Final Report Daad13 02 C 0015 Part5 App G K

SPIE-2004-5585 46

Characterization of chemical warfare G-agent hydrolysis products by surface-enhanced Raman spectroscopy

Frank Inscore, Alan Gift, Paul Maksymiuk, and Stuart Farquharson*

Real-Time Analyzers, 87 Church Street, East Hartford, CT 06108

ABSTRACT The United States and its allies have been increasingly challenged by terrorism, and since the September 11, 2001 attacks and the war in Afghanistan and Iraq, homeland security has become a national priority. The simplicity in manufacturing chemical warfare agents, the relatively low cost, and previous deployment raises public concern that they may also be used by terrorists or rogue nations. We have been investigating the ability of surface-enhanced Raman spectroscopy (SERS) to detect extremely low concentrations (e.g. part-per-billion) of chemical agents, as might be found in poisoned water. Since trace quantities of nerve agents can be hydrolyzed in the presence of water, we have expanded our studies to include such degradation products. Our SERS-active medium consists of silver nanoparticles incorporated into a sol-gel matrix, which is immobilized in a glass capillary. The choice of sol-gel precursor allows controlling hydrophobicity, while the porous silica network offers a unique environment for stabilizing the SERS-active silver particles. Here we present the use of these silver-doped sol-gels to selectively enhance the Raman signal of the hydrolyzed products of the G-series nerve agents. Keywords: chemical warfare agent detection, CWA, hydrolysis, SERS, Raman spectroscopy

1. INTRODUCTION The potential use of chemical and biological warfare agents by terrorist organizations directed against U.S. military and Coalition forces in the Middle East, and civilians at home, is an issue that has generated considerable concern in the post 9/11 era. The ability to counter such attacks, requires recognizing likely deployment scenarios, among which includes poisoning water supplies with chemical warfare agents (CWAs). The G-series nerve agents are a particular concern due to their extreme toxicity (LD50 man for GB = 25 mg/kg, GD = 5 mg/kg, GF = 5mg/kg ),1 persistence (hydrolysis half-life of 1-3 days),2 relatively high solubility (5-25 g/L, see Table 1), and their previous use in Iraq3 and Japan.4 The nerve agents, isopropyl methylphosphonofluoridate (GB), pinacolyl methylphosphonofluoridate (GD), and cyclohexyl methylphosphonofluoridate (GF) initially hydrolyze to isopropyl methylphosphonic acid (IMPA), pinacolyl methylphosphonic acid (PMPA), and cyclohexyl methylphosphonic acid (CMPA), respectively, and subsequently, at a much slower rate, to a common final, stable product methylphosphonic acid (MPA, see Figure 1).5,6 Clearly any analysis designed to detect nerve agents in poisoned water must not only be able to detect µg/L concentrations,7 but also be able to detect and distinguish the resultant hydrolysis products. In addition, the ability to quantify the relative amounts of the initial agent and its hydrolysis products would provide a means to estimate when the water supply was poisoned. It is also worth noting that an analyzer capable of measuring these hydrolysis products at such low concentrations would also be valuable in establishing prior presence of nerve agents through soil and groundwater analysis,8,9 verify successful destruction during decommissioning operations,5,10,11 and establishing extent of exposure during an attack.12 Several technologies have recently been investigated as potential at-site analyzers for chemical agents, as well as their hydrolysis products.6,13 This includes liquid chromatography combined with mass spectrometry (LC/MS),9,14-17 infrared spectroscopy18,19,20 and Raman spectroscopy (RS).21 However, LC/MS remains a labor intensive technique, infrared is limited by the strong absorption of water which obscures much of the spectrum, while Raman spectroscopy does not have sufficient sensitivity.21 In the past few years, we and others have explored the potential of surface-enhanced Raman spectroscopy (SERS) to detect CWAs,22-28 and their degradation products.29 The utility of SERS is based upon the extreme sensitivity of this technique and the ability to identify molecular structure through the abundant vibrational information provided by Raman spectroscopy. SERS employs the interaction of surface plasmon modes of metal particles with target analytes to increase scattering efficiency by as much as 1 million times.30

SPIE-2004-5585 47

In our studies, we have employed metal-doped sol-gels to promote the SERS effect. The porous silica network of the alkoxide sol-gel matrix offers a unique environment for immobilizing and stabilizing SERS-active metal particles of both silver and gold.31-34 The choice of metal and Si-alkoxide composition provides a means for chemically selecting the target analyte to be enhanced based on charge and polarity. Electropositive silver or electronegative gold particles can selectively enhance the Raman signals of negative or positive chemical species, respectively, while different alkoxides (or combinations of) can be used to select for polar or non-polar molecules. Previously, we used glass vials internally coated with the SERS-active sol-gel to measure cyanide, HD, VX, and MPA.28 More recently, we have developed glass capillaries filled with the SERS-active sol-gel that can be attached to a syringe to perform simple and rapid sample extraction and SERS analysis.35 This paper employs these extractive and SERS-active capillaries to examine the ability of SERS to measure and distinguish the hydrolysis products of GB, GD, and GF. Both Raman and surface-enhanced Raman spectra are presented along with preliminary vibrational mode assignments. Table 1. Properties of chemical agents and their primary hydrolysis products investigated in the present study.2

Chemical Agent Hydrolysis ½ life Water Solubility at 25°C Sarin (GB) 39 hr (pH 7) completely miscible IMPA stable (can hydrolyze to MPA) 4.8 g/L MPA very stable (resistant to further degradation) >1000 g/L Soman (GD) 45 hr (pH 6.6) 21 g/L (@20°C) PMPA stable (can hydrolyze to MPA) no data Cyclosarin (GF) slower than GB 3.7 g/L CMPA no data (can hydrolyze to MPA) no data

PO

O

PO

O

PO

O

PO

O

PO

O

PO

O

F

F

F

GB

GD

GF

OH

OH

HF

HF

HF

2-propanol

2-pinacolyl

cyclohexanol

MPA

MPA

MPA

IMPA

PMPA

CMPA

H2O

H2O

H2O

+

+

+

+

+

+

OH

H2O

H2O

H2O

Figure 1. Hydrolysis pathways for G-Series nerve agents.

2. EXPERIMENTAL The hydrolysis degradation chemicals measured in this study (IMPA, PMPA, CMPA) were obtained as analytical reference materials from Cerilliant (Round Rock, TX) and used without further purification. MPA and all chemicals used to prepare the silver-doped sol-gel coated capillaries were acquired from Sigma-Aldrich (St. Louis, MO) and used as received. For the purpose of safety, samples were prepared in a chemical hood, transferred to a sampling device and sealed prior to being measured. All samples were measured initially by Raman in their pure state at room temperature; MPA as a solid powder, with IMPA, and PMPA as neat liquids. CMPA was obtained in forensic quantities (1 mg/mL in MeOH), and was not amenable to RS studies at these concentration levels. Methanol or water (HPLC grade) was used to prepare solutions of the target chemicals for SERS measurements at a

SPIE-2004-5585 48

concentration of 1 mg/mL from solid powders or 0.1% v/v from neat liquids unless noted otherwise. Lower concentrations were prepared from these solutions by serial dilution, and all solutions were stored at 10°C until needed. The Raman and SERS spectra of the target chemicals presented here were all measured in capillaries. SERS-active capillaries were prepared using the following general methodology. A silver-doped sol-gel solution, prepared according to previous published procedures from a mixture of two precursor solutions,31 was drawn via a syringe into pre-cleaned 1-mm diameter capillaries. This procedure was modified for the SERS-active capillaries, in particular by replacing TMOS with an alkoxide mixture composed of tetramethyl orthosilicate (TMOS), octadecyltrimethoxysilane (ODS), and methyltrimethoxysilane (MTMS) at a v/v/v ratio of 1/1/5. A 50 µL sample from each of the prepared analyte solutions was drawn into a SERS-active capillary for measurement. The capillaries were mounted horizontally on an XY positioning stage (Conix Research, Springfield, OR), such that the focal point of an f/0.7 aspheric lens was positioned just inside the glass wall. The probe optics and fiber optic interface have been described previously.35 A Fourier transform Raman spectrometer (Real-Time Analyzers, model IRA-785, East Hartford, CT) equipped with a 785 nm diode laser (Process Instruments Inc. model 785-600, Salt Lake City, UT) and a silicon photo-avalanche detector (Perkin Elmer model C30902S, Stamford, CT) was used to deliver 100 mW of power to the SERS and RS samples and generate spectra with 8 cm-1 resolution.

3. RESULTS AND DISCUSSION

The SERS spectra of chemicals are often different than their Raman spectral counterparts due to the surface interactions that can enhance various vibrational modes to different extents. Therefore the Raman spectra were measured and included in this study to aid interpretation of the corresponding SERS spectra. The simplest chemical specific to the G series nerve agents is methylphosphonic acid, which has been well characterized by IR and Raman spectroscopy,36,37 and subsequent normal coordinate analysis for assigning the vibrational modes.38 The Raman spectrum of MPA contains 10 discernable peaks between 350 and 1650 cm-1 (Figure 2B). Four PO3 bending modes are observed at 408, 462, 491 (shoulder) and 504 cm-1. The PC symmetric stretch is the most intense peak observed at 774 cm-1. A CH3 rocking mode occurs at 892 cm-1 with little intensity, while the PO3 stretching mode produces a peak to 956 cm-1. Two additional CH3 and PO3 modes produce peaks at 1004 and 1054 cm-1, also with moderate intensity. The 10th mode in this region is a CH3 bending mode which occurs at 1424 cm-1.

The SERS spectrum of MPA (Figure 2A) is considerably simpler than that of the solid powder Raman spectrum, with weak peaks observed at 469, 521, 958, 1003, 1038, and 1420 cm-1. These SERS spectral peaks can all be assigned to the modes observed at similar frequencies in the Raman spectrum, albeit the 521 and 1038 cm-1 peaks have shifted significantly from the 504 and 1054 cm-1 Raman spectral peaks. The most characteristic SERS spectral peaks are the

Figure 2. A) SERS and B) Raman spectra of MPA. Conditions: A) 0.1 mg/ml in water, TMOS/ODS/MTMS sol-gel in capillary, 1-min acquisition time. B) solid, 5-min acquisition time.

