Field-rugged sensitive hydrogen peroxide sensor based on Tunable Diode Laser Absorption Spectroscopy (TDLAS)

Transcription

1 Field-rugged sensitive hydrogen peroxide sensor based on Tunable Diode Laser Absorption Spectroscopy (TDLAS) M.B. Frish*, J. R. Morency, M.C. Laderer, R.T. Wainner, K.R. Parameswaran, W.J. Kessler, and M.A. Druy Physical Sciences Inc., 2 New England Business Center, Andover, MA 181 ABSTRACT This paper reports the development and initial testing of a field-portable sensor for monitoring hydrogen peroxide (H 2 O 2 ) and water (H 2 O) vapor concentrations during building decontamination after accidental or purposeful exposure to hazardous biological materials. During decontamination, a sterilization system fills ambient air with water and peroxide vapor to near-saturation. The peroxide concentration typically exceeds several hundred ppm for tens of minutes, and subsequently diminishes below 1 ppm. The H 2 O 2 / H 2 O sensor is an adaptation of a portable gas-sensing platform based on Tunable Diode Laser Absorption Spectroscopy (TDLAS) technology. By capitalizing on its spectral resolution, the TDLAS analyzer isolates H 2 O 2 and H 2 O spectral lines to measure both vapors using a single laser source. It offers a combination of sensitivity, specificity, fast response, dynamic range, linearity, ease of operation and calibration, ruggedness, and portability not available in alternative H 2 O 2 detectors. The H 2 O 2 range is approximately - 5, ppm. The autonomous and rugged instrument provides real-time data. It has been tested in a closed-loop liquid/vapor equilibrium apparatus and by comparison against electrochemical sensors. Keywords: TDLAS, Hydrogen Peroxide, Sensors, Gas Analysis, Instrumentation, Diode Lasers, Spectroscopy 1. INTRODUCTION Sterilization with Vapor Phase Hydrogen Peroxide (VPHP or H 2 O 2 ) is a leading technique for decontaminating buildings, offices, laboratories, and hospital rooms exposed, purposefully or accidentally, to hazardous biological materials. 1 VPHP is a highly-reactive chemical but, when compared to other sterilization agents such as chlorine dioxide, it is relatively innocuous for decontaminating areas containing electronics or corrodible surfaces. Decontamination with VPHP is a chemical sterilization technique commonly utilized in the food, pharmaceutical, and medical industries to decontaminate production or processing equipment and tools. During this process, the sterilization system fills the ambient air with water and peroxide vapor, raising the combined vapor concentrations to or just below their dew point. 2 The VPHP concentration generally reaches 2-12 ppm (depending on initial ambient temperature, humidity, and room size) and remains at that concentration for tens of minutes. 3 Subsequently, the concentration must diminish below 1 ppm for the area to be safely occupied. 4 To optimize the decontamination process efficiency, verify the process endpoint, and thus minimize the decontamination cycle time, continuously monitoring and regulating the VPHP concentration during decontamination is necessary. In large rooms or buildings it may be useful to monitor several locations concurrently. Presently, however, there is no field-portable, automated and autonomous, rugged, reliable, accurate, fast and sensitive instrumentation suitable for realtime H 2 O 2 measurement over the full concentration range of interest for monitoring and control while decontaminating *frish@psicorp.com; phone ; fax ; psicorp.com

