S. C. WALTS ET AL. Extinction Cross-Section Measurements of Bacillus globigii Aerosols Suzanne C. Walts, Craig A. Mitchell, Michael E. Thomas, and Donald D. Duncan In a continuing series of experiments designed to determine the spectral extinction cross section of bacterial aerosols, two White cell transmissometers were constructed to obtain stable, long path length measurements. A laser transmissometer at 543 nm was used to calibrate a broadband transmissometer (400–850 nm). Spectral transmittance was measured as a function of particle concentration, and extinction cross sections were calcu- lated. Visible band measurements of Bacillus globigii aerosols indicated a slight increase in the extinction cross section with increasing wavelength. The extinction cross section was estimated to be 2.58 (±0.25) ϫ 10Ϫ8 cm2 at the 543-nm wavelength. INTRODUCTION The need to detect biological aerosol threats exists in the expression for the range-resolved received power, both in the open atmosphere and within buildings Prec(R), for an aerosol lidar system: and other enclosed spaces. Current sensor technolo- gies include mass spectroscopy, immuno-assays, lidar, −2␤ ()RR c PREe()= ext ⍀ ␴ NR () . (1) and infrared radiometry. However, the need for standoff rec rec b 2 remote sensing and speed for early detection has placed Here, an emphasis on optical techniques. A variety of active E = the source pulse energy, and passive optical sensors are being evaluated that ext = the range-dependent extinction “coeffi - require knowledge of the extinction and volume back- cient” (or the extinction cross section per ␤ scatter cross sections as a function of wavelength, parti- unit volume; ext(R) = CextN(R), where Cext cle size, and chemical makeup (growth media variations is the extinction cross section), ⍀ and different aerosolizing procedures). rec = the solid angle subtended by the receiver It is now being recognized that aerosol lidars will play an telescope, important role in the remote detection of biological agents. b = the backscatter cross section per solid angle Of the remote sensors being considered, aerosol lidars have of the target aerosol, the most mature technology. Information obtained from our N(R) = the range-dependent aerosol number measurements is critical to the design of an optimal lidar concentration, and system. The signifi cance of these measurements is seen c = the speed of light. 50 JOHNS HOPKINS APL TECHNICAL DIGEST, VOLUME 25, NUMBER 1 (2004) EXTINCTION CROSS-SECTION MEASUREMENTS OF Bg AEROSOLS To properly design and characterize the performance constructed within the APL bio-aerosol chamber (Fig. 1). of the lidar system, the extinction coeffi cient and back- Each White cell consists of three identical 6-in.-dia. scatter cross section of various bacterial aerosols as a f/4 mirrors. In the broadband White cell, light from a function of concentration and wavelength must be fi ber-coupled lamp (Cole-Parmer 9741-50) is brought quantifi ed. Knowledge of the spectral dependence of to a focus with a 5ϫ microscope objective next to the these optical properties may allow discrimination of bio- fi eld mirror and allowed to expand to fi ll the fi rst fold logical pathogens from naturally occurring background mirror. It is then focused onto the fi eld mirror and aerosols such as pollen, haze, and dust. expands again to fi ll the second fold mirror. After once Past measurements of the anthrax simulant Bacillus more focusing onto the fi eld mirror, the light repeats globigii (Bg) have been obtained on fi lms1 and in short- this sequence until the last refl ection off the second path, high-concentration aerosols.2,3 Although these fold mirror is focused with a 10ϫ microscope objective are useful measurements, the results are more qualita- onto an optical fi ber that leads to an ultraviolet/near- tive than quantitative. High concentrations tend to infrared spectrometer (StellarNet EPP2000C). Output cause particles to agglomerate into clusters that are not of the lamp is monitored with a silicon photodetector representative of actual distributions in the atmosphere. (Thorlabs PDA55) positioned near the output of the Thus, cross-section measurements are needed on realis- fi ber (not shown). tic aerosol concentration levels that contain single par- Although the path length within the broadband ticles. Unfortunately, the transmission losses for propa- transmissometer is signifi cantly longer than that in gation through such distributions are low, complicating previous measurements, it is still desirable to calibrate the measurements. the broadband measurement with a laser transmissom- The tactic we have chosen for improving the mea- eter measurement. One reason for doing so is that the surement precision is to use very long path lengths. Since laser source offers signifi cantly higher brightness and a the required straight-line path lengths are impractical smaller focused spot on the fi eld mirror. Thus, longer in a laboratory setting, we