A B

A B

Figure 3. A) SERS and B) Raman spectra of IMPA. Conditions as in Fig. 2, but: A) 0.1 % v/v in MeOH, B) neat liquid.

SPIE-2004-5585 49

intense 756 cm-1 peak and the unique peak at 1300 cm-1. The former peak clearly corresponds to a nearly pure PC symmetric stretch, while the latter is likely a CH3 twist. The next hydrolysis product studied was isopropyl methylphosphonic acid. Like MPA, both the Raman and SERS spectra of IMPA are dominated by a peak in the 700 cm-1 region, specifically at 728 and 716 cm-1, respectively (Figure 3). However, these peaks are not simply a PC stretch, but include a considerable amount of the backbone CPOCC mode created by the addition of the isopropyl group. Both spectra also contain moderate peaks at 782 and 772 cm-1 that may also be PC containing backbone modes, as has been suggested by a theoretical treatment for sarin.39 It is also worth noting that the Raman spectrum of IMPA is very similar to that of a published spectrum of sarin.21 A number of the peaks assigned to PO3 modes for MPA have shifted moderately from the Raman to the SERS spectra for IMPA, and includes the following respective peaks; 510 and 508 cm-1, 938 and 931 cm-1, and 1006 and 1004 cm-1. The latter peak likely contains significant methyl character. Similarly, the methyl rocking and bending modes observed for MPA are now at 880 and 874 cm-1, and 1420 and 1416 cm-1 in the respective Raman and SERS spectra of IMPA. Not surprisingly, the isopropyl group not only increased the intensity of these bands, but also gives rise to a CH deformation, and additional CH3 and CH2 wagging modes, at 1359 and 1349 cm-1, 1390 and 1388 cm-1 and 1453 and 1451 cm-1, in the respective Raman and SERS spectra. The isopropyl group also gives rise to a CC bend at 421 and 424 cm-1, and a CC stretch at 1179 and 1173 cm-1 in the respective Raman and SERS spectra. In the Raman spectrum of IMPA a peak also appears at 1104 cm-1 that is characteristic of CO or CC stretches, while in the SERS spectrum a peak appears at 1055 cm-1 and is assigned to a PO3 stretch, as was the 1038 cm-1 peak in the MPA SERS spectrum. The Raman spectrum of pinacolyl methylphosphonic acid, like IMPA, contains an increasing amount of CC and CHn character (Figure 4B). This includes new peaks at 541, 934, 977, 1212 and 1264 cm-1 that are assigned to a CC3 wag, a CC3 bend, a CCC bend, and two CC stretching modes based on a theoretical treatment for soman.39 The 1300 to 1500 cm-1 region again contains a number of CHn bending modes, and the peaks are assigned accordingly. The most obvious change in the spectrum is that the PC plus backbone mode in the IMPA spectrum has split into two distinct peaks at 732 and 761 cm-1. The SERS spectrum for PMPA is dominated by these latter peaks, except that they overlap considerably producing a peak centered at 750 cm-1 with a shoulder at 729 cm-1 (Figure 4A). The remaining SERS peaks are evident, but have little intensity, except for the CC3 wag at 543 cm-1, the PO3 stretch at 1037 cm-1, and the CH2 bend at 1444 cm-1. Cyclohexyl methylphosphonic acid was only available as 1 mg/mL in methanol and a Raman spectrum at this concentration could not be obtained. The SERS spectrum in many ways is like IMPA with the addition of cyclohexane modes (Figure 5). This includes peaks at 622, 1023, and 1262 cm-1, that are attributed to ring CC stretching modes, and a peak at 811 cm-1 that is assigned to a ring CH2 bending mode. The most intense peak observed at 747 cm-1 is again assigned to a PC stretch plus backbone mode.

In general, the SERS spectra for these alkyl methylphosphonic acids have two common features, the PC stretch produces the most intense peak, more so than the Raman spectra when compared to the intensity of the other peaks, and the most

Wavenumber (cm-1) Figure 5. SERS spectrum of CMPA. Conditions as in Fig. 3, but A) 1 mg/mL in MeOH.

Figure 4. A) SERS and B) Raman spectra of PMPA. Conditions as in Fig. 3.

A B

SPIE-2004-5585 50

substantial shift in peak frequencies occurs for PO3 modes when compared to the Raman spectra. The increased intensity of the PC mode suggests that it is perpendicular to the surface, based on previous research that has shown that modes couple to the plasmon field more effectively in this orientation.40 The shift in the PO3 frequencies suggests strong surface interactions through this group. Taken together, the SERS data suggests that these molecules are oriented with the PO3 group interacting with the silver surface and the methyl group away from the surface. In the case of MPA, especially for the doubly deprotonated anion, the three oxygens could form the base of a tripod on the surface. This orientation may become less likely for the other molecules as the alkoxide groups replace the hydroxide group with surface interaction through the other two oxygens. This change in orientation along with increasing amounts of backbone character to the PC stretch could explain the shift and splitting of this mode. Table 2. Tentative vibrational mode assignments for Raman and SERS peaks for VX and its hydrolysis products.

MPA IMPA PMPA CMPA Tentative Assignmentsa RS SERS RS SERS RS SERS SERSb

408 421 424 PO3 bend 462c,d 469 441 442 441 PO3 bend 491c 475 PO3 bend 504c 521 510 508 514 495 C-PO3 bend 541e 543 549 C-C3 bend 622 Ring breathing 728 716 732 729sh PC stretch and backbone 774 756 782 772 761 750 747 PC stretch and backbone 799 792 CH bend 811 Ring CH2 880e 874 869e 863 857 CCC bend 892c,d 902 888 896 CH3 rock 934e 929 C-C3 bend 956c,d 958 938 931 PO3 stretch 977e CCC stretch 1004 1003 1006 1004 1015 1000 PO3 or CH3 bend 1023 Ring breathing sym 1054 1038d 1055 1052 1037 1050 PO3 stretch 1079 1073 CCC bend 1104 1116 OC or CC stretch 1143 1132 1150 CC stretch 1179 1173 1212e 1206 CC stretch 1224 1236 1243 CH2 bend or above 1264e 1257 1262 CC stretch 1300 1291 CH3 bend 1359 1349 1355 1347 CH deformation 1374 CHn bend 1390 1388 1390 1394 1393 CH3 rock 1424c,d 1420 1420 1416 1420 1415 1416 CH3 bend (bound to P) 1453 1451 1447 1444 1443 CH2 rock

a - Assignment terminology is simplified since assignments refer to multiple molecules. b - no Raman spectrum measured, c = Ref. 36, d = Ref. 37, e = Ref. 39.

4. CONCLUSION The ability to measure and identify the various hydrolysis degradation products with our SERS-active silver-doped sol- gel coated capillaries has been demonstrated. The SERS spectra of these chemicals were somewhat different than their Raman spectral counterparts, which is attributed to the interaction of these chemicals with the silver. In general, the Raman and SERS spectra for the alkyl methylphosphonic acid hydrolysis products were dominated by one or two peaks between 715 and 765 cm-1, which have been assigned to PC stretching modes with varying amounts of backbone mode

SPIE-2004-5585 51

contributions. The spectral intensity of this mode and the shift in frequency of the PO3 modes in the SERS spectra suggest a strong surface interaction for these molecules. It is clear from the present study that the hydrolysis products can easily be identified as a class by these 700 cm-1 peaks, but quantifying each in a mixture is likely to require chemical separations or chemometric approaches. These approaches, as well as measurements to determine the detection limits and pH dependence of these hydrolysis products are in progress.

5. ACKNOWLEDGMENTS The authors are grateful for the support of the U.S. Army (DAAD13-02-C-0015, Joint Service Agent Water Monitor program), and the National Science Foundation (DMI-0215819), and would like to thank Dr. Steve Christesen for helpful discussions, and Mr. Chetan Shende for sol-gel chemistry development.

6. REFERENCES 1. Committee on Toxicology. Review of Acute Human-Toxicity Estimates for Selected Chemical-Warfare Agents,

Nat. Acad. Press (Washington, D.C.) 1997. 2. Munro, N.B., Talmage, S.S., Griffin, G.D., Waters, L.C., Watson, A.P., King, J.F., and Hauschild V. “The Sources,

Fate, and Toxicity of Chemical Warfare Agent Degradation Products”, Environ. Health Perspect., 107, 933-974 (1999).

3. Hoenig, S.L. Handbook of Chemical Warfare and Terrorism, Greenwood Press (Westport, CT) 2002. 4. Nozaki, H. and Aikawa, N. “Sarin poisoning in Tokyo subway”, Lancet, 345 1446-1447 (1995). 5. Wagner, G. and Yang, Y. “Rapid nucleophilic/oxidative decontamination of chemical warfare agents”, Ind. Eng.

Chem. Res., 41, 1925-1928, (2002). 6. Creasy, W., Brickhouse, M., Morrissey, K., Stuff, J., Cheicante, R., Ruth, J., Mays, J., Williams, B., O’Connor, R.,

and Durst, H. “Analysis of chemical weapons decontamination waste from old ton containers from Johnston atoll using multiple analytical methods”, Environ. Sci. Technol., 33, 2157-2162, (1999).

7. McKone, T.E., Huey, B.M., Downing, E., and Duffy, L.M., Editors. Strategies to Protect the Health of Deployed U.S. Forces: Detecting, Characterizing, and Documenting Exposures, Nat. Acad. Press (Washington, D.C.) p.207, 1999.

8. Johnston, R.L., Hoefler, C.M., Fargo, J.C., and Moberley, B. “The Defense Nuclear Agency’s Chemical/ Biochemical Weapons Agreements Verification Technology Research, Development, Test and Evaluation Program and its Requirements for On-Site Analysis”, AT-ONSITE, 5-8 (1994).