2 buildings. A significant challenge for currently-available VPHP measurement technologies is spanning the full range of concentrations encountered during a decontamination procedure without cross-sensitivity to ubiquitous water vapor. In spectroscopic VPHP sensors, the cross-sensitivity results from a complex interlacing of H 2 O 2 and H 2 O spectral features. We are addressing this need with the optical spectroscopic technology known as Tunable Diode Laser Absorption Spectroscopy (TDLAS). TDLAS is a robust configurable trace gas sensing technology deployed for industrial process monitoring and control, quality assurance, environmental sensing, plant safety, and infrastructure security Novel adaptations of mature TDLAS platforms address applications that demand sensing multiple target gases in spectroscopically-complex gas mixtures with high spatial or temporal resolution in harsh environments. TDLAS sensors offer a combination of sensitivity, specificity (i.e. freedom from cross-sensitivity to non-target gases), fast response, dynamic range, linearity, ease of operation and calibration, ruggedness, and portability not available in alternative H 2 O 2 detectors. Capitalizing on these virtues, we have adapted a commercial TDLAS platform to create a sensitive field-portable H 2 O 2 analyzer that provides accurate real-time H 2 O 2 and H 2 O measurements. Unlike other spectroscopic H 2 O 2 analyzers that probe spectral bands, TDLAS capitalizes on the exquisite spectral resolution of Distributed Feedback (DFB) lasers to probe individual spectral lines. Rather than being confounded by spectral complexity from the multitude of neighboring and overlapping H 2 O 2 and H 2 O spectral lines, the TDLAS analyzer isolate H 2 O 2 spectral lines from neighboring H 2 O lines. It eliminates, within the sensor detection limits, cross-sensitivities between the two vapor species and measures both vapors using a single DFB laser source. The H 2 O 2 measurement range is about -5, ppmv with sub-ppm resolution. The H 2 O range is about 5% with better than.5% resolution. Benchtop tests described herein using a closed-loop liquid/vapor equilibrium apparatus demonstrated sensor response. Tests during room decontamination provided accurate calibration and comparison with electrochemical sensors while demonstrating real-time continuous data. 2. SENSOR DESCRIPTION Figure 1 is a block diagram of the main TDLAS H 2 O 2 / H 2 O sensor components. It comprises three modules: 1) The Control Unit; 2) The Sensor Head, and 3) The System Interface. Figure 2 shows the components separated on a laboratory benchtop, and the sensor system assembled in a hand-portable unit. Serial Interface System Interface Control Unit Optical Fiber Detector Signal Sensor Head Keypad & User Display Analyte Gas Flow Figure 1. TDLAS analyzer system block diagram. K-363

3 Figure 2. H 2 O 2 / H 2 O analyzer components and hand-portable assembly (cover removed). The Control Unit is housed within a 3 kg, 22cm x 28cm x 8 cm box that attaches to the Sensor Head by an umbilical cable bearing an optical fiber and a few wires. The Control Unit includes a full Wavelength Modulation Spectroscopy (WMS, described below) system implemented on a single 15cm x 15cm printed circuit board, pictured in Figure 3, with digital signal processing and an embedded microcontroller. The source laser is installed on the board. The H 2 O 2 / H 2 O sensor employs a DFB-type diode laser in a 14-pin butterfly package with a thermoelectric cooler and optical fiber output. Laser control, thermal control, signal processing, and data reporting functions are performed on the WMS board. The board draws less than 1.5 W of power. Four AA-sized Li-ion batteries provide up to 24 hours of continuous operation without recharge. Figure 3. WMS Board in compact package.

4 The Sensor Head, illustrated in Figure 4, comprises: 1) A Measurement Section formed from a Herriot-style multi-pass optical cell 12 in an aluminum tubular enclosure through which sample gases flow; and 2) a Transceiver Section that includes laser beam launch and receive optical components, a photodetector, and pre-amplifier. A sealed glass window mechanically separates the two sections while transmitting the laser beam between them. A pair of diaphragm pumps (seen in Fig.2) draws ambient air through a pressure-reducing orifice into and through the Measurement Section. Pressure within the Measurement Section is about 1 torr. The Measurement Section volume is approximately 7 ml. The gas exchange period is approximately 1s. All wetted components external to the Measurement Section are PTFE. Figure 4. Semi-transparent drawing of Sensor Head. The optical fiber conducts the laser beam from its origin at the Control Unit to the Transceiver. The fiber terminates at a collimator within the transceiver and launches the laser beam through the window into the Measurement Section. There, the laser beam traverses the sample gas, covering an optical path length of about 2.7 m within the Herriot cell, which has a physical length of 12 cm and 2.5 cm diameter. The laser beam exits the Herroit cell and returns through the window to the Transceiver, where it impinges upon a photodetector which converts the beam, attenuated by target gas absorption, into an electrical signal. The signal processor portion of the Control Unit unit interprets the electrical signal and outputs information via a local display and a remote communications signal to the System Interface. 3. TDLAS TECHNIQUE The measurement principle is absorption spectroscopy. Many gaseous molecules absorb light at specific wavelengths (called absorption lines). For example, Figure 5 presents the absorbance spectra for CO 2 and H 2 O across a 1 m path at standard temperature and pressure, calculated using the HITRAN database. 13 To measure target gas number density, N, the DFB laser emits a well-defined but adjustable or tunable wavelength having a linewidth narrower than the target gas absorption line width. The target gas partially attenuates the laser beam according to the Beer-Lambert law, I ν = I ν exp[-s(t) g(ν-ν,p) N l] I ν exp[-α(ν)] (1)

