Article pubs.acs.org/est
Temperature Control during Regeneration of Activated Carbon Fiber Cloth with Resistance-Feedback David L. Johnsen and Mark J. Rood* Department of Civil and Environmental Engineering, University of Illinois, 205 North Mathews Avenue, Urbana, Illinois 61801, United States
ABSTRACT: Electrothermal swing adsorption (ESA) of organic compounds from gas streams with activated carbon fiber cloth (ACFC) reduces emissions to the atmosphere and recovers feedstock for reuse. Local temperature measurement (e.g., with a thermocouple) is typically used to monitor/control adsorbent regeneration cycles. Remote electrical resistance measurement is evaluated here as an alternative to local temperature measurement. ACFC resistance that was modeled based on its physical properties was within 10.5% of the measured resistance values during electrothermal heating. Resistance control was developed based on this measured relationship and used to control temperature to within 2.3% of regeneration set-point temperatures. Isobutane-laden adsorbent was then heated with resistance control. After 2 min of heating, the temperature of the adsorbent with isobutane was 13% less than the adsorbent without isobutane. This difference decreased to 2.1% after 9 min of heating, showing desorption of isobutane. An ACFC cartridge was also heated to 175 °C for 900 cycles with its resistance and adsorption capacity values remaining within 3% and 2%, respectively. This new method to control regeneration power application based on rapid sensing of the adsorbent’s resistance removes the need for direct-contact temperature sensors providing a simple, cost-efficient, and long-term regeneration technique for ESA systems.
■
for treating a 68 m3/min gas stream with 710 ppmv toluene) because adsorbent mass is reduced (e.g., required ACFC mass was 3% of required GAC mass in ref 11) and electrothermal regeneration can be used, reducing energy consumption compared with steam regeneration.10 Adsorption systems with annular ACFC cartridges continuously capture and recover organic vapors and gases as a liquid for reuse or disposal.1,12−14 Regeneration of these cartridges involves electrothermal heating to a specified cartridge temperature, which is dependent on the cartridge’s electrical resistivity and geometry.1,15−17 The resistivity of flat rectangular ACFC samples was used to predict the resistance of ACFC cartridges based on the cloth’s resistivity, temperature, and geometry.1 This technique allows for design of the appropriate size and shape of ACFC cartridges to achieve required
INTRODUCTION Activated carbon fiber cloth (ACFC) has higher adsorption capacities than typical granular activated carbons (GACs) for select volatile organic compounds (VOCs) due to its large porosity, surface area, and micropore volume.1−3 Thermal regeneration techniques for adsorbate-laden ACFC include steam, microwave, electromagnetic induction, and electrothermal (Joule) heating.1,4−8 Electrothermal heating is a promising regeneration technique used for >40 years that involves passing electrical current through the adsorbent to achieve resistive heating.9 Benefits of electrothermal heating include rapid heating, elimination of aqueous condensate present with steam heating, and direct application of power to the adsorbent independent of the gas flow passing through the adsorbent allowing for accurate control of regeneration cycles.9,10 However, ACFC-electrothermal swing adsorption (ESA) has increased cost per unit mass of adsorbent compared to GACs and requires an inert carrier gas during regeneration cycles when the adsorbate is reactive with oxygen and at concentrations between its lower and upper explosive limits.10 ACFC systems are cost competitive with GAC systems (e.g., © 2012 American Chemical Society
Received: Revised: Accepted: Published: 11305
July 31, 2012 September 10, 2012 September 11, 2012 September 11, 2012 dx.doi.org/10.1021/es303099t | Environ. Sci. Technol. 2012, 46, 11305−11312
Environmental Science & Technology
Article
Figure 1. Apparatus for electrical resistance measurements comprised of ACFC rectangles (A), electrodes (B), electrical connections (C), and thermocouple (D).