9. D’Agustino, P.A, Hancock, J.R., and Provost, L.R. “Determination of sarin, soman and their hydrolysis products in soil by packed capillary liquid chromatography-electrospray mass spectrometry”, J. Chromatography A, 912, 291-299 (2001).

10. Yang, Y., Baker, J., and Ward, J. “Decontamination of chemical warfare agents”, Chem. Rev., 92, 1729-1743 (1992).

11. Christesen, S., MacIver, B., Procell, L., Sorrick, D., Carrabba, M., and Bello, J. “Nonintrusive analysis of chemical agent identification sets using a portable fiber-optic Raman spectrometer”, Appl. Spec., 53, 850-855 (1999).

12. Hui, D.-M. and Minami, M. “Monitoring of fluorine in urine samples of patients involved in the Tokyo sarin disaster, in connection with the detection of other decomposition products of sarin and the by-products generated during sarin synthesis”, Clin. Chim. Acta, 302, 171-188 (2000).

13. “The Chemical Weapons Convention Redefines Analytical Challenge”, Analytical Chemistry News & Features, June 1, 397A (1998).

14. Sega, G.A., Tomkins, B.A., and Griest, W.H. “Analysis of methylphosphonic acid, ethyl methylphosphonic acid and isopropyl methylphosphonic acid at low microgram per liter levels in groundwater” J. Chromatography A, 790, 143-152 (1997).

15. Creasy, W.R. “Postcolumn Derivatization Liquid Chromatography/Mass Spectrometry for Detection of Chemical-Weapons-Related Compounds” Am. Soc. Mass Spectrom., 10, 440-447 (1999).

16. Katagi…, J. Chromatography A, 833, 169-179 (1999). 17. Liu, Q., Hu, X., and Xie, J. “Determination of nerve agent degradation products in environmental samples by liquid

chromatography–time-of-flight mass spectrometry with electrospray ionization”, Analytica Chimica Acta, 512, 93-101 (2004).

SPIE-2004-5585 52

18. Hoffland, L.D., Piffath, R.J., and Bouck, J.B. “Spectral signatures of chemical agents and simulants”, Optical

Engineering, 24, 982-984, (1985). 19. Braue, E.H.J., and Pannella, M.G. “CIRCLE CELL FT-IR Analysis of Chemical Warfare Agents in Aqueous

Solutions”, Applied Spectroscopy, 44, 1513-1520, (1990). 20. Kanan, S. and Tripp, C. “An infrared study of adsorbed organophosphonates on silica: a prefiltering strategy for the

detection of nerve agents on metal oxide sensors”, Langmuir, 17, 2213-2218, (2001). 21. Christesen, S.D. “Raman cross sections of chemical agents and simulants”, Appl. Spec., 42, 318-321 (1988). 22. Lee, Y. and Farquharson, S. “Rapid chemical agent identification by SERS”, SPIE, 4378, 21-26 (2001). 23. Farquharson, S., Maksymiuk, P., Ong, K., and Christesen, S. “Chemical agent identification by surface-enhanced

Raman spectroscopy”, SPIE, 4577, 166-173 (2001). 24. Spencer, K.M., Sylvia, J., Clauson, S. and Janni, J. “Surface Enhanced Raman as a Water Monitor for Warfare

Agents in Water”, SPIE, 4577, 158-165 (2001). 25. Premasiri, W., Clarke, R., Londhe, S., and Womble, M. “Determination of cyanide in waste water by low-resolution

surface enhanced Raman spectroscopy on sol-gel substrates”, J. Ram. Spec., 32, 919-922 (2001). 26. Tessier, P., Christesen, S., Ong, K., Clemente, E., Lenhoff, A., Kaler, E., and Velev, O. “On-line spectroscopic

characterization of sodium cyanide with nanostructured Gold surface-enhanced Raman spectroscopy substrates”, App. Spectrosc., 56, 1524-1530 (2002).

27. Christesen, S.D., Lochner, M.J., Ellzy, M., Spencer, K.M., Sylvia, J., and Clauson, S. “Surface Enhanced Raman Detection and Identification of Chemical Agents in Water”, 23rd Army Science Conf., Orlando, 2002.

28. Farquharson, S., Gift, A., Maksymiuk, P., Inscore, F., Smith, W., Morrisey, K., and Christesen, S. “Chemical agent detection by surface-enhanced Raman spectroscopy”, SPIE, 5269, 16-22 (2004).

29. Farquharson, S., Gift, A., Maksymiuk, P., Inscore, F., and Smith, W. “pH dependence of methyl phosphonic acid, dipicolinic acid, and cyanide by surface-enhanced Raman spectroscopy”, SPIE, 5269, 117-125 (2004).

30. Weaver, M.J., Farquharson, S., and Tadayyoni, M.A. “Surface-enhancement factors for Raman scattering at silver electrodes”, J. Chem. Phys., 82, 4867-4874 (1985).

31. Lee, Y. and Farquharson, S. “SERS Sample Vials Based on Sol-Gel Process for Trace Pesticide Analysis”, SPIE, 4206, 140-146 (2000).

32. Farquharson, S. and Lee, Y. “Trace Drug Analysis by Surface-Enhanced Raman Spectroscopy”, SPIE, 4200-16 (2000).

33. Lee, Y., Farquharson, S., and Rainey, P.M. “Surface-Enhanced Raman Sensor for Trace Chemical Detection in Water”, SPIE, 3857, 76-84 (1999).

34. Lee, Y, Farquharson, S., Kwong, H., and Shahriari, M. “Surface-Enhanced Raman Sensor for Surface-Enhanced Raman Spectroscopy”, SPIE, 3537, 252-260 (1998).

35. Farquharson, S., Gift, A., Maksymiuk, P., and Inscore, F. “Rapid dipicolinic acid extraction from Bacillus spores detected by surface-enhanced Raman spectroscopy”, Appl. Spec. 58, 351-354 (2004).

36. Nyquist, R. “Vibrational spectroscopic study of (R-PO3)2¯”, J. Mol. Struct., 2, 123-135, (1968). 37. Van der Veken, B.J. and Herman, M.A. “Vibrational analysis of methylphosphonic acid and its anions: I.

Vibrational spectra”, J. Molec. Struct., 15, 225-236 (1973). 38. Van der Veken, B.J. and Herman, M.A. “Vibrational analysis of methylphosphonic acid and its anions: II. Normal

coordinate analysis”, J. Molec. Struct., 15, 237-248 (1973). 39. Hameka, H. and Jensen, J. “Theoretical prediction of the infrared spectra of nerve agents”, CRDEC-TR-326, 1992. 40. Suh, J.S. and Moskovitz, M. “SERS of amino acids and nucleotide bases adsorbed on silver” J. Am. Chem. Soc.

108, 4711-4718 (1986).

Applied Spectroscopy, 59, 2005 654

Surface-enhanced Raman spectra of VX and its hydrolysis products STUART FARQUHARSON,∗ ALAN GIFT, PAUL MAKSYMIUK, AND FRANK INSCORE Real-Time Analyzers, East Hartford, CT 06108

∗ Author to whom correspondence should be sent.

Detection of chemical agents as poisons in water supplies, not only requires µg/L sensitivity, but also requires the ability to distinguish their hydrolysis products. We have been investigating the ability of surface-enhanced Raman spectroscopy (SERS) to detect chemical agents at these concentrations. Here we expand these studies and present the SERS spectra of the nerve agent VX (ethyl S-2-diisopropylamino ethyl methylphosphonothioate) and its hydrolysis products; ethyl S-2-diisopropylamino methylphosphonothioate, 2-(diisopropylamino) ethanethiol, ethyl methylphosphonic acid, and methylphosphonic acid. Vibrational mode assignments for the observed SERS peaks are also provided. Overall, each of these chemicals produces a series of peaks between 450 and 900 cm-1 that are sufficiently unique to allow identification. SERS measurements were performed in silver-doped sol-gel filled capillaries that are being developed as part of an extractive point sensor. INTRODUCTION In the post 9/11 era the use of chemical and biological warfare agents by terrorist organizations directed against U.S. and Coalition forces in Afghanistan and Iraq, as well as civilians at home is an undeniable possibility. Countering future attacks requires recognizing likely deployment scenarios, among which includes poisoning of water supplies. In this instance, the nerve agent ethyl S-2-diisopropylamino ethyl methylphosphonothioate (VX) is of particular concern, because in addition to an oral LD50 of 0.012 mg/kg in humans, it is reasonably soluble (150g/L), and somewhat persistent with a hydrolysis half-life greater than 3 days.1 Furthermore, one of its hydrolysis products, ethyl S-2-diisopropylamino methylphosphonothioate (EA2192), is considered just as deadly, more soluble and more persistent (Table I).2 In fact, VX can hydrolyze according to two different pathways (Fig. 1, Reaction Pathways 1 and 2).3,4 In one case, 80% of VX is converted to 2-(diisopropylamino) ethanethiol (DIASH), which is stable in water, and ethyl methylphosphonic acid (EMPA), which further hydrolyzes to form methylphosphonic acid (MPA) and ethanol. In the other case, 20% of VX is converted to EA2192 and ethanol, and as previously indicated, EA2192 eventually hydrolyzes and forms DIASH and MPA. Previously, we5-8and others 9-11 reported the surface-enhanced Raman spectra of VX, EA2192, and MPA as preliminary data to demonstrate the potential of developing a portable analyzer capable of measuring µg/L concentrations of chemical agents in less than 10 minutes. The expected success of surface-enhanced Raman spectroscopy (SERS) is based on the enormous increase in Raman scattering efficiency when a

Table I. Hydrolysis half-lifea and water solubilityb for VX and its primary hydrolysis products.

