H 14 |
115 |
105 |
150 |
123 |
Table 3. Temperature measurements by sample number.
Sample |
T1 © |
T2 © |
T3 © |
T avg © |
Time (s) |
H 08 |
51 |
56 |
47 |
51.3 |
3 |
H 04 |
75 |
68 |
37 |
60.0 |
3 |
H 22 |
53 |
59 |
35 |
49.0 |
4 |
H 01 |
47 |
38 |
35 |
40.0 |
3 |
H 25 |
68 |
56 |
38 |
54.0 |
5 |
H 23 |
89 |
76 |
45 |
70.0 |
3 |
H 19 |
52 |
56 |
36 |
48.0 |
3 |
H 13 |
60 |
52 |
36 |
49.3 |
5 |
H 18 |
68 |
60 |
37 |
55.0 |
3 |
Results for Experiment 2
This experiment revealed very little correlation between the two parameters. It is possible that the temperature range into which all the samples fall in the wider spectrum is negligible, so resistance may be of little importance for this investigation. There was also a problem in taking the measurements, as R1 was measured first, then R2, then R3. The cooling of the samples was extremely rapid, and although the whole data collection procedure took only a few seconds, the sample may cool substantially in even less time than that.
Experiment 3: Uniformity of E-field within Horn Antenna
Purpose
The purpose of Experiment 3 was to determine how uniformly the E-field is transmitted at the opening of the 20 dB gain horn to determine where the most energy is applied to the test sample.
Setup
Figure 4 illustrates the layout of wooden substrate samples used in this experiment.
Figure 4. Temperature measurement positions on
wooden substrate samples.
Procedure
The following is a copy of the instructions for the experimental procedure:
1. Using the same horn and cage setup used in Experiment 2, place an untreated piece of wood in front of the horn and trace the outline of its opening on the board.
2. Heat on maximum power until smoke or other signs of sufficient heat appear.
3. Measure each box evenly, record, then measure and record again.
4. Repeat for different temperatures.
Data
Table 4 summarizes the data collected in this experiment.
Table 4. Temperature measurement data for wooden substrates.
TRIAL | |||||||
Position |
T1 © |
T2 © |
T avg © |
T1 © |
T2 © |
T avg © | |
1 |
37 |
35 |
36 |
35 |
34 |
34.5 | |
2 |
66 |
64 |
65 |
56 |
55 |
55.5 | |
3 |
35 |
34 |
34.5 |
33 |
34 |
33.5 | |
4 |
66 |
60 |
58 |
38 |
38 |
38 | |
5 |
89 |
87 |
88 |
77 |
75 |
76 | |
6 |
53 |
46 |
49.5 |
37 |
35 |
36 | |
7 |
56 |
50 |
53 |
40 |
40 |
40 | |
8 |
74 |
71 |
72.5 |
65 |
65 |
65 | |
9 |
40 |
40 |
40 |
38 |
39 |
38.5 | |
Results for Experiment 3
Figure 5 shows a plot of data from the experiment to help visualize the output energy within the horn antenna.
Figure 5. Energy output to test sample through 20 dB gain horn antenna.
The measurements collected in Experiment 3 indicate that the samples received the most energy in the center while output to the corners dropped off substantially to as little as 40 percent of the energy being produced in the center. This finding is important for developing a hand-held horn applicator, because a device with these characteristics is ideal for sweeping a large surface.
Experiment 4: Modified Horn Configuration
Purpose
The purpose of Experiment 4 was to modify the horn antenna for operation outside of the steel cage and test it to simulate the eventual application on a veritable paint/susceptor-covered wall.
Setup
The setup for this experiment is illustrated in Figure 6.
Figure 6. Modified horn microwave system.
Procedure
The following is a copy of the procedure followed for Experiment 4:
1. Place sample in frame and fix edges of horn flush with surface.
2. At 300 W, vary cycling times and repetitions, recording temperatures after each cycle and time per cycle, as well as observations.
3. Repeat at 450 W and 600 W.
Data
The data gathered from these tests are presented in Table 5.
Sample |
Power (W) |
Cycles |
Max Temp
|
Scraping
|
Observations |
A17 |
300 |
4, 4s |
114 |
7 |
Scraped fair, did not burn |
H11 |
300 |
4, 5s |
131 |
9 |
Scraped well, bubbled but no burn |
H20 |
300 |
3, 10s |
112 |
7 |
Difficult to scrape, bubbles but no burn |
H21 |
300 |
3, 12s |
121 |
12 |
Scraped well, bubbled and no burn |
H27 |
300 |
5, 4s |
136 |
14 |
Scraped very well, some bubble, no burn |
H09 |
450 |
2, 15s |
140 |
15 |
Scraped well, but smoked and burnt |
H10 |
450 |
4, 10s |
153 |
10 |
Scraped well, but smoked and burnt |
H17 |
450 |
4, 5s |
130 |
14 |
Scraped well, a little smoke but no marks |
A22 |
600 |
1, 12s |
112 |
6 |
Did not scrape, burnt and smoked |
H24 |
600 |
3, 4s |
135 |
15 |
Scraped well, a little smoke and bubbles |
H26 |
600 |
4, 3s |
130 |
17 |
Scraped very well, no marks, bubbled well, a little smoke |
*Data in the CYCLES column is presented in the following format: number of cycles separated by a 2-second delay, exposure time of each cycle.
Results of Experiment 4
The results of this experiment indicated that higher powers drastically increase the chances of damaging the substrate by burning even though they reduce the time required for paint removal. However, the most important result is that removal of paint is possible using such a device at 300 W, 450 W, and 600 W output.