have elected to “fold” these path lengths can be achieved than with broadband measurement paths with a device known as a White cell sources. Another reason is that, to suppress background (named after the designer, J. U. White4). noise, the beam must be mechanically chopped and lock-in techniques used to measure the output power of the laser and the signal transmitted through the White EXPERIMENTAL CONFIGURATION cell. Since this cannot be accomplished easily with the In previous measurements of the spectral cross sec- current broadband system (the required integration tion of Bg at APL,5 a short path length and an unsta- time of the spectrometer does not allow the use of a ble platform contributed to unacceptable noise levels modulated source), simultaneous laser measurements for the broadband transmissometer. To address these (with a HeNe laser at 543 nm) are made in a second shortcomings, broadband and laser White cells were White cell. Fold mirrors Fold mirrors Fan Agent introduction Test chamber APS Broadband Field mirror Si detector Field mirror source Spectrometer HeNe laser Fan Fan Broadband White cell Laser White cell Figure 1. Experimental confi guration in the APL bio-aerosol chamber. Particle size distribution and concentration are measured with an aerodynamic particle sizer (TSI Inc. Model APS 3320). (Figure not to scale.) JOHNS HOPKINS APL TECHNICAL DIGEST, VOLUME 25, NUMBER 1 (2004) 51 S. C. WALTS ET AL. RESULTS linear fi t to these data. The extinction cross section is The broadband and laser transmissometers were oper- the slope of this curve as indicated by the relationship ated simultaneously to measure the transmittance as the aerosol concentration was varied. Each laser transmit- ␤ ext tance value at 543 nm was calculated according to the C = , (4) ext N formula VV/ where N is the particle number density. For this set of T = sig ref , measurements, the extinction cross section was esti- 00 (2) Ϫ8 2 VVsig/ ref mated to be 2.58 (±0.25) ϫ 10 cm . The observed variation could be caused by changes in the chamber’s where V is the White cell signal detector voltage and environmental condition (e.g., humidity), sample prepa- sig ration, sample dissemination, and so on. Variations on a Vref is the laser reference detector voltage. Before intro- ducing the aerosolized particles into the chamber, base- particular day are often within ±2%. 00 Results from the laser transmissometer were used to line voltages (and)VVsig ref were recorded. All detector values were obtained by averaging data points collected calibrate the broadband laser transmissometer measure- every 0.7 s for 1 min. For each particle concentration ments. Figure 3 shows the spectral transmittance for dif- level, quantities of Bg were introduced into the chamber ferent particle concentration levels. The laser calibra- and the mixture was allowed to stabilize for several min- tion points at 543 nm are indicated by black boxes. Note utes. Fans distributed the aerosol uniformly throughout that the transmittances are in the single scatter regime the chamber. (see the boxed insert). The single-wavelength extinction coeffi cient was Several experimental spectrometer and laser spectral then estimated by using the relationship cross sections are plotted in Fig. 4, representing data over a 1-year period. Also shown is the corresponding Mie theory prediction.7 For this calculation, the bulk index 1 of refraction was obtained from Tuminello et al.1 Even ␤ =− ln(T ) , ext L (3) though the computed extinction cross section is scaled by a factor of 1.2, agreement with the computed spectral where L is the path length of the transmissometer. For shape is remarkably good. These spectra indicate little this test, a path length of 90 m was achieved with the wavelength dependence throughout the visible. How- White cell. Figure 2 shows the extinction coeffi cient as a ever, the cross section is expected to change dramati- function of Bg concentration. The measured data points cally as the wavelength increases, beginning in the near are indicated in color, and the black curve represents a infrared. Cross sections also depend on the particle size dis- tribution of the aerosolized bacteria. Figure 5 shows a 0.0030 representative size distribution for Bg particles released into the aerosol chamber as measured by the APS. The 0.0025 0.0020 105 205 ) 0.95 3 307 Ϫ 0.0015 396 0.90 510 0.0010 615 Extinction coefficient (1/m) 714 0.85 807 (cm Concentration 0.0005 Transmittance 903 0.80 0 0200 400 600 800 1000 0.75 Particle concentration (1/cm3) 480 520 560 600 640 680 Wavelength (nm) Figure 2. Experimental Bg extinction coeffi cient collected on two different dates (indicated in red and green) at 543 nm versus par- Figure 3. Transmittance versus wavelength for varying concen- ticle concentration. The line is the least-squares fi t to the data and trations of Bg. Curves are broadband spectrometer results and ϫ Ϫ8 2 its slope is the extinction cross section, Cext = 2.58 10 cm .