considered when using the electrical resistance of an adsorbent to determine adsorbent temperature. ACFC has been shown to maintain VOC adsorption capacity for >300 heating cycles, suggesting consistent capture and recovery of VOCs can occur for numerous adsorption and regeneration cycles.24 ACFC’s electrical resistance must also remain consistent to determine temperature of the adsorbent to control regeneration cycles on a long-term basis. This study presents a novel ESA regeneration method, referred to as resistance-feedback control, which uses remote voltage and current measurements to determine the adsorbent’s electrical resistance. Resistance-feedback control is used to control the adsorbent’s temperature during regeneration instead of using localized temperature sensors. To develop and evaluate this method, first, the electrical resistance of ACFC was characterized from ambient temperature to typical ESA regeneration temperatures. From this relationship between temperature and resistance, resistance-feedback control was developed and used to electrothermally regenerate ACFC with and without adsorbed isobutane to determine the effect of adsorbed material on electrical resistance. Last, ACFC was heated for 900 automated cycles to evaluate the consistency of resistance values during electrothermal heating, which is necessary for long-term monitoring and control of regeneration cycles with this method. Thus, this study introduces a new method to control electrothermal heating of an adsorbent based on rapid remote electrical resistance measurements, which eliminates the need for direct-contact temperature sensors and demonstrates consistent resistance values after repeated electrothermal heating cycles, showing that this costeffective and simple ACFC-ESA regeneration method can operate for extended durations for practical applications.
temperatures, which can reduce regeneration power requirements and provide a simpler and lower cost configuration. Temperature control during electrothermal heating typically requires local temperature sensors to determine the amount of power applied to the ACFC to achieve the set-point temperature. Thermocouples and resistance temperature detectors (RTDs) are commonly used to measure ACFC temperature and perform best with direct adsorbent contact for rapid measurement of the adsorbent’s temperature.14 This measurement requires the sensor to be electrically isolated from the ACFC during electrothermal heating, increasing measurement complexity. Direct-contact temperature sensors can lose contact with the ACFC or create a short circuit during electrothermal regeneration that damages the thermocouple or adsorbent, requiring maintenance and replacement of these components while reducing system run time. Thermocouples and RTDs also provide point measurements, which may not be representative of the average adsorbent temperature as the ACFC temperature distribution during electrothermal heating has been shown to have a standard deviation ranging from 3 to 60 °C when heating different types of ACFC to 60−130 °C.18 Alternative sensors that have measured carbonaceous adsorbent temperatures include an infrared camera, a dilation liquid thermometer, and a fiber optic sensor.15,18,19 Each of these sensing techniques has been effective but requires the purchase and maintenance of the sensor. However, during electrothermal regeneration of ACFC, electrical resistance is typically determined based on measured current and voltage values. A method that treats an entire ACFC cartridge as an RTD to control its temperature based on remote measurements reduces the need for local temperature sensors, thus potentially reducing the costs for maintenance and the replacement of system components while increasing system run time. In addition to temperature, electrical resistance of activated carbon is dependent on the adsorbed material. Electrical resistance has been shown to increase, decrease, or initially increase and then decrease with increasing adsorbed mass.20−23 For example, the resistance of activated carbon rods made of cellulosic material decreased by 3.1% and increased by 0.9% after adsorption of isobutane and ethylene oxide, respectively.23 Thus, the type and mass of adsorbed adsorbate should be
■
ADSORBENT PROPERTIES, APPARATUS, AND METHODS Resistivity of Flat Sheets of ACFC. ACC 5092-15 (American Kynol, Inc.) was the ACFC used in this study and has an areal density (mass/bulk surface area) of 176 g/m2, BET surface area of 1335 m2/g, total pore volume of 0.64 cm3/g, microporosity of 96.6%, and average micropore width of 0.76 nm.25,26 Resistivity was determined based on the average resistance measurements from five rectangular ACFC samples 11306
