Environ. Sci. Technol. 2011, 45, 738–743
Control of Electrothermal Heating during Regeneration of Activated Carbon Fiber Cloth DAVID L. JOHNSEN, KAITLIN E. MALLOUK, AND MARK J. ROOD* Department of Civil and Environmental Engineering, University of Illinois, 205 North Mathews Avenue, Urbana, Illinois 61801, United States
Received September 29, 2010. Revised manuscript received November 17, 2010. Accepted November 29, 2010.
Electrothermal swing adsorption (ESA) of organic gases generated by industrial processes can reduce atmospheric emissions and allow for reuse of recovered product. Desorption energy efficiency can be improved through control of adsorbent heating,allowingforcost-effectiveseparationandconcentration of these gases for reuse. ESA experiments with an air stream containing 2000 ppmv isobutane and activated carbon fiber cloth (ACFC) were performed to evaluate regeneration energy consumption. Control logic based on temperature feedback achieved select temperature and power profiles during regeneration cycles while maintaining the ACFC’s mean regeneration temperature (200 °C). Energy requirements for regeneration were independent of differences in temperature/ power oscillations (1186-1237 kJ/mol of isobutane). ACFC was also heated to a ramped set-point, and the average absolute error between the actual and set-point temperatures was small (0.73%), demonstrating stable control as set-point temperatures vary, which is necessary for practical applications (e.g., higher temperatures for higher boiling point gases). Additional logic that increased the maximum power application at lower ACFC temperatures resulted in a 36% decrease in energy consumption. Implementing such control logic improves energy efficiency for separating and concentrating organic gases for post-desorption liquefaction of the organic gas for reuse.
Introduction Volatile organic compounds (VOCs) are regulated as ozone precursors because they contribute to the formation of ground-level ozone, which is hazardous to human health. In 2008, 14.5 Tg of VOCs was emitted into the atmosphere from anthropogenic sources in the United States (1). A variety of applications utilize VOCs for production of materials (e.g., packaging materials and coatings), and they emit dilute concentrations of VOCs (200 ppmv, activated carbon adsorption can achieve >99% removal (3) and allows for reuse of the recovered VOC. Activated carbon fiber cloth (ACFC) is an * Corresponding author phone: (217) 333-6963; fax: (217) 3336968; e-mail:
[email protected]. 738
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 45, NO. 2, 2011
effective adsorbent for VOCs and has a larger adsorption capacity for dilute VOCs than granular activated carbon due to its microporous structure and large micropore volume (4-6). Desorption of VOCs from adsorbents can be achieved by thermal swing adsorption (TSA). Steam heating is a common TSA desorption method that has disadvantages including water corrosion of system components, heat loss to system components, low concentration of recovered material, and potential water contamination with the adsorbate requiring further water purification (7, 8). Electrothermal swing adsorption (ESA) is another form of TSA studied since the 1970s and involves applying electrical voltage across the activated carbon, which acts as an electrical resistor that dissipates heat as current passes through the adsorbent (9). The desorption gas concentration can thus be controlled independently from the carrier gas flow rate, providing increased control of the concentration ratio, which is the outlet adsorbate concentration generated during a desorption cycle divided by the inlet adsorbate concentration during the preceding adsorption cycle. ACFC adsorption systems have shown maximum concentration ratios of 63, 100, 240, and 1050 (10-13). Additional benefits of ACFC-ESA are that it does not require steam, energy is delivered directly to the adsorbent, and the heat and mass transfer rates of the adsorbent are rapid (10, 11, 14). ACFC has also been shown to maintain a high adsorption capacity for >300 electrothermal heating cycles (15). The long lifetime of ACFC with a high adsorption capacity has allowed a dual-vessel ACFCESA system (16) to continuously capture and recover organic vapors as liquid and organic gases as liquid with postdesorption temperature/pressure control (12). Another benefit to ESA is that it allows for rapid feedback control of applied voltage for heating the adsorbent, which has been utilized to provide constant VOC concentrations during desorption cycles (between 496 ( 3 ppmv (mean ( standard deviation) and 4962 ( 32 ppmv) (17). Careful control of the adsorbent’s temperature is also desirable when high VOC