Chemical Agent Hydrolysis Half-life Water Solubility VX >3 days (pH 7) 150 g/L EA2192 > 10 x VX ∞ sol. DIASH stable ca. 1000 g/L EMPA >8 days 180 g/L MPA very stable >1000 g/L

a = Ref. 1, b = Ref. 2, c at 25°C

molecule interacts with the surface plasmon modes of metal nanoparticles, such as gold or silver,12 which will provide the necessary sensitivity. Typical enhancements on the order of 1 million times have been reported for MPA,6 and calculated limits of detection (LOD) at 50 to 100 µg/L,8,9 are close to the required 10 µg/L LOD for nerve agents in water.13 The expected success of SERS is also based on the unique set of Raman spectral peaks due to the specific molecular vibrations of each chemical that will allow unequivocal identification of the nerve agents and their hydrolysis products. Towards fulfilling this second expectation, we have measured the SERS spectra of VX and its hydrolysis products; EA2192, DIASH, EMPA, and MPA, and provide preliminary vibrational mode assignments. In this study, a silver-doped sol-gel has been incorporated into a glass capillary to both chemically extract the target analytes and promote the SERS effect.14 EXPERIMENTAL DIASH and EMPA were obtained as analytical reference materials from Cerilliant (Round Rock, TX) and used without further purification. MPA and all chemicals used to prepare the silver-doped sol-gel coated capillaries were acquired from Sigma-Aldrich (St. Louis, MO) and also used as received. For the purpose of safety, all samples were prepared in a chemical hood, transferred to a capillary and sealed prior to being measured. The Raman spectra of VX and EA2192 were measured as a pure liquid and a pure solid, respectively at the U.S. Army’s Edgewood Chemical Biological Center. The Raman spectra of EMPA was measured as a pure liquid, while both DIASH and MPA were measured near the point of saturation as 1 g/mL in HPLC grade water samples. In the case of surface-enhanced Raman spectral measurements, EMPA was prepared as 0.1% v/v in methanol, DIASH as 1 mg/mL in methanol, VX as 1% v/v in water, MPA as 0.1 mg/mL in water, and EA2192 as 1 mg/mL in water. VX and EA2192 were measured in 2-ml glass vials internally coated with a layer of silver-doped sol-gel (Real-Time Analyzers, Simple SERS Sample Vials, East Hartford, CT), while MPA, EMPA, and DIASH were measured in 1-mm diameter glass

Applied Spectroscopy, 59, 2005 655

FIG. 1. Hydrolysis pathways for VX.3,4

capillaries filled with silver-doped sol-gel. The latter were prepared according to previously published methods,15 except for the following modification: the alkoxide, tetramethyl orthosilicate (TMOS), was replaced by an alkoxide mixture composed of TMOS, methyltrimethoxysilane (MTMS), and octadecyltrimethoxysilane (ODS) in a v/v/v ratio of 1/1/5. This latter alkoxide combination produced a more non-polar sol-gel that better extracted the MPA, EMPA, and DIASH from the solvent. Both SERS-active sampling devices were mounted horizontally on an XY positioning stage (Conix Research, Springfield, OR), such that the focal point of an f/0.7 aspheric lens was positioned just inside the glass wall. The probe optics and fiber optic interface have previously been described.15 In all cases a 785 nm diode laser (Process Instruments Inc. model 785-600, Salt Lake City, UT) was used to deliver ~100 mW of power to the SERS samples and 100 to 300 mW to the Raman spectroscopy samples. A Fourier transform Raman spectrometer (Real-Time Analyzers, model IRA-785) equipped with a silicon photo-avalanche detector (Perkin Elmer model C30902S, Stamford, CT) was used to collect both the Raman and SERS spectra at 8 cm-1 resolution and at 5-min and 1-min acquisition times, respectively, except in the case of the Raman spectra of VX and EA2192. These two measurements, performed at Aberdeen, used a 785 nm diode laser to deliver 100 to 150 mW to the sample. A dispersive spectrometer and a silicon-based CCD detector were used to acquire 1 cm-1 resolution spectra in 1-min acquisitions (InPhotonics, Norwood, MA).16 All samples were measured within 1 hour of preparation to ensure minimum hydrolysis. Only in the case of VX, with the shortest hydrolysis half-life, would any significant product form in this time frame (< 1%). Furthermore, once the samples were introduced into the vials or capillaries they were measured within 10 minutes. For the vials, this appears to be sufficient time for the sample to diffuse through the sol-gel to the silver surface, as no time dependence was observed for the spectra. For the capillaries, the sample is drawn through the sol-gel minimizing the amount of diffusion required to reach equilibrium, and again no time dependence was observed for the spectra.

RESULTS AND DISCUSSION

The assignment of SERS peaks to vibrational modes is less straightforward than for Raman spectral peaks due to the metal-to-molecule surface interactions that shift and enhance various modes to different extents. For this reason, the Raman spectra for all of the chemicals investigated were also measured and included in the spectral analysis. The analysis begins with methyl phosphonic acid, the final hydrolysis product, since it is the simplest molecule, and the vibrational modes have been assigned.17-19 This approach provides greater confidence in the assignments of the more complex molecules, in particular VX. It should be realized that ethanol is also a hydrolysis product, but is SERS-inactive and consequently not included in this study. Table II summarizes the assignments of the measured spectral peaks to vibrational modes for a 1 g/mL aqueous MPA solution. Six of the possible 24 vibrational modes for this molecule with Cs symmetry occur in the solution Raman spectrum between 350 and 1650 cm-1 (Fig. 2A). The dominant spectral feature at 763 cm-1 is assigned to the symmetric PC stretch, which in essence bonds methyl and phosphate tetrahedral-like structures. Moderately intense peaks at 444 and 954 cm-1 are assigned to a symmetric PO3 bend and a symmetric PO3 stretch, respectively. The other three peaks of moderate intensity at 488, 883, and 1423 cm-1 are assigned to a PO3 bend, a CH3 rock, and a CH3 bend, respectively. The SERS spectrum of 0.1 mg/mL MPA is very similar to the Raman spectrum in general appearance (Fig. 2B), dominated by the peak at 756 cm-1, which is again assigned to the symmetric PC stretch. This peak has gained intensity relative to all of the other peaks, suggesting that this mode is perpendicular to the surface, based on previous research that has shown that modes couple to the plasmon field more effectively in this orientation.20 While shifts in the peaks at 954 and 1423 cm-1 to 958 and 1420 cm-1, respectively, are minor, shifts in the peaks at 444 and 488 cm-1 to 469 and 521 cm-1, respectively, are more substantial. Nevertheless, these latter peaks are consistent with Raman spectra of monobasic anion of methylphosphonic acid (MPA-), which have been reported at 462 and 507 cm-1, respectively.18 This is further

SP

O

O

N

SHN

OHP

O

O

SP

O

OH

N

OHP

O

OH

+VX

EMPA

EtOH

DIASH

H2O

MPA+ EA2192

EtOH

H2O

H2O

MPA

+

+

DIASH

Pathway 1

Pathway 2

Applied Spectroscopy, 59, 2005 656

supported by recent pH dependent SERS studies of MPA, that show that MPA- is the predominant species at neutral pH and very low concentrations.8 Two additional peaks appear at 1038 and 1300 cm-1. The former has also been reported for the Raman spectrum of MPA- at 1040 cm-1 and has been assigned to a symmetric PO2 stretch, while the latter peak has been observed in infrared spectra at 1310 cm-1, and assigned to a symmetric CH3 bend.18 Taken together, the shift in the frequency of these PO3 peaks and the increased intensity of the PC mode, the SERS data suggests that MPA is oriented with the PO3 group interacting with the silver surface and the methyl group away from the surface. FIG. 2. A) Raman and B) SERS spectra of MPA. Conditions: A) 1g/mL MPA in water, 300 mW of 785 nm, 5-min acquisition time, B) 0.1 mg/ml in water, MTMS/ODS/TMOS sol-gel in glass capillary, 100 mW of 785 nm, 1-min acquisition time.

The next simplest hydrolysis product of VX is ethyl methylphosphonic acid, formed according to Pathway 1. The replacement of a hydroxy with an ethoxy group quickly increases the number of predicted vibrational modes to 42, decreases the symmetry of the molecule as well as the purity of the modes, and adds a CPOCC backbone. In addition to the appearance of several new peaks, the dominant PC symmetric stretch at 763 cm-1 is replaced by a peak at 730 cm-1 in the Raman spectrum (Fig. 3A), which is now assigned as a backbone stretch containing PC and OCC character. The asymmetry of this peak suggests an additional, underlying peak, which may also be due to a backbone mode. The CH3 rock and bending modes that occurred for MPA at 883, 1300 (SERS) and 1423 cm-1, are still apparent at 893, 1293 and 1420 cm-1, while additional CH2 rock, and CH3 and CH2 bending modes occur at 792, 1454 and 1480 cm-1. The MPA PO3 bending modes at 444 and 488 cm-1 are replaced by PO2 bending modes at 475 and 503 cm-1, while a new peak at 1047 cm-1 is assigned to a PO2 stretch, as was the 1038 cm-1 peak in the MPA SERS spectrum. The second most intense peak in the Raman spectrum at 1098 cm-1 is characteristic of CO or CC stretches, and is assigned as such without differentiation. Changes, similar to MPA, occur in the SERS spectrum of EMPA (Fig. 3B). Again, the PC stretch, or at least the PC containing backbone modes, which are now resolved at 727 and 746 cm-1, are enhanced the most. However, this enhancement relative to the other peaks, is less than for MPA, since the

modes are no longer pure PC and can not be oriented completely perpendicular to the surface. Nevertheless, interaction with the silver is still most favored through the oxygen atoms, which not only shifts the PO2 stretch from 1047 to 1059 cm-1, but also produces significant enhancement. The remaining POn and CHn modes shift by less than 10 cm-1 and are less enhanced by interaction with silver.

FIG. 3. A) Raman and B) SERS spectra of EMPA. Conditions as in Fig. 2, but A) neat liquid, 100 mW of 785 nm, 5-min, B) 0.1 % v/v in MeOH.