Experiment 5: Graphite Susceptor Testing
Purpose
This experiment was intended to determine the effects of microwave exposure on samples coated with a graphite susceptor.
Part A Procedure
The following steps were used:
1. Coat samples with Slip Plate No. 3, a dry film graphite-based lubricant, and allow to dry (several hours).
2. Measure initial temperature of sample, Ti.
3. Heat in microwave oven on high power for indicated time; measure final temperature, Tf.
4. Immediately scrape sample.
Part A Data
Sample 65A
Exposure: 3 sec
Ti: 20 °C
Tf: 20 °C
Comments: No effects were observed and the sample was not scraped.
Sample 63A
Exposure: 6 sec
Ti: 20 °C
Tf: 88 °C
Comments: A thin wisp of smoke was observed. The surface of the sample bubbled in a region approximately 1 inch in diameter. The paint was very difficult to remove; only a tiny area could be removed using the corner of the scraper. The sample appeared to cool rapidly.
Sample 62A
Exposure: 9 sec
Ti: 20 °C
Tf: 174 °C
Comments: A large flame was observed during microwave exposure. The surface was bubbled and blistered with slight discoloration. Paint was removed from about 50 percent of the sample. The paint was soft but still very difficult to remove. A white residue remained on the wood even after scraping. In the areas where the wood was charred, no residue remained.
Sample 65A
Exposure: 12 sec
Ti: 20 °C
Tf: 130 °C
Comments: No fire was observed. Small uniform bubbling was evident across the surface of the sample. A large area of paint was removed (about 60 percent). The paint scrapings were removed in large pieces. A white residue remained but there was no damage to the wood.
Part B Procedure
The following steps were used:
1. Coat samples with Slip Plate No. 3 using an airless power sprayer and allow to dry.
2. Measure initial temperature, Ti, and three resistance values, R.
3. Heat in microwave oven on high power for indicated time; measure final temperature, Tf.
4. Immediately scrape sample.
Part B Data
Sample 62A R1: ~520W R2: ~340W R3: ~303W
Exposure: 3 sec, R avg.: 388W
Ti: 20 °C
Tf: 20 °C
Comments: No observable surface changes were noted.
Sample 69A R1: ~425W R2: ~490W R3: ~430W
Exposure: 6 sec, R avg: ~448W
Ti: 20 °C
Tf: 150 °C
Comments: Sample smoked, but did not burn. Surface bubbled. Bubbles concentrated on 2/3 of sample with one large blister at the end of the sample at R3. It was observed that the bubbles consisted of graphite only; the paint did not bubble up. The paint was soft but difficult to remove. About 15 percent of the sample was scraped to the bare wood. Most of the paint was removed from areas that had a high degree of bubbling. Some of the graphite was removed from other areas. No white residue was left, but the wood was gouged by the scraper.
Sample 57A R1: ~330W R2: ~400W R3: ~455W
Exposure: 9 sec, R avg: ~395W
Ti: 20 °C
Tf: 109 °C
Comments: Smoke was observed during microwave exposure; a small spot fire occurred at the very end of exposure between R2 and R3. The graphite was discolored in the location of the fire. The surface was uniformly bubbled. The graphite was removed in small areas over 50 percent of the sample. The paint was difficult to remove. The paint was stripped to the wood in two thin stripes below the burned area. The wood was charred.
Sample 55A R1: ~305W R2: ~385W R3: ~350W
Exposure: 12 sec, R avg: 347W
Ti: 20 °C
Tf: 118 °C
Comments: A large fire engulfed the sample after 10 seconds, and the microwave device was turned off. Microwave restarted, but it is likely that no more power was absorbed due to a 2-second delay in microwave generation. The surface had numerous small bubbles and one large blister. Once again, the blister was graphite only; the paint was softened but did not bubble. The paint was difficult to remove. About 15 percent of paint was removed. A yellow-white residue was left behind. Some fire damage was revealed at the corner and one end of the sample.
Results for Experiment 5
The results for this experiment are included and interpreted with the results for Experiment 6 (next section).
Experiment 6: Optimization of Hand-Held Horn Generator System
Purpose
The purpose of this experiment was to find an optimal method of applying micro-wave power to graphite-coated painted wood surfaces using the prototype equipment.
Setup
Figure 7 shows a diagram of the prototype portable paint-removal system.

Figure 7. CERL portable microwave paint removal system.
Procedure
The following text is a copy of the procedure followed.
1. Coat the large painted samples (28.5 cm X 35 cm) with the graphite susceptor material (see Figure 8).
2. Test different methods and time exposures of application.
3. Record the method used, the duration of exposure, and the results (including final temperature and ease of paint removal).
Figure 8. Wood substrate painted with lead-based paint and
coated with susceptor.
Data
Table 6 presents the measurements and data collected for Experiment 6.