concentrations are required for post-desorption combustion for disposal or condensation for recovery. High VOC concentrations have been achieved for activated carbon monoliths (18) and ACFC (19) by increasing the desorption temperature, although a practical upper temperature limit for increasing the VOC concentration has been demonstrated (20). Another study showed that increasing the heating duration and electrical energy application increases the total mass desorbed, while increasing the electrical current with constant power has no effect (21). Experiments were also performed in which 7 V of direct current (dc) was initially applied to ACFC for desorption and then 4 V of dc was applied once the adsorbent reached 110 °C, and the desorption rate was higher than applying a constant 4 V of dc, suggesting that increased power application during initial heating achieves higher desorption concentrations (22). While ESA has been analyzed at specific temperatures for select adsorbents/adsorbates, differences in desorption properties resulting from maintaining a time-weighted average adsorbent temperature while oscillating the realtime adsorbent temperature or power applied to the adsorbent have not been characterized. Feedback control techniques which increase the rate of initial heating through increasing power application have the potential to increase desorption rates and reduce energy requirements to regenerate an adsorbent, but such an approach has not been characterized. Thus, this work focuses on three areas that have not been well characterized for adsorbent regeneration: 10.1021/es103303f
2011 American Chemical Society
Published on Web 12/15/2010
FIGURE 1. ACFC adsorption/desorption vessel gas flow. the effects of temperature oscillations, power oscillations, and the rate of initial temperature rise on the energy efficiency and VOC concentration profile during regeneration. This information assists in selection of the proper temperature and power profiles necessary for a specific application such as post-desorption condensation, which requires high VOC concentrations.
Adsorbent Properties, Apparatus, and Methods Adsorbent Properties. ACFC is a semiconductor, which allows for its electrothermal regeneration (14, 19, 23). The ACFC used in this study is ACC 5092-15 (American Kynol, Inc.) made from a phenolic resin. ACFC-15 was selected for its large surface area, micropore volume, and microporosity of 1322 m2/g, 0.621 cm3/g, and 94.4%, respectively (24). Mass was determined by cutting the ACFC into 20 cm × 260 cm rectangles, heating the rectangles to 125 °C in air at atmospheric pressure for >8 h to remove adsorbed water, and then weighing the rectangles with a gravimetric balance (Mettler, BB2400). Areal density was then determined by dividing the mass of the samples by their respective area. Annular cartridges were assembled by rolling the ACFC around 1.9 cm diameter stainless steel cylinders that were located at each end of the cartridge (12). Apparatus and Methods. Experiments were performed with a 3.5 L interior volume vessel containing two annular ACFC cartridges electrically connected in series with a total of 183.3 ( 0.85 g of ACFC (Figure 1) (12). Gas flow rates were controlled with mass flow controllers (air, Aalborg, model GFC571S; isobutane/nitrogen, Tylan Inc.). Gas streams for adsorption cycles were generated by mixing isobutane (Aeropres Corp., 97.8% isobutane, vapor withdrawal) with house compressed air that passed through silica gel and a high-efficiency particulate air filter to remove water and particles. A 50 SLPM gas stream containing 2000 ppmv isobutane was directed into the vessel until the exhaust stream reached a breakthrough concentration of 1000 ppmv isobutane (i.e., 50% breakthrough) as measured with a photoionization detector (RAE Systems, Inc., PDM-10A). After breakthrough, the vessel was purged of O2 with 15 L of N2 (S.J. Smith, 99.95% N2) at 5 SLPM. The N2 flow rate was then reduced to 0.5 SLPM, and a feedback controller was activated to heat the ACFC. All controllers were designed with Labview 6.1 software and controlled dc voltage to a silicon-controlled rectifier (SCR; Robicon, model 440 102.10), which supplied up to 120 V of alternating current (ac) that was then reduced with a variable-voltage transformer (Variac, Powerstat) and applied to the ACFC for heating. Direct contact type K thermocouples (0.081 cm diameter, Omega Inc.) measured