The other major hydrolysis product of VX according to Pathway 1 is 2-(diisopropylamino) ethanethiol. The normal Raman spectrum can be analyzed in terms of an alkanethiol and an alkyl substituted tertiary amine. For example, the former chemical type produces a CSH bending mode and two CS stretching modes between 650 and 750 cm-1, and an SH stretching mode at 2570 cm-1.21,22 DIASH contains peaks at 667, 721, 738, and 2569 cm-1 (Fig. 4A), which are assigned to these respective modes. The latter chemical type produces one NC3 breathing mode in the 400-500 cm-1 region and a second breathing mode near 950 cm-1, an NCC bending mode near 570 cm-1, an NC stretching mode near 1200 cm-1, and in concert CH bending modes near 740 and 1450 cm-1.23,24 DIASH contains peaks at 481, 945, 585, 1184, 738, and 1441 cm-1, which are assigned to these respective modes. Note that the assignment of the peak at 738 cm-1 has been assigned to both a CS stretch and a CH bend. Also the most intense peak in the spectrum appears at 814 cm-1 and is attributed to a backbone mode consisting of SC stretching and NC3 breathing modes. The Raman spectrum also contains two low frequency peaks at 416 and 435 cm-1 that are attributed to CC or CN bending modes, while more than 12 moderately intense peaks appear between 1000 and 1400 cm-1, which are variously assigned to CC or CN stretches, or CHn bending modes. The SERS spectrum of DIASH is dominated by the nitrogen and sulfur containing modes (Fig. 4B), specifically peaks at 482, 587, 811, and 938 cm-1 can be attributed to modes at similar frequencies in the Raman spectrum. This is expected for the sulfur modes, since DIASH can couple strongly to the silver surface through a deprotonated sulfur. Deprotonation is supported by the absence of the 667 and 2569 cm-1 peaks assigned to the CSH and SH modes, respectively, in the SERS spectrum. It is also believed that this interaction shifts the CS mode from 738 to 698 cm-1. A similar shift of 26 cm-1 has been observed for simple

Ram

an In

tens

ity (r

elat

ive)

Wavenumber (∆cm-1)

B A

450 650 850 1050 1250 1450 1650

Ram

an In

tens

ity (r

elat

ive)

Wavenumber (∆cm-1)

B A

450 650 850 1050 1250 1450 1650

Applied Spectroscopy, 59, 2005 657

alkanethiols in the Raman and SERS spectra.25-27 It is also believed that the 738 cm-1 peak of moderate intensity in the SERS spectrum of DIASH is the CH bend component of the Raman peak. An additional peak occurs in the SERS spectrum at 1032 cm-1 that likely contains some S character. The enhancement of the two NC3 modes at 482 and 938 cm-1 is somewhat surprising since these modes are sterically excluded by the isopropyl groups from interacting with the surface. Consequently, the enhancement is attributed to a molecular orientation with these modes perpendicular to the surface, which is easily attained.

FIG. 4. A) Raman and B) SERS spectra of DIASH. Conditions as in Fig. 3, but A) 1g/mL in water, B) 1 mg/mL in MeOH. The last hydrolysis product studied in this series is EA2192, and most of the observed Raman peaks can be assigned to the same modes assigned for the Raman peaks of MPA, EMPA and DIASH. Specifically, the Raman peaks at 418, 484, 587, 814, 1132, 1183, 1219, 1306, 1343, 1399, and 1460 cm-1 (Fig. 5A), can be assigned to the following DIASH modes; a CC or CN bending mode, an NC3 breathing mode, an NCC bending mode, the SCNC3 backbone mode, three NC stretching modes, and four CHn bending modes. Similarly, the peaks at 732 and 1418 cm-1 can be assigned to MPA or EMPA modes; an OPC backbone mode and the CH3 wagging mode of the isolated methyl group bound to phosphorous. The PS bond connecting the MPA and DIASH moieties also produces several new peaks. For example, the peaks at 386, 513, and 1054 cm-1 (the latter being the most intense peak in the spectrum) are assigned to SPO bending, PO2S bending and PO2S stretching modes, respectively. The peak at 947 cm-1 is assigned to an NC3 stretch based on the DIASH spectrum, while a less intense peak at 966 cm-1 is assigned to a PO2 stretch based on the MPA spectrum. It is also worth noting that the peaks at 667 and 2569 cm-1 that were observed for DIASH due to SH modes are absent, as expected. Just as the Raman spectrum of EA2192 is dominated by DIASH peaks, so is the SERS spectrum (Fig. 5B). This includes peaks at 481, 584, 693, 811, 939, and 1125 cm-1, assigned to an NC3 breathing mode, an NCC bending mode, the shifted CS stretching mode, the SCNC3 backbone mode, another NC3 stretching mode, and a NCC stretching mode. Three additional peaks of significant intensity occur at 526, 735, and 971 cm-1, and are all attributed to phosphate modes, a

PO2S bend, the OPC stretch, and a PO2 stretch. The appearance of the SC stretching mode at 693 cm-1 indicates that sulfur still interacts with silver significantly. But then, the absence of the PO2S stretching mode at 1054 cm-1 is difficult to explain, and the Raman assignment is therefore, in doubt.

FIG. 5. A) Raman and B) SERS spectra of EA2192. Conditions: A) pure solid, 150 mW of 785 nm, 1-min, 1 cm-1, B) 1 mg/mL in water, 100 mW of 785 nm, 1-min in standard SERS vial.

The Raman spectra of VX and EA2192 are surprisingly different. This may be attributed, at least to some degree, to the fact that VX was measured as a pure liquid, while EA2192 was measured as a solid, the natural states for these two chemicals at room temperature. The change in state can certainly account for the peaks in the VX spectrum to be broader, overlap, and change relative intensity (Fig. 6A). Nevertheless, the following peaks are found at near the same frequency as the EA2192 peaks; 372, 461, 484, 528, 696, 744, 836, 856, 891, 931, 1015, 1101, 1170, 1214, 1300, 1366, 1394, 1443, and 1462 cm-1, and are assigned accordingly (see Table II). The addition of the ethyl group produces two new peaks at 1101 and 1228 cm-1, which are assigned to an OC stretching mode (see EMPA) and a CH2 bending mode. The reappearance of the PC stretching mode at 769 cm-1 suggests that this peak and the 731 cm-1 peak contain significant OPC

FIG. 6. A) Raman and B) SERS spectra of VX. Conditions as in Fig. 5, but A) pure liquid, and B) 1% v/v in methanol.

Ram

an In

tens

ity (r

elat

ive)

Wavenumber (∆cm-1)

B A

450 650 850 1050 1250 1450 1650

Ram

an In

tens

ity (r

elat

ive)

Wavenumber (∆cm-1) 450 650 850 1050 1250 1450 1650

B A

Ram

an In

tens

ity (r

elat

ive)

Wavenumber (∆cm-1) 450 650 850 1050 1250 1450 1650

B A

Applied Spectroscopy, 59, 2005 658

Table II. Tentative vibrational mode assignments for Raman and SERS peaks for VX and its hydrolysis products

MPA EMPA DIASH EA2192 VX Tentative Assignmentsa NR SER NR SER NR SER NR SER NR SER

386 372 376 SPO bend 423 416 418 CC or CN bend 435 CC or CN bend

444b,c 469c 453 456 461 458 POn bend 481d 482 484 481 484 484 NC3 breathing

488b 475 482 499 POn bend 521c 503 505 513 526 528 539 POn(S) bend 585d 587 587 584 NCn bend 645 667 622 PSC bend 667e CSH bend 697f 693 696 CS stretch 730 727 721e 732 735 744 731 PC stretch + backbone (CPOCC) 738d,e 738 CH bend and/or CS stretch

763 756 741sh 746 769 769 PC stretch and/or backbone 792 779 790 CH bend 817 811 814 811 805 SC stretch + NC3 breathing 827 830 831 830 836 820

883b,c 893 891 889 863 856 CH3 bend 904 903 905 891 891 885 OPC stretch / CCN stretch 929 925 946d 938 947 939 931 939 NC3 stretch

954b,c 958 945 966 971 965 POn stretch 1003 1010 1006 1015 1006 POn or CH3 bend 1043 1032 1040 1029 SCCN bend 1038c 1047 1059 1054 PO2(S) stretch

1070 1098 1094 1095 1101 1096 OC or CC stretch 1129 1120 1132 1125 1121 NC stretch 1162 1184d 1205 1183 1170 NC stretch 1224 1219 1214 1220 NC stretch 1228 1237 CH2 bend 1253 1300 1293 1287 1299 1306 1300 1301 CH3 bend 1329 1327 1365 1355 1343 CN bend + CC bend 1366 1365 1366 1397 1399 1394 1400 CH3 bend / NC3 stretch

1423b,c 1420 1420 1416 1418 CH3 bend 1454 1441 1449d 1427 1443 1439 CH2 bend 1451 CHn bend 1480 1461 1459 1460 1464 1462 1462 CHn bend 1493 1547 CH3 bend

a Assignment terminology is simplified since assignments refer to multiple molecules. b = Ref. 17, c = Ref. 18, d = Refs. 22 and 23, e = Refs. 20 and 21, f = Refs. 24-26 character. Most of these assignments are consistent with those of a computer predicted Raman spectrum,28 especially since the VX modes are significantly delocalized and only the primary contributions are listed. The most intense peaks were predicted at 455, 546, 713, 759, 762, 880, 1093, 1216, 1414, 1441, and 1463 cm-1, and assigned to a PS stretch or CPO bend, PO2SC wag, SC stretch, PC stretch, OCC stretch, CC stretch or CH3 rock, OC stretch or CH3 rock, NC stretch, the CH3 bend of the phosphorous methyl group, and two CH bends of the

isopropyl groups. The SERS spectrum of VX is reasonably similar to the Raman spectrum, with corresponding peaks at 376, 458, 539, 731, 939, 1096, 1301, 1439, and 1462 cm-1 readily observed (Fig. 6B). In fact the greatest difference is that the CC and CHn modes are not enhanced, as expected, and little can be said about the orientation of the molecule to the surface, other than the PO2S group interacts sufficiently to be enhanced producing the peak at 539 cm-1. It is worth noting that the