Sample |
Method |
Result |
L01A |
60 sec continuous motion, then 20 sec static in each 1/3, scraping after each 20 sec period. |
2 small burns on an overlap, temperatures to 130 °C. Scraped very well. |
L01B |
30 sec static in each 1/3, two times through scraping after entire process. |
No burns, but spots between overlaps did not get hot enough to scrape. Not a good system. |
L02A |
60 sec continuous motion, scraping large surface after exposure. |
Poor scraping, minimal part of center heated sufficiently. Not a good method. |
L02B |
45 sec continuous, then 20 sec static on each 1/3, scraping after each 20 sec period. |
More wood than paint came up, very difficult to scrape. Not a good method. |
L03A |
30 static seconds in each 1/3 of sample, scraping after each 1/3 exposed. |
Good scraping, small burns on overlaps. Scraped easily and completely. |
L03B |
20 static seconds in each 1/3 of sample, scraping after each 1/3 exposed. |
Not enough exposure. Overlaps scraped well, but other parts did not scrape. |
L04A |
60 sec continuous motion, then 15 static seconds in each 1/3, scraping after each 15 sec period. |
Did not remove exceptionally easily, but removed completely. Burns on overlaps. |
L04B |
20 sec on, 5 sec off, 20 sec on to each 1/3, scraping after process completed to 1/3. |
After the second heating, small portions were removed, while overlaps again burned. |
L05A |
Heat gun for 6 minutes. |
Very long, high temperatures, but still poor scraping. |
L05B |
60 sec continuous motion, then 30 sec on top 1/3, 30 sec on remaining 2/3 (motion), scrape after each 1/3. |
Complete removal, good temperatures, difficult scraping, burn on overlap. |
L06A |
120 sec continuous motion over entire sample, scraping entirety after exposure. |
Scraped easily and completely, but not enough room to test applicator maneuverability. No burning. |
L06B |
90 sec continuous motion over entire sample, scraping entirety after exposure. |
Not quite as high temperatures, but easily scrapable. Again, need more painted surface to truly test system. |
L07A |
120 second continuous motion, scraping after entire 2 minutes. |
Exceptional removal. No discoloration, no burning, easy scraping, came off in large pieces. |
L07B |
150 second continuous motion, scraping after entire exposure. |
Even better. No discoloration, burns, very easy scraping, large pieces. |
L08A |
135 second continuous motion, scraping after entire exposure. |
A little more difficult to scrape, but still easy. Total removal, no discoloration, no burns, large pieces. |
L08B |
120 seconds continuous motion, scraping after entire exposure. |
This seems to be the target time. Again, very easy scraping, no burns, no discoloration, large pieces. |
L09A |
105 seconds continuous motion, scraping after entire exposure. |
Possibly a little too short. Although total removal, no burns or discoloration, scraping more difficult. |
L09B |
105 seconds continuous motion, scraping after entire exposure. |
Better than L09A, but still more difficult than full 2 minute exposure time. |
L10A |
120 seconds continuous motion, scraping after entire exposure. |
Again, the best method found with these samples and the working setup. |
L10B |
120 seconds continuous motion, scraping after entire exposure. |
Complete removal, no burns, no discoloration, large pieces, ideal temperatures. |
Results for Experiments 5 and 6
The results of Experiments 5 and 6 indicate that graphite-based susceptors are suitable for coupling microwave energy to underlying paint. These experiments also indicate that the most consistent, complete, and best method of applying the microwave power is using 2 minutes of continuous motion, moving back and forth across approximately a 1 foot by 5 inch section and scraping after the full exposure. Average temperatures reached about 120 °C consistently, with temperatures as high as 138 °C and as low as 98 °C on the samples tested with this method. These experiments were quite successful in finding a way to remove paint completely and evenly without burning or discoloring the substrate.
Experiment 7: Minimum Scraping Temperature
Purpose
The purpose of Experiment 7 was to determine the minimum temperature that the paint must reach in order to be scraped from a wooden substrate.
Setup
Because temperature was the only variable being tested in this experiment, a standard lab-grade convection oven was used for optimal temperature uniformity and control within the test environment. All temperatures were measured with a standard lab-grade pyrometer.
Procedure
The steps presented below were followed:
1. Heat oven to 100 °C, then place three painted samples (no susceptor) in it.
2. Allow to cook until temperature of wood is equal to the temperature of the oven (100 °C).
3. Remove samples and attempt to scrape; record data.
4. If scrapable, repeat procedure at a temperature 5 °C below the previous recorded data until scraping is not possible.
Data
Table 7 presents the data recorded for this experiment.
Table 7. Data for determining minimum scraping temperature.
Sample # |
Temperature (°C) |
Observations |
U01 |
100 |
Scraping was very easy; similar to the best microwaved samples. The paint came off cleanly and in large portions. Good scrapability at this temperature. |
U02 |
100 |
|
U03 |
100 |
|
U04 |
90 |
Scraping again was extremely easy. The paint was removed with little difficulty leaving a very thin coating of primer and came off in large pieces again. |
U05 |
90 |
|
U06 |
90 |
|
U07 |
80 |
Scraping was possible, but not as easy as the previous two temperatures. The paint did not come off in as large of pieces, but still scraped clean. |
U08 |
80 |
|
U09 |
80 |
|
U10 |
70 |
Scraping not feasible at this temperature. Difficult and did not come off evenly or completely. Scraping possible, but not easy. Completion achieved. |
U11 |
70 |
|
U12 |
75 |
Results for Experiment 7
The results for this experiment are included and interpreted with the results for Experiment 8 (next section).
Experiment 8: Further Experimentation With Fully Coated Samples
Purpose
The purpose of this experiment was to use the system as developed thus far to remove paint from the larger samples fully coated with paint and graphite susceptor.
Setup
Figure 9 illustrates the frame setup used to hold the horn in Experiment 8.
Figure 9. Sample holding frame for laboratory testing of microwave applicator.
Procedure
The steps followed in this experiment were as follows:
1. Using the large, fully coated samples (28.5 cm x 35 cm), place a full sample in the frame and reflector/absorber setup.