5 where: I ν I ν l S(T) o is the launched laser power of wavenumber ν is the received laser power after propagation through the absorbing target gas is the optical pathlength through the gas is the temperature-dependent line-strength (a fundamental spectroscopic property) of the target molecule s absorption line is the frequency (reciprocal of wavelength) at the center of the absorption line g( - o, P) is the lineshape parameter at the laser frequency and pressure P α(ν) is the absorbance N is the target gas molecular number density. N is deduced using explicit knowledge via calibration, calculation, or measurement of each of the other terms in Eq. (1)..7.6 Atmospheric Absorption (1 meter path length) 296 K, 1 atm Water CO Wavelength (nm) Figure 5. Example gas absorption spectra. H-973 for SR The lineshape g( - o, P) is a complex function of temperature and pressure. When the total pressure of the gas sample is sufficiently high, resulting in frequent intermolecular collisions, the lineshape function assumes a Lorentzian form: g( - o ) = [1/ g o P] [1/({( - o )/g o P} 2 + 1)] (2) where g o = broadening coefficient (cm -1 /atm). Eq. (2) shows that the linewidth, g o P, (i.e., the range of wavenumbers spanned between the half-maxima of the lineshape) is proportional to the total pressure of the gas sample. At low pressures, the lineshape function is Gaussian and independent of pressure. Because linewidths decrease with pressure, operating the TDLAS sensor with the sample gas at reduced pressure facilitates sampling individual target gas absorption lines that are distinct from absorption lines from other gases, thus enabling selectivity for a target trace gas absent spectral interferences from other background gases and reducing cross-sensitivities. The TDLAS sensor exploits the laser s fast tuning capability to rapidly modulate the wavelength, causing it to sweep back and forth across an absorption line at a precise modulation frequency, ω m. 14,15 The response to the wavelength

6 modulation is an amplitude modulation at the detector, illustrated in Figure 6. The amplitude modulation arises from two sources: 1) the transmitted laser power follows the wavelength modulation and thus modulates sinusoidally at frequency ω m ; and 2) the absorption due to target gas. Since the wavelength crosses the target gas absorption line twice for each modulation cycle, the amplitude modulation due to absorption occurs at precisely twice the modulation frequency, 2ω m. Via lock-in amplification, the signal processor measures the average values of the amplitude modulations at ω m and 2ω m. These are called the F1 and F2 signals respectively. F1 is proportional to the received laser power, while F2 is proportional to the received laser power and the target gas absorbance. Thus, the ratio F2/F1 is proportional to the absorbance only, and is independent of the received laser power as long as it exceeds sensor noise. The modulation and lock-in detection technique is called as wavelength modulation spectroscopy (WMS). WMS typically measures absorbances of 1-5 or less with one second or faster response, and for H 2 O 2 and H 2 O provides ~1 ppm-m detection limits Molecular Transmission Spectrum o Wavelength Wavelength Modulation o + o o + m 2 m Time 3 m 2.6 = WMS 1KHz Modulation Time (µs) No Target Gas With Target Gas Figure 6. WMS spectroscopy: Laser wavenumber ν modulates sinusoidally across a spectral absorption line with modulation depth ± δ and modulation frequency ω m, creating an amplitude modulated detector signal. J-4176a 4. H 2 O 2 SPECTROSCOPY AND CALIBRATION Figure 7 shows the F2 signals vs. wavelength obtained by measuring the vapor at 85 torr above a small liquid reservoir (seen in Fig.2) containing nominally 3% by weight H 2 O 2 in H 2 O solution. The solution spectrum is compared with a tap water spectrum obtained under identical conditions. The difference between the two spectra yields the H 2 O 2 spectrum. The indicated H 2 O 2 and H 2 O lines were selected for the WMS measurements. It is notable that at the center of the selected H 2 O 2 line, the contribution to F2 due to water vapor is very close to zero, as required to preclude cross-sensitivity.