Applied Spectroscopy, 59, 2005 659

SERS spectra of VX and EA2192 are not that similar. In particular, the NC3 modes have little intensity in the VX spectrum. More interestingly, perhaps, is the similarity between the EA2192 and DIASH SERS spectra. The principle difference being the addition of the PC stretching mode at 735 cm-1. This may simply be due to the fact that both molecules interact through the sulfur with the metal surface to similar extents resulting in similar orientations. However, it is also possible that the EA2192 spectrum is of DIASH. This is possible if EA2192 either hydrolyzed or photodegraded. Since the sample was prepared and measured within 1 hour, and the hydrolysis half-life is on the order of weeks,1 the former explanation seems unlikely. Since the peak intensities did not change during these measurements, photodegradation catalyzed by silver also seems unlikely. Further experiments are required to clarify this point. CONCLUSION

We have reported the SERS spectra of VX and its hydrolysis products, EA2192, DIASH, EMPA, and MPA. Tentative vibrational mode assignments for the observed SERS peaks have also been provided. This was accomplished with the aid of the corresponding Raman spectra for these chemicals. Overall the SERS spectra consisted of unique peaks at approximately 460, 530, 730, 760, and 890 cm-1, assigned to POnX (X= O or S) and PC and PS backbone modes. The contribution of these modes had sufficient variability that each chemical could be uniquely identified by its SERS spectrum in this low frequency region. However, quantifying each of these chemicals in an aqueous mixture may require chemical separations or chemometric approaches. Such approaches, along with establishing detection limits and pH dependence for these chemicals are currently being pursued. ACKNOWLEDGMENTS The authors are grateful for the support of the U.S. Army (DAAD13-02-C-0015, Joint Service Agent Water Monitor program). The authors would also

like to thank Dr. Steve Christesen for helpful discussions, and Mr. Chetan Shende for sol-gel chemistry development. ____________________________ 1. Y. Yang., Acc. Chem. Res. 32, 109 (1999). 2. Y. Yang, J. Baker and J. Ward, Chem. Rev. 92, 1729 (1992). 3. W. Creasy, M. Brickhouse, K. Morrissey, J. Stuff, R. Cheicante, J. Ruth,

J. Mays, B. Williams, R. O’Connor, and H. Durst, Environ. Sci. Technol. 33, 2157 (1999).

4. Q. Liu, X. Hu, and J. Xie, Anal. Chim. Acta 512, 93 (2004). 5. Y. Lee and S. Farquharson, SPIE-Int. Soc. Opt. Eng. 4378, 21 (2001). 6. S. Farquharson, P. Maksymiuk, K. Ong, and S. Christesen, SPIE-Int. Soc.

Opt. Eng. 4577, 166 (2001). 7. S. Farquharson, A. Gift, P. Maksymiuk, F. Inscore, W. Smith, K.

Morrisey, and S. Christesen, SPIE-Int. Soc. Opt. Eng. 5269, 16 (2004). 8. S. Farquharson, A. Gift, P. Maksymiuk, F. Inscore, W. Smith, SPIE-Int.

Soc. Opt. Eng. 5269, 117 (2004). 9. K. M. Spencer, J. Sylvia, S. Clauson, and J. Janni, SPIE-Int. Soc. Opt.

Eng. 4577, 158 (2001). 10. P. Tessier, S. Christesen, K. Ong, E. Clemente, A. Lenhoff, E. Kaler, and

O. Velev, Appl. Spectrosc. 56, 1524 (2002). 11. S. D. Christesen, M. J. Lochner, M. Ellzy, K. M. Spencer, J. Sylvia, and

S. Clauson, 23rd Army Science Conference, Orlando (2002). 12. D. L. Jeanmaire and R. P. Van Duyne, J. Electroanal. Chem. 84, 1 (1977). 13. T. E. McKone, B. M. Huey, E. Downing, and L. M. Duffy, Strategies to

Protect the Health of Deployed U.S. Forces: Detecting, Characterizing, and Documenting Exposures (National Academy Press, Washington, D.C., 2000) p.207.

14. S. Farquharson and P. Maksymiuk, Appl. Spectrosc. 57, 479 (2003). 15. S. Farquharson, A. Gift, P. Maksymiuk, and F. Inscore, Appl. Spectrosc.

58, 351 (2004). 16. S. Christesen, B. MacIver, L. Procell, D. Sorrick, M. Carrabba, and J.

Bello, Appl. Spectrosc. 53, 850 (1999). 17. R. A. Nyquist, J. Mol. Struct. 2, 123 (1968). 18. B. J. Van Der Veken and M. A. Herman, J. Mol. Struct. 15, 225 (1973). 19. B. J. Van Der Veken and M. A. Herman, J. Mol. Struct. 15, 237 (1973). 20. J. S. Suh and M. Moskovitz, J. Am. Chem. Soc. 108, 4711 (1986). 21. M. Hayashi, Y. Shiro, H. Murata, Bull. Chem. Soc. Jpn. 39, 112 (1966). 22. T. Torgrimsen and P. Kleboe, Acta Chem. Scand. 24, 1139 (1970). 23. C. Crocker and P. L. Goggin, J. Chem. Soc. Dalton Trans. 5, 388 (1978). 24. C. Gobin, P. Marteau, and J.-P. Petitet, Spectrochim. Acta 60, 329 (2004). 25. T. H. Joo, K. Kim, and M. S. Kim, J. Phys. Chem. 90, 5816 (1986). 26. C. H. Kwon, D. W. Boo, H. J. Hwang, and M. S. Kim, J. Phys. Chem. B

103, 9610 (1999). 27. A. Kudelski, Langmuir 19, 3805 (2003). 28. H. Hameka and J. Jensen, ERDEC-TR-065 (1993).

SPIE-5585 2004 53

Detect-to-treat:

development of analysis of Bacilli spores in nasal mucus by surfaced-enhanced Raman spectroscopy

Frank E. Inscore, Alan D. Gift, and Stuart Farquharson*

Real-Time Analyzers, Inc., East Hartford, Connecticut 06108

ABSTRACT As the war on terrorism in Afghanistan and Iraq continue, future attacks both abroad and in the U.S.A. are expected. In an effort to aid civilian and military personnel, we have been investigating the potential of using a surface-enhanced Raman spectroscopy (SERS) sampling device to detect Bacillus anthracis spores in nasal swab samples. Such a device would be extremely beneficial to medical responders and management in assessing the extent of a bioterrorist attack and making detect-to-treat decisions. The disposable sample device consists of a glass capillary filled with a silver-doped sol-gel that is capable of extracting dipicolinic acid (DPA), a chemical signature of Bacilli, and generating SERS spectra. The sampling device and preliminary measurements of DPA extracted from spores and nasal mucus will be presented.

Keywords: Dipicolinic acid; Bacillus spores; Anthrax; Surface-enhanced Raman spectroscopy.

1. INTRODUCTION In the autumn of 2001 the threat of conventional suicide-bombing terrorism and bioterrorism within the United States became a grave reality. Consequently, future terrorist attacks both at home and abroad against civilian and military personnel alike are undeniable possibilities. In the case of using anthrax causing spores as a terrorist weapon, much was learned from the distribution of endospores through the U.S. postal system.1-6 For example, it was established that detection of exposure within the first few days allowed successful treatment of victims using Ciproflaxin, deoxycycline and/or penicillin G procaine.5 However, the National Naval Medical Center who processed 3,936 nasal swab samples from the Capitol Hill, DC and Brentwood, NJ postal facility employees, required 2-3 days of growing microorganisms in culture media to establish that all but six employees were uninfected.6 The remaining six employees were also uninfected, but the samples required further analysis. This process was reported as “extremely time-consuming and labor-intensive”. This re-emphasizes the much stated need for methods to rapidly detect Bacillus anthracis spores so that emergency responders and management can assess the extent of the event and make detect-to-treat decisions. Nevertheless, the challenges are formidable considering that the Center for Disease Control (CDC) estimates that inhalation of 10,000 anthracis endospores or 100 nanograms will be lethal to 50% of an exposed population (LD50).7

Although polymerase chain reactions (PCR)8,9 and immunoassays5,10,11 have been developed to augment or replace the standard laboratory method of culture growth, they still have significant limitations. PCR still requires hours to perform and each analyzer is limited to the number of samples that can be measured, while the latest immunoassays designed to detect the response of immunoglobulin G to the protective antigen of B. anthracis are only 80% specific and require at least 10 days after infection to be detected.5 As an alternative to these methods, several researchers have been investigating the analysis of calcium dipicolinate (CaDPA) as a B. anthracis signature.12-14 This approach is viable because only spore forming bacteria contain CaDPA, and the most common, potentially interfering spores, such as pollen and mold spores, do not. It has been long known that Raman spectra of Bacilli spores are dominated by bands associated with CaDPA15 and that these spectra may provide a suitable anthrax signature at the genus level.16 With this in mind, we have been investigating the potential of using a surface-enhanced Raman spectroscopy (SERS) sampling device to detect spores in nasal swab samples. The design, intended for medical responders, employs disposable SERS-active capillaries (one per analysis) that can be easily analyzed using a portable Raman analyzer.17 This approach is based on our previous SERS measurements of dipicolinic acid (DPA), the acid of CaDPA, both in water18,19 and extracted from B. cereus spores.20