2. Use different methods and exposure durations.
3. Record method, time, and observations.
Data
Table 8 presents the data collected for Experiment 8.
Table 8. Data for fully coated samples.
Sample |
Method |
Result |
F01 |
3 minutes continuous motion over entire surface. |
Approximately 2/3 paint removed, some wood with it. No burns, clean. |
F02 |
2 minutes over top ½, scrape, then 2 minutes over bottom ½, scrape again. |
100% Removal. No burns, clean removal, large pieces. Very good. |
F03 |
4 minutes continuous motion over entire surface. |
90% Removal. Difficult to scrape, some wood with it. No burns, clean. |
F04 |
1½ minutes over top ½, scrape, then 1½ minutes bottom, scrape. |
100% Removal with a very small portion of wood. No burns, good. |
F05 |
1:45 over top ½, scrape, 1:45 over bottom ½, scrape again. |
100% Removal, no burns, discoloration, large pieces, very good. |
Results for Experiments 7 and 8
The results of Experiments 7 and 8 indicate that 75 °C is the absolute minimum temperature at which the paint can be removed, but ideally a temperature of 80 °C should be attained. These temperatures are helpful in determining when to pull the applicator off of the paint and when to begin scraping. Furthermore, the experiments reveal that 1 sq ft of susceptor-coated paint can be removed in approximately 2½ to 3 minutes. The paint is removed easily in large pieces with a hand-held scraper.
Laboratory experiments using the microwave oven revealed that the amount of paint removed increases with increasing microwave exposure time. However, increased microwave exposure time, in general, increases the chances of a spot fire and damage to the substrate. It was found that paint can be removed more easily and in larger quantities by maintaining the sample at a relatively high temperature while avoiding the peak temperatures that resulted in combustion. This result can be accomplished by cycling the power on and off.
Experiment 9: Leakage from CERL Microwave Applicator
Purpose
The purpose of this experiment was to determine the levels of microwave leakage to which the user may be exposed to while using the CERL system.
Setup
A rubber shielding skirt, designed by HVS Technologies, Inc., was installed on the applicator, as shown in Figure 10. The figure also illustrates where leakage measurements were taken on the microwave applicator device. The measurements were made with a Simpson 380-2 Microwave Leakage Detector.
Figure 10. Diagram of shielding skirt and microwave energy leakage measurement positions.
Procedure
The following procedure was followed for Experiment 9:
1. Set applicator down on setup used in previous experiment (see Figure 9), but use an uncoated sample.
2. Start the system at a time sufficient to take measurements at strategic locations using the leakage detector.
3. Record results.
Data
Table 9 lists the results of the microwave leakage tests with the shielding skirt in place.
Table 9. Results of microwave leakage measurements.
Location # |
Leakage (mW/cm3) |
Location # |
Leakage (mW/cm3) |
Location # |
Leakage (mW/cm3) |
Location # |
Leakage (mW/cm3) |
1 |
<1 |
6 |
1 |
11 |
<1 |
16 |
<1 |
2 |
<1 |
7 |
1 |
12 |
<1 |
17 |
<1 |
3 |
1 |
8 |
<1 |
13 |
<1 |
18 |
4 |
4 |
<1 |
9 |
1 |
14 |
<1 |
19 |
<1 |
5 |
<1 |
10 |
<1 |
15 |
<1 |
20 |
2 |
Results for Experiment 9
This experiment revealed that the leakage from the portable microwave applicator with shielding skirt falls within safe exposure levels as specified in IEEE C95.1 (5 mW/cm3 at a range of 5 cm). The final CERL prototype device further protects the user from microwave exposure by incorporating safety switches that prevent the unit from operating unless it is held steadily and in full contact with the substrate.
Optimization of Susceptor Materials
The ability of a material to absorb microwave energy is related to the electrical conductivity (s) through the loss tangent, tan d = s/(we), where w is the frequency and e is the permittivity or dielectric constant. Using the loss tangent, the microwave power absorption can be calculated for a given electrical conductivity or resistivity. Table 10 lists the dielectric properties of the experimental susceptor materials. Initial work used graphite and graphite-based compounds as the susceptor material.
Table 11 lists the electrical resistivity of the selected graphite susceptors.
Table 10. Dielectric properties of susceptors.
Substrate |
Dielectric Constant |
Loss Tangent |
Conductivity (S/cm) |
Frequency
|
SiC |
107 |
0.686 |
0.041 |
1.0 |
Carborundum |
60 |
0.580 |
0.058 |
3.9 |
Graphite |
5 x 107 |
7.0 x 104 |
2.45 |
Table 11. Properties of selected graphite materials.
Coating |
Vehicle |
Consistency |
Texture |
Resistivity (ohm-cm) |
Sq. R (ohm/sq.) |
Slip plate No. 4* |
Water |
Medium |
Even |
0.4709 |
43 |
GW 220 |
Water |
Watery |
Flaky |
66.2 |
9.33 |
GW 330 |
Water |
Medium |
Even |
0.641 |
157.67 |
GW 430 |
Water |
Thick |
Brush Strokes |
0.89 |
113.167 |
Slip Plate No. 3* |
Oil |
Medium |
Even |
1.150 |
491.167 |
Slip Plate No. 1* |
Oil |
Thick |
Brush Strokes |
4.572 |
864.33 |
Slip Plate Spray* |
Oil |
Spray |
Even |
5.802 |
518.33 |
* Superior Graphite Co.