7 Figure 7. H 2 O 2 and H 2 O spectra. When the average laser frequency is tuned to coincide with the H 2 O 2 absorption line peak, then F2 is proportional to H 2 O 2 concentration. Eq. (3) yields concentration upon measuring F2 and F1: [H 2 O 2 ] = C*(F2 F2 o )/F1 (3) where C and F2 o are constants determined by a two-point (zero and span) calibration. Because F1, F2 and F2 o are proportional to transmittance of the optical system, the ratio (F2 F2 o )/F1 is proportional to concentration and independent of optical transmittance. F2 o represents an offset resulting from non-linearity of laser launch power vs. laser injection current as well as from wavelength-dependent transmittance of optical components other than the target gas. Speckle, interference patterns, fiber bend and coupling losses, and other effects of laser coherence are included in the latter contribution to the offset. These effects are sensitive to movement of optical components and may vary over time in response to changing ambient temperature or mechanical stress. The offset drift due to these changes is generally the limit to measurement accuracy. In practical industrial TDLAS sensors, the offset drift is typically equivalent to an absorbance of about 1-5. The H 2 O 2 sensor was calibrated by comparison with an electrochemical sensor during testing with a room decontamination system, described in Section 7 below. Knowing the calibration enables quantitative evaluation of offset drift, cross-sensitivity, and noise, thus yielding sensor accuracy and precision as described next. 5. DETECTION LIMIT The assembled sensor was attached to a closed circulation apparatus as shown schematically in Figure 8. The liquid solution was nominally 3% by weight H 2 O 2. Henry s law indicates that H 2 O 2 concentration should be approximately 5 ppm, but because H 2 O 2 decomposes rapidly, the actual concentration was lower, measured with the calibrated TDLAS as ~25 ppm.

8 PTFE Tubing Glass Flask Liquid Solution Sensor Head Pump K-3142 Figure 8. Closed circulation calibration configuration. Figure 9 shows the measured H 2 O 2 concentration and compares with measurement of zero gas. The zero gas indicates a noise-equivalent detection limit of.6 ppm at.1 Hz, representing the measurement precision Saturated Air re-circulating through 3% peroxide solution bubbler sec averaging 5-5 Dry air (w/1% CO 2 ) Avg =.4 ppm =.6ppm Time (s) Figure 9. Preliminary H 2 O 2 calibration and zero data, sensor tuned to sense H 2 O 2 vapor. 6. DRIFT AND CROSS-SENSITIVITY Figure 1 illustrates drift when drawing room air (~.6% H 2 O) through the sensor and compares with drawing zero gas. The drift is limited to less than 2 ppm. We believe that the initial 2 ppm measurement observed immediately after activating the pump is not true offset drift, but is actually residual H 2 O 2 that had adsorbed on sensor tubing during prior operation and de-adsorbed when pumping reduced the Measurement Section pressure. Subsequent fluctuations of about ±.3 ppm, not present when flowing zero gas, may be due to water vapor cross-sensitivity or to residual H 2 O 2.

9 Pump off Draw room air Draw zero gas Time (sec) Figure 1 Drift while drawing room air or zero gas with sensor tuned to measure [H 2 O 2 ]. The data of Figure 11 attempt to resolve the ambiguity in identifying the source of apparent drift. These data compare measurements of room air (approximately.6% H 2 O) with wet air drawn through the appartus of Figure 8 with the flask containing tap water only. The wet air has concentration fluctutaions reaching up to 3% H 2 O (saturation), as seen in Figure 12. From these data, we conclude that, when measuring H 2 O 2, cross-sensitivity to water vapor in saturated room air is less than ±.5 ppm. Drift due to residual H 2 O 2 and coherent optical effects is < 2 ppm. Since drift is signifcantly smaller (~.1 ppm) when flowing zero gas, we tentatively conclude that the observed drift is most likely measurement of residual H 2 O Room air Wet air Time (sec) 9 1 Figure 11. Sensor response with room air or nearly-saturated air flowing and sensor tuned to measure [H 2 O 2 ] Room air Bubbler CO 2 + room air leak Room air T=32.5C I~76mA m= 4 ma Time (s) Figure 12. Sensor response to changes in H 2 O concentration when tuned to sense H 2 O vapor.

10 7. TESTS WITH DECONTAMINATION APPARATUS Pictured in Figure 13, the H 2 O 2 / H 2 O sensor was positioned in a small room, approximately 3 m x 5 m floor area, employed as a test site for a Bioquell Q1 decontamination suite. The Q1 unit includes an electrochemical H 2 O 2 sensor and a relative humidity sensor (which senses the sum of [H 2 O 2 ] + [H 2 O]). The calibration of the TDLAS sensor was set to agree with the Q1 unit when the H 2 O 2 concentration reached its peak value during the decontamination cycle. Figure 14 compares the TDLAS measurements with the Q1 measurements throughout the duration of a decontamination cycle. Figure 13. TDLAS sensor in decontamination test site.