SPIE-5585 2004 54

2. EXPERIMENTAL Lyophilized B. cereus spores, prepared according to literature,16 were supplied by the University of Rhode Island and used as received. Dipicolinic acid (2,6-pyridinedicarboxylic acid), dodecylamine (DDA), and all chemicals used to prepare the silver-doped sol-gel coated capillaries were obtained from Sigma-Aldrich (Milwaukee, WI) and used without further purification. The SERS-active capillaries were prepared according to previous published procedures for the Simple SERS Sample Vials using a silver amine precursor and an alkoxide precursor with the following modifications.17 The alkoxide precursor employed a combination of methyltrimethoxysilane (MTMS) and tetramethyl orthosilicate (TMOS) in a v/v ratio of 6/1, which was mixed with the amine precursor in a v/v ratio of 1/1. Approximately 15 microL of the mixed precursors were then drawn into a 1-mm diameter glass capillary coating a 15-mm length. After sol-gel formation, the incorporated silver ions were reduced with dilute sodium borohydride. The serial diluted samples of DPA were prepared in HPLC grade water. B. cereus samples were prepared using ~0.1 mm3 particles with a typical mass of 0.1 mg. The sample masses were consistent with a previous determination of spore density at 0.081 g/mL that indicated a high degree of entrained air. These particles were carefully divided into 3 or 10 equal specks prior to the addition of DDA or nasal mucus (see RESULTS AND DISCUSSION). DPA or B. cereus spores were artificially added to nasal mucus samples that were collected in 20 mL glass vials by expulsion. The DPA in mucus samples were prepared by mixing equal volumes of 1mg/mL DPA in water and mucus. The B. cereus in mucus samples were prepared by adding a finely diced 0.1 mg spore sample to 100 microL of mucus. For each of the spore samples, either specks or 100 microL of spore containing mucus, 100 µL drop of a 50 mM DDA solution in ethanol, pre-heated to 78 oC, was added and allowed to digest the spore coat for 1 minute. The resultant solutions, as were the DPA in water samples, were drawn into SERS-active capillaries for analysis. This was accomplished by mounting the capillaries horizontally to an XY positioning stage (Conix Research, Springfield, OR) just inside the focal point of an f/0.7 aspheric lens. The probe optics and fiber optic interface have been described previously.20 A Fourier transform Raman spectrometer (Real-Time Analyzers, model IRA-785, East Hartford, CT) equipped with a 785 nm diode laser (Process Instruments Inc. model 785-600, Salt Lake City, UT) and a silicon photo-avalanche detector (Perkin Elmer model C30902S, Stamford, CT) was used to deliver 100 mW of power to the SERS samples and generate spectra with 8 cm-1 resolution.

3. RESULTS AND DISCUSSION Previously we reported SERS spectra of dipicolinic acid at a series of concentrations obtained in 2-mL glass vials internally coated with a silver-doped sol-gel as the SERS-active media.19 This included samples as low as 1 mg/L using 100 mW of 785 nm and 1-min acquisition time. For this concentration the signal was barely discernable above the noise

Figure 1. SERS spectra of DPA in water at A) 1 mg/L and B) 10 microg/L (100 pg in 10 microL sample) using the SERS-active capillaries, 100 mW of 785 nm and 1- min acquisition time.

for the 1008 cm-1 peak (signal-to-noise, S/N =5.6), and a limit of detection (LOD, defined as a S/N of 3) was estimated just below the measured value at 540 microg/L. One limitation of these vials is that the sample must diffuse through the porous sol-gel to the silver surface for SERS to occur. Since this might limit sensitivity or require allowance for diffusion, we have developed sol-gel filled capillaries. A syringe allows drawing the sample through the sol-gel in a couple of seconds forcing analyte-to-surface interactions. In an effort to establish that these SERS-active capillaries provide better sensitivity, a set of serially diluted solutions of DPA in HPLC grade water were prepared and measured. Figure 1 shows that, as desired, a significantly better DPA SERS spectrum was obtained for 1 mg/L using the capillaries rather than the vials. In fact 10 microg/L samples repeatedly produced spectra (Figure 1B). Intense peaks are observed at 815, 1008, and 1382 cm-1, moderate

A B

SPIE-5585 2004 55

peaks are observed 657, 758, 1049, 1182, 1428 cm-1, and 1567 cm-1. Several of these peaks have been previously assigned based on the Raman spectrum of DPA as follows:15,16,20 the 1008 cm-1 peak to the symmetric ring stretch, the 1382 cm-1 peak to the O-C-O symmetric stretch, the 1428 cm-1 peak to the symmetric ring C-H bend, and the 1567 cm-1 peak to the asymmetric O-C-O stretch. The 10 microg/L sample was used to estimate an LOD of 1 microg/L (S/N equaled 33 for the 1008 cm-1 peak). This was consistent with the fact that attempted measurements of 1 microg/L samples did yield spectra, but not in every case. It is also worth noting that only 10 microL samples were used to generate the spectra, or in the case of the 10 microg/L sample, 100 pg of DPA. Previously, the SERS-active capillaries were used to measure DPA extracted from ~10 microg of Bacillus cereus spores, and preliminary spectra were reported.20 The procedure is described here (Figure 2). Three 0.1 mg samples of B. cereuswere weighed and then each diced into ~ 10 equal parts (~10 microg or 10 million spores), which allowed performing 30 measurements. To each particle 100 microL of 50 mM DDA in ethanol at 78 oC was added. After 1 minute the solution was drawn into a SER-active capillary, which was then mounted above a laser excitation beam such that the surface-enhanced Raman spectrum could be acquired. Figure 2E shows a representative spectrum for one of these capillaries using a 1-min acquisition time. The primary DPA peaks at 657 cm-1, 815 cm-1, 1008 cm-1, 1382 cm-1, and 1428 cm-1 are easily seen. Again, the S/N of the 1008 cm-1 peak, which was measured as 120, was used to estimate an LOD of 250 ng or 25,000 B. cereus spores in 100 microL DDA. Since it is known that B. cereus spores contain 10-15% DPA (as calcium dipicolinate),21 and that the majority of the DPA is extracted by hot DDA,14 this LOD can be compared to DPA in water. Accordingly, the 10 microg of spores per 100 microL DDA is approximately equivalent to 10 mg of DPA per L water, and consequently the LOD is equivalent to 250 microg/L, which is considerably less sensitive than the 10 microg/L measured for DPA in water. In an effort to measure fewer spores, anhydrous ether was used to disperse spores on a surface to the point of being invisible to the unaided eye. In this series of experiments a 0.1 mg B. cereus sample was divided into three near equivalent specks. To each speck 600 microL of ether was added and allowed to dry. The dispersed spores and ether produced a solvent ring ~5 cm in diameter with a significant portion of the spores at the edge. A non-cotton swab was used to collect the residual spores in the center 1/3rd of this area. The swab was added to a vial containing 100 microL of 50 mM DDA in ethanol heated to 78 oC. After 1-min, ~ 10 microL of this solution was extracted into a SERS-active capillary and measured as before. The peaks in the SERS spectrum, acquired in 1-min, are ~ 1/5th the intensity of those in the previous experiment, suggesting a collected sample of ~2 microg (Figure 2F). The measured S/N of 25 for the

Figure 2. Sample preparation includes A) three initial 0.1 mg B. cereus spore samples, B) addition of 100 microL 78 oC 50 mM DDA to ~10 microg portion, C) drawing 10 microL into SERS-active capillary, and D) mounting capillary in Raman sample compartment. E) SERS spectrum of representative 10 microg sample using 150 mW of 785 nm and 1-min acquisition time. F) SERS spectrum of representative 2 microg sample using 100 mW of 785 nm and 1-min acquisition time.

A B

C D

E F

SPIE-5585 2004 56

1008 cm-1 peak suggests an LOD of 250 ng. Although this LOD is equivalent to the previous experiment, this experiment has at least lowered the measured amount of spores by a factor of 5. In either case, comparison to the measurement of 10 microg/L DPA, suggests that these procedures include considerable losses in extracting the DPA from the spores and transferring it to the silver surface. Conversely, if the efficiency of these procedures can be improved then 1 ng or 100 spores should be able to be detected. were observed, even when the sample was kept at 78 oC for 10 minutes. Several possibilities may explain this result. It is possible that chemicals within mucus 1) react with or coat the spores protecting them from digestion by the DDA, 2) react with DDA making it ineffective in digesting the spores, 3) effectively clog the sol-gels preventing released DPA from reaching the silver particles, 4) react with the silver particles and deactivate their Raman signal enhancing properties, 5) react with DPA making it unavailable for measurement, or 6) any combination of these possibilities. The successful measurement of DPA in nasal mucus suggests that possibilities 3 and 4 are not the major reason for being unable to detect DPA extracted from spores contained in mucus. Experiments are currently being designed and tested to determine which of these possibilities is hindering the measurement.

4. CONCLUSION Towards the goal of developing a simple SERS-active sample device to measure Bacillus anthracis spores in nasal mucus, we have measured 100 pg dipicolinic acid in a 10 microL water sample, suggesting that as few as 100 spores could be measured. However, only 0.2 microg of B. cereus spores in a 10 microL sample were measured lowering expectations to 20,000 spores. Furthermore, SERS spectra were not obtained for B. cereus spores artificially added to nasal mucus. Current research is aimed at determining the factors that hindered this last measurement, and at developing the appropriate separation methods to overcome this limitation. However, it is worth noting that the presented method can be used to detect spores on surfaces, and may have value in determining the extent of facility contamination.

ACKNOWLEDGEMENTS The authors are grateful for the support of the National Science Foundation (DMI-0296116 and DMI-0215819) and the U.S. Army (DAAD13-02-C-0015, Joint Service Agent Water Monitor program). The authors are indebted to Chetan Shende for preparing the SERS-active capillaries. The authors would also like to thank James Gillespie, Nicholas Fell, and Augustus Fountain for providing important background information, and Professor Jay Sperry of the University of Rhode Island for supplying B. cereus spores.

Figure 3. SERS spectra of A) 0.5 mg/mL DPA in a 50/50 nasal mucus/water mixture and B) 1 mg/ml DPA in HPLC water for comparison. Conditions as in Fig. 1, but A) 5-min.