Dixon Ticondaroga Co.
The graphite susceptor only arcs after it has been heated to a high temperature. One reason for this is that under a strong electric field, as a material gets hotter it will begin to boil off electrons. These electrons, under this strong electric field, can then initiate an arc across the material's surface. Hot spots of a very small size may contribute greatly to this undesirable effect. Hot spots can be caused by minuscule inhomogeneities in the susceptor or by field concentrations at the applicator-susceptor interface.
Some methods for reducing the occurrence of arcing include (1) increasing the work function of the surface and (2) reducing the field strength. Reducing the field strength can be accomplished a number of ways:
· The power of the magnetron can be reduced (but only at the expense of reducing the overall rate of paint stripping).
· The power density can be reduced by increasing the size of the aperture and heating a larger area.
· The time-average power density can be lowered by repeatedly scanning a small aperture over a larger area until the whole area is hot enough to scrape away the paint.
These three methods for lowering the effective power field strength while preheating the substrate were investigated.
Another way to possibly address this issue would be to slightly modify the applicator to prevent field concentrations that may initiate hot spots or small arcs that lead to larger burning arcs. This line of investigation was not pursued in the current study, however.
Due to arcing and burning of some samples using the graphite susceptor, the use of a self-regulating susceptor material was investigated. Such a material would absorb energy well within the desired temperature range but would become less absorbing at higher temperatures, thus preventing burns. Polyaniline is a conducting polymer that exhibits this behavior within approximately the desired temperature range. It can be used to produce a susceptor coating with a direct current (DC) resistivity in the range 50 to 500 W. Its resistivity increases with temperature, and becomes virtually insulating at 180 °C. To measure this property, polyaniline was coated onto glass with two electrodes positioned 1 inch apart. These leads were connected to an ohm meter, and the specimen was placed in a lab-grade convection oven. Figure 11 and 12 show the measured resistance-versus-temperature curves. This behavior is irreversible in that the polyaniline loses its conductivity permanently when heated to the applicable temperatures.
After the polyaniline is applied to the substrate it must be "activated" by rinsing with ethanol before microwave energy is applied. This activation rinse raises the DC conductivity into a functional range for energy absorption. The mechanism by which this activation step works is not fully understood. It appears that rinsing with ethanol removes a top insulating layer from the material, but this does not really account for the activation phenomenon. It was considered that the underlying polyaniline might be conductive before activation but that the top insulating layer prevented DC conductivity measurement. However, the microwave absorption of unactivated polyaniline was found to be negligible, indicating that the underlying polyaniline is nonconductive until activated with ethanol. To verify the irreversibility of the conductivity loss when the polyaniline was heated above 180 °C, the glass sample used in this experiment was again rinsed with ethanol after heating, but no change in the resistivity of the sample was detected after rinsing.
Figure 11. Resistance of a polyaniline film versus temperature.
Figure 12. Resistance of a polyaniline film versus temperature showing rapid transition from conductor to insulator above 180 °C.
Experiments Using Chemical Stabilizers with Susceptor Materials
Further laboratory experiments were conducted in order to determine if chemical stabilizers applied in conjunction with susceptors would render nonhazardous paint scrapings without affecting the ease of removal. The susceptors tested were mixtures of graphite and glue, and mixtures of graphite and Slip Plate #3. The chemical stabilizers tested were Lead-XÒ and PreTox 2000Ò. Experiments were specifically designed to determine the optimum scheme for mixing or layering the susceptors with the stabilizers.
Initial screening of susceptor candidates was accomplished by measuring resistances of the applied susceptors after they were allowed to dry. The resistances of both brush-applied and spray-applied susceptors were determined in this manner. It was found that the spray-applied susceptors had resistances approximately one order of magnitude lower than the brush-applied susceptors. The graphite/glue mixtures had extremely high resistances, and thus did not yield suitable susceptors.
As in previous experiments, the susceptor/stabilizer combinations chosen from the initial screening were applied to samples of lead-based paint on glass substrates, which were subsequently heated in a microwave oven at 1000 watts for 3 minutes in cycles of 5 seconds on / 5 seconds off, and then removed and scraped. The on/off duty cycle simulates microwave heating that would be acheived using the CERL portable microwave applicator. These experiments were performed in order to ascertain the ability of the susceptor/stabilizer combination to yield nonhazardous waste when subjected to microwave heating. Glass substrates were chosen because the paint can be completely removed from the glass substrates very easily. The resulting paint waste scrapings were analyzed using the EPA toxicity characteristic leaching procedure (TCLP) in order to determine if the waste was nonhazardous. In order to be classified as nonhazardous, the waste must leach less than 5 ppm lead as determined by TCLP.
Experiment 10: Microwave Oven Test of Susceptors and Stabilizers
Specifications
1. Paint layer
· In-house lead-based paint.
2. Lead-abatement layer
· PreTox2000 (samples PT1-PT3)
· Slip Plate No. 3 (samples L1-L3)
· Slip Plate No. 3 and Lead-X (samples L4-L6).
3. Susceptor layer
· Slip Plate No. 3 graphite.
Procedure
1. Prepare lead abatement layer
· Prepare 15 percent by weight mixture of Slip Plate No. 3 and Lead-X.
2. Prepare samples
· Paint nine glass plates with lead-based paint.
· Spray three of the lead-painted panels Slip Plate No. 3.
· Spray three of the lead-painted panels with the Slip Plate No. 3 graphite and Lead-X mixture.
· Coat three of the lead-painted panels with PreTox2000 using a brush.