the ACFC temperature during heating. The concentration of isobutane in the exhaust gas stream during regeneration cycles was continuously measured with a flame ionization detector (FID; MSA Inc., series 8800) because the concentration was above the photoionization detector’s measurement range. Root mean square (rms) voltage and rms current applied to the ACFC were measured with a potentiometer (Omega, Inc.) and a current transformer (Split Core) connected to an ammeter (Omega, Inc.), respectively. All measurements were continuously stored at 1 Hz with National Instruments Fieldpoint hardware. These measurements were utilized to determine the mass of isobutane desorbed from the ACFC, which is proportional to the area under the desorption concentration profile (Figure 4C) (Mdesorbed, g): tf
Mdesorbed )
∑M i)1
[
PQN2Yiso
W
RT
]
(ti - ti-1)
(1)
where MW ) molar mass (g/mol), P ) pressure (atm), R ) ideal gas constant ((L · atm)/(K · mol)), T ) temperature (K), QN2 ) flow rate of N2 during desorption (SLPM), t ) time (min), tf ) final time (min), and Yiso ) isobutane to N2 mole ratio at ti. Total energy consumption (E, J) for heating was determined by integrating the power applied over the desorption cycle: tf
E)
∑ IV(t - t i
i-1)
(2)
i)1
where I ) rms current (A) and V ) rms voltage (V). Since physical adsorption is a reversible process, the energy released during physical adsorption is equal to the energy required for desorption of the adsorbate (independent of heat loss to the adsorbent and carrier gas), so the heat of adsorption can be estimated through measurement of the heat of desorption (25). The following method was used to estimate the heat of desorption. ACFC loaded with adsorbate was heated for a desorption cycle to determine energy consumption, which includes energy for both desorption of the adsorbate (evaporation and overcoming intermolecular attraction) and heat losses to the system. ACFC was then heated at the same conditions, except without adsorbate, to determine the energy losses to the system. The energy required to desorb the adsorbate was determined by subtracting the energy required to heat ACFC without adsorbate from the energy required to heat ACFC with adsorbate. The VOL. 45, NO. 2, 2011 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
739
FIGURE 2. Control loop for the ACFC regeneration controller based on temperature feedback.
TABLE 1. Controller Constants for ACFC Heating control constant
PID control with temperature feedbacka
secondary PID control with current feedbackb
proportional integral derivative
0.70 0.05 0.05
0.2000 0.0005 0.0500
a
PID and controller.
PID
delay
controllers.
b
Secondary
PID
average heat of desorption (kJ/mol) for a desorption cycle was then determined by dividing the energy required to desorb the adsorbate (independent of heat losses to the system) by the number of moles of desorbed adsorbate: heat of desorption )
E50% - E0% MW Mdesorbed
(3)
where Ei% ) energy consumption for heating ACFC adsorbed to i% breakthrough (kJ). Also, at 50% breakthrough, >96% of the total equilibrium mass is adsorbed, suggesting a close approximation to equilibrium conditions. Regeneration Controllers. Feedback controllers were developed in the software and apply a 0-5 V dc signal to the SCR on the basis of the set-point and measured temperatures of the ACFC. The SCR then applies voltage to the ACFC that is proportional to its input voltage, resulting in ACFC heating. Control of the voltage affects the resulting applied power and outlet isobutane concentration profile during the desorption cycle. Select temperature feedback controllers were
tested for heating the ACFC to a set-point of 200 °C, including on/off, PID (proportional integral derivative), PID delay, and PID delay with a secondary PID controllers. The on/off, PID, and PID delay are primary controllers that receive temperature feedback from a thermocouple and control voltage applied to ACFC to achieve and maintain a temperature set-point (Figure 2). The on/off controller applies 5 V of dc to the SCR when the ACFC temperature is lower than the set-point temperature and 0 V of dc when the ACFC’s temperature is higher than the set-point temperature. The PID controller, which applies 0-5 V of dc to the SCR, was initially tuned with the Ziegler-Nichols method (26), and additional adjustments were made to the constants, with final values provided in Table 1. The PID delay controller applies 5 V of dc to the SCR until the ACFC is within 2% of the set-point temperature, and then the PID controller is activated. During initial ACFC heating from ambient temperature to the set-point temperature, each controller applies the maximum output voltage determined by the SCR. ACFC is a semiconductor such that increased temperature results in decreased resistance and thus an increase in current. The maximum current applied for desorption heating was limited to 30 A as a safety protocol to avoid damaging the system’s hardware. Thus, the SCR maximum voltage setting is determined to maintain a current of