dx.doi.org/10.1021/es303099t | Environ. Sci. Technol. 2012, 46, 11305−11312
Environmental Science & Technology
Article
(7.4 cm × 3.9 cm) from different sections of the same lot of material in a 1.85 L interior volume Pyrex vessel (Figure 1). Each rectangle (Figure 1(A)) was supported across its width by stainless steel electrodes (Figure 1(B)). The vessel was purged with 2 SLPM N2 (25 °C and 98.5 kPa were considered standard conditions) for 30 min, while voltage (Figure 1(C)) was applied across the ACFC with a direct current (DC) power supply (Tenma, model 72-2085) to maintain the ACFC at 200 °C to desorb volatile adsorbates (e.g., water vapor) from the ACFC. ACFC temperature was measured with a Type K thermocouple (Figure 1(D), 0.25 mm diameter, Omega Inc.) contacting the ACFC at the center of the sample. The ACFC was then heated from 50 to 210 °C in 20 °C increments, and current and voltage were measured concomitantly with a multimeter (Fluke, model 45). Current and voltage were used to calculate the resistance of the ACFC (without considering electrode contact resistance), and then resistivity (ρ(T), Ω m) was calculated based on the relationship between resistance, temperature, and geometry.1
ACFC cartridges occurred in a 1.5 L interior volume Pyrex vessel (Figure 2). For adsorption cycles, a 50 SLPM air stream
Figure 2. Schematic (A) and photograph (B) of the ACFC cartridge in a vessel.
containing 2000 ppmv isobutane was directed through the vessel until the vessel’s exhaust stream reached a breakthrough concentration of 1000 ppmv isobutane as measured with a photoionization detector. For regeneration cycles, 0.5 SLPM N2 passed through the vessel, and a resistance- or temperaturefeedback controller controlled the power application for heating to desorb isobutane from the ACFC. ACFC temperature was measured with a thermocouple located at the center of the outer layer of cloth. Root mean square voltage and current were measured with a potentiometer and a current transformer connected to an ammeter, respectively. All measured values were stored at 1 Hz. The following section describes experiments to determine cartridge resistance as a function of temperature. ACFC was heated to temperature set-points for 5 min to allow time for voltage and current stabilization. Then a 100 sample arithmetic mean was calculated for measured voltage and current values and used to calculate resistance. Equations 1−3, that were used to determine the resistivity of a flat sheet of ACFC, were also used to model the resistance of the ACFC cartridge based on the ACFC’s resistivity, temperature, and geometry to evaluate closure between the measured and modeled resistance values. Resistance-Feedback Controller. Resistance-feedback control converts a temperature set-point to a resistance setpoint to control ACFC heating based on the ACFC’s real-time resistance values, determined using remote amperage and voltage measurements. This method eliminates the need for local temperature measurements. In this study, ACFC voltage and current values were used as feedback for proportional plus integral control of the voltage applied to the ACFC to achieve and maintain the temperature set-point (Figure 3). Conditioning was used to reduce variability in the real-time calculated resistance values to achieve stable temperature control based on resistance. Conditioning involved the removal of outlier resistance values (4 Ω, constituting 0.97 for all cartridges showing linearity in the measured values. This relationship is valuable because it allows for ACFC temperature prediction with remote resistance measurements. Dynamic Resistance-feedback Control. A linear regression relating measured resistance (R, Ω) and temperature 11308
dx.doi.org/10.1021/es303099t | Environ. Sci. Technol. 2012, 46, 11305−11312
Environmental Science & Technology
Article
(T, °C) values was determined for cartridge 1 (Figure 4, R = −6.9 × 10−3T + 3.4, r2 = 0.993) so that temperature set-points can be converted to resistance set-points for resistance-feedback control. This linear regression of the measured data was used instead of modeled values to achieve more accurate control because the AAE between the regression and the measured resistance values was smaller (0.8%) than the AAE between the modeled and measured resistance values (10.5%). The linear regression was used to determine resistance set-points ranging from 2.67 to 1.98 Ω, corresponding to typical regeneration temperatures from 100 to 200 °C. Resistance-feedback control was used to heat cartridge 1 to select temperature set-points for 5 min intervals (Figure 5).