· Spray the three PreTox2000-coated panels with Slip Plate No. 3 graphite.
3. Heat samples and scrape
· Place a sample in a microwave oven.
· Heat the sample for 3 minutes in on/off cycles of 5 seconds.
· Scrape the paint from the panels (be sure to collect all the scrapings).
· Grind up the scrapings and send them out for TCLP testing.
Data
Tables 12 and 13 list parameters and data collected in Experiment 10.
Several promising candidates for susceptor/stabilizer systems emerged from the microwave oven heating experiments. These susceptor/stabilizer systems were applied to lead-based paint samples on wooden substrates, and then heated using the CERL portable microwave applicator. After being heated with the microwave applicator, the samples were scraped as before, and the scrapings (lead-based paint, susceptor, and stabilizers) were analyzed in accordance with TCLP.
Table 12. Microwave-assisted paint removal laboratory experiments with PreTox2000 as the stabilizer and graphite as the susceptor using microwave oven and glass substrates.
Sample |
Gl |
Gl+Pb |
Gl+Pb+PT2 |
Gl+Pb+PT2+Gr |
TCLP Results |
PT1 |
1011.86 g |
1027.79 g |
1035.40 g |
1039.11 g |
<0.06 PPM |
PT2 |
1014.10 g |
1028.72 g |
1036.54 g |
1040.31 g |
<0.06 PPM |
PT3 |
1016.53 g |
1027.53 g |
1068.55 g |
1072.10 g |
<0.06 PPM |
Gl=glass |
Pb=LBP |
Gr=graphite |
PT2=PreTox2000 |
Table 13. Microwave-assisted paint removal laboratory experiments with Lead-X as the stabilizer and graphite as the susceptor using microwave oven and glass substrates.
Sample |
Gl |
Gl+Pb |
Gl+Pb+Gr |
Gl+Pb+Gr/Ldx |
TCLP Results |
L1 |
1011.12 g |
1019.41 g |
1021.85 g |
NA |
350 PPM |
L2 |
1028.68 g |
1038.04 g |
1040.27 g |
NA |
260 PPM |
L3 |
1006.95 g |
1016.23 g |
1018.84 g |
NA |
300 PPM |
L4 |
1008.22 g |
1017.31 g |
NA |
1010.27 g |
300 PPM |
L5 |
999.51 g |
1008.55 g |
NA |
1011.53 g |
320 PPM |
L6 |
1017.16 g |
1025.86 g |
NA |
1029.29 g |
380 PPM |
Gl=glass |
Pb=LBP |
Gr=graphite |
Ldx=Lead-X |
Notes: Samples L1-L3 have a susceptor coating of approximately 2 mils. Samples L4-L6 have a susceptor coating of approximately 1 mil. The first four columns of each table reflect data taken to keep track of mass of coatings on glass panels.
Experiment 11: Test of CERL Microwave Applicator Using Both Susceptors and Stabilizers
Specifications
1. Paint layer
· In-house lead-based paint
2. Lead-abatement layer
· PreTox2000
3. Susceptor layer
· Slip Plate No. 3 graphite and graphite powder
Procedure
The following procedure was followed in Experiment 11:
1. Prepare samples.
· Paint four wooden panels with lead-based paint (brush on).
2. Prepare susceptor layer.
· Find the mass of the volume of Slip Plate No. 3 to be used
· Add graphite powder until the total mass of the mixture is 17.4 percent powder and 83.6 percent Slip Plate No. 3
3. Apply lead abatement layer (brush on).
4. Apply susceptor layer (airless sprayer).
5. Perform microwave-assisted paint removal process and send waste out for TCLP.
Data
Table 14 summarizes the data collected for Experiment 11.
Table 14. Microwave-assisted paint removal tests using PreTox2000 and Slip Plate No. 3/graphite powder on painted wood substrate.
Sample |
Temp © |
Resistance (ohms) | |||
Left |
Center |
Right |
TCLP Results | ||
F2A |
90 128 |
300 |
210 |
260 |
<0.06 PPM |
F2B |
30 110 |
500 |
200 |
320 |
<0.06 PPM |
F3A |
70 150 |
390 |
550 |
320 |
0.08 PPM |
F3B |
60 120 |
400 |
320 |
400 |
0.08 PPM |
F4A |
35 210 |
350 |
235 |
320 |
<0.06 PPM |
F4B |
80 220 |
310 |
300 |
520 |
<0.06 PPM |
Comments
F2A: Heated very evenly. Substrate was very warm, and initial scraping damaged it a little. Everything came off. There doesnt appear to be any lead paint remaining on the surface. Smoke appeared near the end of 2 minute application.
F2B: Heated fairly unevenly. Initially cooler side didnt scrape well, but eventually all scraped very well. Everything came off, but some residue of lead paint visible in board. Board heated also, and smoke appeared near end of 2 minutes.
F3A: Scraped well. Some discoloration in the center of board where it was hottest. Very little wood scraped up in an area with approximate dimensions of 1 in. by 0.5 in.
F3B: Scraped extremely well. Some discoloration in center of board where it was hottest.
F4B: Surface became extremely hot. Area where graphite bubbled was difficult to scrape. All paint scraped. The substrate suffered some heat damage. Scraping with the grain seems to work best.
F4A: Same as F4B, except it seemed that by first scraping against grain and then scraping with the grain, the results were achieved more quickly. Burn spots occurred in regions where the surface temperature was greater then 190 °C after heating.