11 TDLAS Bioquell Time (min) TDLAS 25 Bioquell Time (min) TDLAS Bioquell Time (min) TDLAS Bioquell Time (min) Figure 14. TDLAS compared with Bioquell Q1 sensor measurements. The TDLAS unit switched between H 2 O 2 (top) and H 2 O (bottom) measurement at 75s intervals. Bioquell data ends after 7 minutes. The TDLAS sensor was tuned and calibrated during the initial ten minutes after starting the decontamination cycle. Note that the Q1 RH sensor measures [H 2 O 2 ] + [H 2 O], while the TDLAS [H 2 O] measurement includes no contribution from [H 2 O 2 ]. 8. CONCLUSION We have developed and demonstrated a hand-portable sensor for accurately and sensitively measuring vapor phase H 2 O 2 and H 2 O using TDLAS. H 2 O 2 measurement precision is approximately.1 ppm with.1 Hz bandwidth, and absolute accuracy is better than 2 ppm in the presence of saturated water vapor at room temperature (nominally 3% H 2 O). The measured accuracy limit combines the effects of sensor offset drift, cross-sensitivity to H 2 O, and residual H 2 O 2 desorbing from sensor components. Data acquired to date indicate that the latter is the primary source of measurement error. Future controlled calibrations will resolve the various contributions to the error limit. 9. ACKNOWLEDGMENT The authors thank Bioquell Inc. for enabling sensor tests at the Bioquell facility. This work was supported by EPA Grant EP-D REFERENCES [1] Rogers, J.V., et.al., Environmental Technology Verification Report: Bioquell, Inc. CLARUS C Hydrogen Peroxide Gas Generator, prepared by Battelle, Columbus, OH (24). [2] Waitling, D., Method and apparatus for hydrogen peroxide sterilization, European Patent EP B1 (23) [3] Waitling, D., Ryle, C., Parks, M., and Christopher, M., Theoretical Analysis of the Condensation of Hydrogen Peroxide Gas and Water Vapour as Used in Surface Decontamination, PDA J Pharm Sci Technol November/ 56: (22). [4] Fisher Scientific, Hydrogen Peroxide Materials Data Safety Sheet, Revision 3 (23). [5] Bomse, D.S., Diode Lasers: Finding Trace Gases in the Lab and the Plant, Photonics Spectra, 29(6) (1995).

12 [6] Frish, M. B. and Klein, F., Trace gas monitors based on tunable diode laser technology: an introduction and description of applications, 5th International Symposium on Gas Analysis by Tunable Diode Lasers, Freiburg, Germany, VDI Berichte (1998). [7] Druy, M., Frish, M.B., and Kessler, W.J., From laboratory technique to process gas sensor - the maturation of tunable diode laser absorption spectroscopy, Spectroscopy, 21(3), 14-18, (26). [8] Frish, M. B., et al., The evolution and application of trace gas analyzers based on tunable diode laser absorption spectroscopy, Invited Presentation, 19th International Forum on Process Analytical Chemistry (IFPAC), Washington, DC (25). [9] Frish, M.B., White, M.A., and Allen, M.G., Handheld laser-based sensor for remote detection of toxic and hazardous gases, SPIE Paper No , Water, Ground, and Air Pollution Monitoring and Remediation Conference, Boston, MA. (2). [1] Frish, M.B., Wainner, R.T., Stafford-Evans, J., Green, B.D., Allen, M.G., Chancey, S., Rutherford, J., Midgley, G., and Wehnert, P., Standoff Sensing of Natural Gas Leaks: Evolution of the Remote Methane Leak Detector (RMLD), Invited Paper in Conference on Lasers and Electro-optics/Quantum Electronics and Laser Science and Photonic Applications, Systems and Technologies 25, Optical Society of America, Washington DC (25). [11] Frish, M.B., Laderer, M.C., Wainner, R.T., Wright, A.O., Patel, A.H., Stafford-Evans, J., Morency, J.R., Allen, M.G., Green, B.D., The Next Generation of TDLAS Analyzers, SPIE Paper No , Optics East, Boston, MA (27). [12] Herriot, D., Kogalnik, H., and Kompfner, R., Off-axis Paths in Spherical Mirror Interferometers, Applied Optics 3(4), 523 (1964). [13] Rothman, L.S., et.al., The HITRAN 28 molecular spectroscopic database, Journal of Quantitative Spectroscopy and Radiative Transfer, 11, (29). [14] Bjorklund, G.C., Frequency modulation spectroscopy: a new method for measuring weak absorptions and dispersions, Opt. Lett., 5 (198). [15] Bomse, D.S., Stanton, A.C., and Silver, J.A., Frequency Modulation and Wavelength Modulation Spectroscopies: Comparison of Experimental Methods Using a Lead-Salt Diode Laser, Applied Optics, 31 (1992).