Figure 6. Temperature (standard deviations as vertical bars) and resistance values when heating cartridge 1 with and without adsorbate using resistance-feedback control. Horizontal solid lines represent the temperature and resistance set-point values during the regeneration cycle.
Table 2. Resistance-Feedback Control Responses from Heating Cartridge 1 to 2.32 Ω, as Shown in Figure 6a no adsorbate loading comparison initial value (Ω, °C) set-point (Ω, °C) settling time (s) AAE (%)b
with adsorbate loading
resistanced temperatured resistancee c
NA 2.32 33 1.0
30 150 100 0.7
NA 2.32 36 0.8
temperaturee 30 150 874 12.7
a
For each condition (with and without adsorbate loading), values are provided for the resistance response (directly controlled and measured) and the temperature response (measured during resistance control). bACFC without adsorbate first reached the set-point temperature in 2 min. AAE values for heating both with and without adsorbate were thus calculated for 1 min starting after 2 min of heating. cNot applicable. dDotted resistance and temperature lines from Figure 6. eSolid resistance and temperature lines from Figure 6.
Figure 5. Resistance-feedback control when heating cartridge 1 without adsorbate. Black lines describe set-points or predicted values, and blue/red data points describe measured values. Labeled AAE (%) values, at each set-point, were calculated starting when resistance reached each set-point value.
The ACFC was heated, and after initially reaching each resistance set-point, measured resistance values were within 1% AAE of the set-point resistance values showing accurate resistance control. Once the resistance value reached the setpoint value, measured ACFC temperature was compared to modeled temperature that was predicted based on the resistance set-point. Measured and modeled temperatures agreed within 2.3% for each set-point. No overall trends were observed between AAE and temperature set-point. This close agreement between expected and measured values suggests that remote resistance measurements can be used to determine and control ACFC temperature. Resistance-feedback Control with Adsorbed Isobutane. Cartridge 1, without and with adsorbate (0.1074 ± 0.0002 g isobutane/g ACFC), was electrothermally heated for 40 min with resistance-feedback control to a set-point of 2.32 Ω, which is equivalent to 150 °C for ACFC without adsorbate (Figure 6). For each experiment, controller performance was evaluated based on both resistance response and temperature response (Table 2). The controller’s resistance response was determined by comparing the difference between the measured resistance and the 2.32 Ω set-point value, and the temperature response was determined by comparing the difference between the measured temperature and the 150 °C temperature that is predicted for heating ACFC with no adsorbate to this 2.32 Ω resistance set-point. However, adsorbed mass alters the ACFC’s resistance. Because power application was controlled to maintain ACFC resistance at 2.32 Ω for all tests, differences
between temperature profiles for the cases with and without isobutane were attributed to adsorbed mass. For electrothermal heating to 2.32 Ω based on resistance feedback, the regenerated ACFC (without isobutane) achieved the predicted set-point temperature within 2 min. The difference in temperature of the adsorbent with and without isobutane was 13.0% at 2 min of heating and 2.1% at 9 min of heating (Figure 6). The difference in temperature during the remaining 31 min of regeneration averaged 1.2%. The larger differences in temperature at the beginning of the regeneration cycle likely occurred because the remaining adsorbed isobutane altered the ACFC’s electrical resistance (with respect to temperature). Then, as the isobutane desorbed from the ACFC, the temperature profiles converged. Thus, the settling time (described in the Methods section) for the temperature to achieve 150 °C is increased (874% larger) when isobutane is present compared to when no adsorbate is present. For the same experiments, the settling time for the resistance to achieve 2.32 Ω was similar (within 8.5%) for ACFC with or without initial adsorbate, which is expected because power application was directly controlled to achieve this resistance set-point. Although the settling time for temperature increased with the presence of isobutane, the maximum AAE value between measured and predicted temperatures was still