Summary of Results
The combined results for Experiments 10 and 11 show that the optimal combination of easy removal and nonhazardous waste product was achieved when PreTox2000 was chosen as the stabilizer, and a mixture of Slip Plate No. 3 and graphite powder was chosen as the susceptor. Optimal conditions were achieved when:
· The PreTox was brush-applied directly over the lead-based paint and allowed to dry for 24 hours.
· Slip Plate No. 3 and graphite powder were mixed together in a weight ratio of 5 parts. Slip Plate No. 3 to 1 part graphite to form the susceptor, and then spray-applied as a separate layer over the PreTox as shown in Figure 13.
· The susceptor topcoat was allowed to dry for 24 hours.
Using this scheme, it was determined that 0.5 sq ft of lead-based paint could be removed in about 3 minutes. This estimate includes heating times of approximately 2 minutes using the applicator, required to obtain a surface temperature of about 100 °C, and a scraping time of approximately 1 minute.
A system developed by HVS Technologies, State College, PA, under contract to CERL, was designed to provide high power to a relatively small area with the goal of rapidly and uniformly heating that area (Hollinger, Varadan, and Varadan 1996). The HVS system (Figure 14) can provide up to 1000 watts of power. A low-loss coaxial cable connects the generator to a hand-held or robotically operated applicator. In the applicator, a stub tuner is used to match the impedance of the microwave to that of the susceptor applied over the painted substrate. In the HVS applicator, an aperture is used to focus the microwaves onto a small area (2.54 cm by 2.54 cm) of substrate. This aperture is a specially designed ceramic window that couples the microwave energy from the waveguide to the painted surface. The aperture window material was selected to have low dielectric loss, high temperature resistance, excellent thermal shock resistance, and good mechanical strength.
The applicator is equipped with four safety switches that are closed when the applicator is pressed against a flat surface. No microwave energy is produced unless all four safety switches and the activator switch are closed. The activator switch is located on the hand grip. Surrounding the applicator is an electromagnetic shielding skirt made of a conducting polymer. On the base of the applicator is an exhaust tube, which is connected to a vacuum system and a high-efficiency particle air (HEPA) filter. The front panel on the generator contains an indicator light, an on-off switch, a reflected power meter, and a forward power meter. The power delivered to the substrate is the difference between the forward and reflective power. The amount of reflected power can be minimized by impedance matching.
Investigations of susceptor/stabilizer combinations by HVS revealed that lower susceptor resistances could be obtained by mixing 28 wt percent Desulco Graphite Powder 9033 with 43.2 wt percent latex paint and 28.8 wt percent water to make a susceptor that can be brush-applied. When this susceptor was brush-applied over PreTox2000 (which itself was brush-applied over LBP), the average resistance was 90.1 ohms. When the susceptor was brush-applied directly over the LBP, and the PreTox2000 was subsequently brush-applied over the susceptor, the average resistance was found to be 166.7 ohms. Neither of these systems has been subjected to TCLP testing, but because the graphite-based susceptor is similar to that used by CERL and the stabilizer (PreTox2000) is the same, there is good reason to expect that the HVS treatment will yield waste products with TCLP results similar to those obtained in the CERL experiments. Due to the lower surface resistances of the HVS susceptor/stabilizer systems, paint heating time might be reduced.
Field Demonstrations of Microwave-Assisted Paint Removal Technologies
Demonstrations of microwave-assisted paint-removal technology were performed on the abandoned lockmasters house at the Army Corps of Engineers Lock and Dam #6, on the Kentucky River near Lexington. HVS was contracted to conduct the demonstration, so the company used its own system in this phase of the study. The onsite work was conducted on 10, 11, and 12 December 1997. Initially the area chosen to be stripped was the interior surface of an outer wall on the enclosed porch. The substrate appeared to be plywood.
Ten square feet for field testing of the HVS applicator and about five square feet for the CERL microwave removal system were marked off, and susceptors were applied. HVS personnel used two different susceptors, covering 5 sq ft with each. The first was a latex-based graphite susceptor of a refined composition, developed by HVS. The second was an intrinsically conductive polymer (polyaniline) dissolved in the solvents cellosolve and xylene. Both susceptors were applied using paint brushes (although both are also sprayable). These were allowed to dry overnight.
As explained previously, the polyaniline must be activated before it is conductive. To do this, chemical sorbent socks (3 in. diameter by 48 in. long) were placed on the floor against the wall on which polyaniline had been applied, and ethanol was spayed onto the polyaniline from a squirt bottle. The activation with ethanol causes the polyaniline to shrink slightly which results in mud cracking. Resistivity measurements of the susceptors were taken as shown in Figure 15.
Figure 15. Resistivities (ohms) of the susceptors coated on the interior porch wall, measured using HVS probe.
The HVS microwave paint stripper was used first on the HVS graphite susceptor. However, when attempting to scrape off the paint after heating, the scraper dug into and tore up the substrate. It was discovered that the substrate in this area was not solid wood as expected, but was compressed paper board. Several attempts were made to scrape both the graphite and the polyaniline-treated areas with the same result. The decision was made to look for substrate specimens known to be solid wood. A wooden door and door frame (Figures 16 and 17) were chosen for HVS to strip. The CERL applicator was demonstrated on a window frame. Both susceptors (graphite and polyaniline) were applied and allowed to dry overnight.
The graphite susceptor on the door had a higher resistance than it did on the porch wall (compare Figures 15 and 16). The overall average on the wall was 266 W while on the door it was more than double that (646 W). This was probably because the susceptor in the can had partially dried and formed a gel that was too thick to apply. Water was added, but this made it thinner than desirable. This overthinning apparently impaired the susceptor's conductivity, thus resulting in higher resistivity measurements. The gelling of the susceptor may have been caused by the cold ambient temperatures, which averaged from the mid 30s to low 40s (degrees Fahrenheit) during the test dates. Friday was the coldest day.
Despite the high resistivities explained above, it was possible to strip the paint using the HVS microwave paint stripper. However, more time was needed to heat the paint, and the period of time after heating during which it was possible to scrape the paint was very short on the order of 5 seconds. The slow heating can be attributed both to the cold weather and to the high resistivity of the susceptor. The short period of time available to remove the heated paint was due solely to the cold weather. Typically an area 6 in. long by 1.5 to 1.75 in. wide was heated for about 30 seconds, then scraped immediately with a 1.25 in. wide putty knife. Both the multiple-scan method developed by CERL and the single-scan method were used. It was believed that the multiple scanning method might work better to gradually heat the substrate as well as the paint. However, the single-scan method seemed to work better and was certainly less tiring for the operator. Arcing of the graphite was not a problem with the N131C graphite, even during prolonged heating of the same spot. With prolonged heating the paint itself would melt enough to start absorbing microwave power and start smoking. In this condition the paint would almost fall off of the wood, but the time required to soften the paint this way was too long to be efficient.

The polyaniline susceptor exhibited low resistivities and worked well during microwave paint stripping, but as with the graphite-treated samples, the cold weather caused slow heating rates and very short paint scraping time limits. One drawback to the polyaniline susceptor is that it irreversibly loses its conductivity as it is heated. Therefore, once it is heated, if all of the heated paint is not scraped off, the remaining susceptor is very difficult to reheat. Even though the microwave applicators produce a very uniform field, the energy output does taper off some at the outer edges. Furthermore, the paint near the edges of the heated spot cools more rapidly because it is adjacent to unheated paint and substrate, and these act as heat sinks. Another complicating factor is that during scraping, some of the susceptor is removed at the edges without removing all of the paint under it. Because of this phenomenon, an estimated 50 percent or more of the scraping time was of necessity dedicated to removing the paint adjacent to previously stripped areas. The center portion of the heated area came off in one scrape, but near the edges of the area the paint was not nearly as soft and required much more effort to remove (Figure 18).

Figure 18. Illustration of the difficult-to-strip boundary between stripped and unstripped areas of polyaniline-treated areas.
A few cases of arcing and burning were noted. All such cases occurred with the graphite susceptor. In the laboratory experiments arcing was associated with low susceptor resistivities, but in this field test the graphite susceptor had a higher-than-desired resistivity due to the gelling and thinning problem noted previously. Therefore, it is likely that the arcing was caused by the presence of small nails or brads that were visually inconspicuous.
It is worth noting, however, that there were two burn marks in the graphite-treated area that did not appear to be caused by metal; both appeared to be related to discontinuities in the paint. In one case the paint had been chipped down to bare wood, and both the bare wood and the paint surrounding it were coated with susceptor. In the other case, there was significant burning right at the paint/no paint interface. In this case, the paint was cracked and the edges appeared to be curled up. Two possible explanations for this are as follows:
1. The paint on the door was quite thick, so in both cases a crevice existed that could have retained excess susceptor that initiated burning during the paint removal process.
2. The irregularity of the surface may have produced electric field concentrations that initiated arcing and burning.
It is interesting to note that no arcing or burning problems occurred with the polyaniline susceptor even though its resistivity in this test was only about 50 percent compared to the graphite-coated samples. It may be that the polyaniline coating was of very uniform resistivity, but a simpler explanation would be that the sample was free of nails or brads of the right size and orientation to cause arcing.

Figure 19 shows two views of the door, before and after LBP removal.
It is concluded that microwave-assisted removal of LBP from wood was successfully demonstrated in these tests. Lead levels on the wood were dramatically reduced on the areas stripped. The new type of susceptor polyaniline was used effectively to heat the paint without any arcing or burning. This is a promising material, but at a current cost of approximately $1500 per liter it is still too expensive for widespread use in routine LBP removal projects.
The graphite susceptor exhibited very little arcing and burning. Those few examples of this problem appear to have been caused by the presence of metal in the substrate and pre-existing physical damage to the paint that exposed the bare substrate to the susceptor. The time taken to strip the 10 sq ft was approximately 5 hours, or a rate of 2 sq ft per hour. The cold weather made it necessary to heat the samples longer than was necessary at ambient temperatures in the laboratory, but it also required the technician to scrape the softened paint more rapidly to prevent re-solidification before removal. Therefore, the overall time required for the process in the field was approximately the same as that achieved in the laboratory.
The paint at the field site seemed to be more tenacious than expected, but this may have been due to the cold temperatures and the resulting inability to reach the same paint temperatures that would be reached in a warmer environment.
This work has resulted in the following publications and patents:
Boy, J., and A. Kumar, Lead-Based Paint Hazard Mitigation in The Encyclopedia of Environmental Analysis and Remediation, Robert A. Meyers, ed. (John Wiley and Sons, Inc., 1998), pp 2501-2516.
Kumar, A., and J. Boy, Microwave-Assisted Removal of Lead-Based Paint from Wooden Structures, Proceedings, Army Science Conference (1998).
U.S. Patent No. 5,268,548, Microwave-Assisted Paint Stripping, Ashok Kumar (7 December 1993).