In an effort to establish baseline sensitivity for spores contained in nasal mucus, several samples were prepared and measured. Although nasal mucus is mostly water, it contains sulfate, sugars, proteins (including albumin), protective enzymes and phagocytes, as well as mucin, a glycoprotein. Consequently, the first samples consisted only of DPA added to nasal mucus to evaluate the potential chemical and spectral interferences that could result from this matrix. Approximately 10 microL of a 0.5 mg/mL DPA in a 50/50 mucus/water mixture was drawn into a SERS-active capillary without any pretreatment and measured. Although the matrix produced a significant offset of the baseline, the primary, characteristic spectral peaks of DPA were easily observed (Figure 3). Next finely divided specks of B. cereus were added to nasal mucus, thoroughly mixed, and treated with hot DDA. Again 10 microL samples were drawn into the SERS-active capillaries and measured. Unfortunately, no peaks

A B

SPIE-5585 2004 57

REFERENCES

1 Jernigan, JA et al. “Bioterrorism-related inhalation anthrax: The first 10 cases reported in the United States”,

Emerg. Infect. Dis. 6, 933-944 (2001). 2 Klietmann, WF, and KL Ruoff “Bioterrorism: implications for the clinical microbiologist,” Clin. Microbiol. Rev. 14,

364-381 (2001). 3 Rotz, LD, AS Khan, SR Lillibridge, SM Ostroff, and JM Hughes, “Public health assessment of potential biological

terrorism agents,” Emerg. Infect. Dis. 8, 225-230 (2002). 4 Dewan, PK et al. “Inhalational Anthrax Outbreak among Postal Workers, Washington, D.C., 2001,” Emerg. Infect.

Dis. 8, 1066-1072 (2002). 5 Bell DM, PE Kozarsky, D. Stephens, “Clinical issues in the prophylaxis, diagnosis, and treatment of anthrax,”

Emerg. Infect. Dis. 8, 222-225 (2002); 6 Kiratisin, P et al. “Large-scale screening of nasal swabs for Bacillus anthracis: Descriptive summary and discussion

of the National Institute of Health’s experience”, J. Clin. Microbio., 3012-3016 (2002) 7 Ingelsby TV, et al. “Anthrax as a biological weapon, 2002: Updated recommendations for management,” J. Amer.

Med. Ass. 287, 2236-52 (2002) 8 Glick, BR, and JJ Pasternak, Molecular biology: Principles and Applications of Recombinant DNA, ASM Press,

Wash. D.C. (1994). 9 Bell CA, Uhl JR, Hadfield TL, David JC, Meyer RF, Smith TF, Cockerill III FR, ”Detection of Bacillus Anthracis

DNA by LightCycler PCR” J. Clin. Microbiol. 40, 2897 (2002). 10 Gatto-Menking DL, Yu H, Bruno JG, Goode MT, Miller M, Zulich AW “Sensitive detection of biotoxoids and

bacterial spores using an immunomagnetic electrochemiluminescence sensor” Biosens. Bioelectron. 10, 501-507 (1995).

11 Quinlan JJ and Foegeding PM, J. Rapid Methods Automation Microbiol. 6: 1(1998) 12 Nudelman R, Bronk BV, Efrima S “Fluorescence Emission Derived from Dipicolinate Acid, its Sodium, and its

Calcium Salts” App. Spectrosc. 54, 445-449 (2000) 13 Rosen DL, Sharpless C, and McBrown LB “Bacterial spore detection and determination by use of terbium

dipicolinate photoluminescence,” Anal. Chem. 69, 1082-1085 (1997) 14 Pellegrino PM, Fell Jr NF, and Gillespie JB “Enhanced spore detection using dipicolinate extraction techniques,”

Anal. Chim. Acta 455, 167-177 (2002) 15 Woodruff WH, Spiro TG, and Gilvarg C “Raman Spectroscopy In Vivo: Evidence on the Structure of Dipicolinate

in Intact Spores of Bacillus Megaterium,” Biochem. Biophys. Res. Commun. 58, 197 (1974) 16 Ghiamati E, Manoharan R, Nelson WH, and Sperry JF “UV Resonance Raman spectra of Bacillus spores” Appl.

Spectrosc. 46, 357- 364 (1992) 17 Farquharson, S and P Maksymiuk, “Simultaneous chemical separation and surface-enhancement Raman spectral

detection using silver-doped sol-gels,” Appl. Spectrosc., 57, 479-482 (2003) 18 Farquharson S, Smith WW, Elliott S and Sperry JF “Rapid biological agent identification by surface-enhanced

Raman spectroscopy,” SPIE 3855: 110-116 (1999) 19 Farquharson, S, A Gift, P Maksymiuk, F Inscore, and W Smith, “pH dependence of methyl phosphonic acid,

dipicolinic acid, and cyanide by surface-enhanced Raman spectroscopy”, SPIE 5269, 117-125 (2004) 20 Farquharson, S., A. Gift, P. Maksymiuk, and F. Inscore, “Rapid dipicolinic acid extraction from Bacillus spores

detected by surface-enhanced Raman spectroscopy”, Appl. Spectrosc., 58, 351- 354 (2004). 21 F.W. Janssen, A.J. Lund, and L.E. Anderson, Science, 127, 26, (1958).

in the form of hoax letters (5–7). Literallytens of thousands of letters containingharmless powders have been mailed tocreate additional fear (8). Consequently,an analyzer must not only be able to dif-ferentiate B. anthracis spores from otherbiological materials, but must be able alsoto identify these harmless powders toeliminate fear and potentially costly shutdowns (9).

In the case of postal-targeted terror-ism, we have been investigating theutility of Raman and surface-enhancedRaman spectroscopy (SERS) to meetthe analytical challenges of speed, sen-sitivity, and selectivity by identifyingvisible and invisible particles on sur-faces, respectively (10–13).

Raman Spectroscopy — BacilliSpores and Hoax MaterialsRaman spectroscopy is attractive because

very small samples can be measured with-out preparation. The sample need only beplaced at the focal spot of the excitationlaser and measured. Moreover, the richmolecular information provided byRaman spectroscopy usually allowsunequivocal identification of chemicalsand biochemicals. As early as 1974, theRaman spectrum of Bacillus megateriumwas measured and shown to be domi-nated by calcium dipicolinate (CaDPA,14). This chemical can be used as a signa-ture since only spore forming bacteriacontain CaDPA, at ~10% by weight(15–17), and the most common spores,such as pollen and mold spores, do not.The ability of Raman spectroscopy tomeasure and identify spores is exempli-fied in Figure 1. Here an ~1-mm3 spec(~100 mg) of Bacillus cereus spores, anontoxic surrogate for B. anthracis spores,was placed on a glass surface, positioned

Inte

nsity

(ar

bitr

ary

units

)

∆cm-1

1000 1500 2000 2500 3000

Exosporium

Outer core wall

DNA

(a)

(b)

Ribosomes

Calcium dipicolinate

Inner core wall

Cortex

Ca2+

Spore coat

Core

Figure 1. Raman spectra of (a) Bacillus cereus spores and (b) calcium dipicolinate.Conditions: 500 mW of 1064 nm at the sample, 5-min acquisition time.mmediately following the September

11, 2001 terrorist attacks, four letterscontaining anthrax causing spores

were mailed through the U.S. postal sys-tem infecting 22 individuals, five fatally.The anxiety caused by this bioterroristattack was exacerbated by the extensivetime required for positive identificationof the Bacillus anthracis spores and theunknown extent of their distributionalong the east coast. The delay in identi-fication was due to the fact that sporeshad to be germinated and grown in cul-ture media to sufficient cell numbers sothat the 16S rRNA gene unique to B.anthracis could be measured. Conse-quently, the Center for Disease Controland Prevention (CDC) employed a com-bination of biological analyses of culturegrown colonies and polymerase chainreactions to differentiate bacilli from

other bacteria and from each other (1,2).From this bioterrorist attack, it became

clear that considerably faster methods ofanalysis were required. This would expe-dite assessment of the scale of an attack aswell as the extent of facility contamina-tion. This information, in turn, could beused to minimize fatalities, because it waslearned that if exposure is detected with-in the first few days, the majority of vic-tims can be treated successfully usingciprofloxacin, doxycycline, and penicillinG procaine (3). However, the challengesof developing such an analyzer are formi-dable considering that the CDC estimatesthat inhalation of 10,000 anthracisendospores or 100 nanograms will belethal to 50% of an exposed population(4). An additional challenge has emergedsince the 2001 attacks, in that a secondarytype of postal-terrorism has proliferated

Detecting Bacillus Spores by Ramanand Surface-Enhanced RamanSpectroscopy (SERS)Raman spectroscopy has been employed to detect Bacillus cereusspores, an anthrax surrogate, collected from a letter as it passedthrough a mail sorting system. Raman spectroscopy also has thecapability to identify many common substances used as hoaxes. Athree-step method also is decribed for the detection of dipicolinicacid extracted from surface spores by SERS.

Stuart Farquharson, Wayne Smith, Carl Brouillette, and Frank Inscore

Stuart Farquharson is president and CEO, Wayne Smith is vice-president ofRaman products, Carl Brouillette is a senior instrument design engineer, andFrank Inscore is a senior Raman applications specialist, all with Real-Time Analyzers,Inc. (East Hartford, CT). E-mail: [email protected].

I

stants, particle size, and irradiation wave-length must be satisfied to generate theplasmon field. Furthermore, the amountof interaction between the metal plas-mon field and molecule will influence theamount of enhancement.

For the present application, it has beenshown that dipicolinic acid (DPA), theacid form of CaDPA, produces an intenseSER spectrum when it interacts with theplasmon field of silver nanoparticles(Figure 4), and it is, therefore, suitable as asignature for bacilli spores (12,13). TheSER spectrum of 1 mg/L DPA in water issimilar in intensity to the Raman spec-trum of a saturated solution of DPA (80g/L DPA in 1 N KOH). Taking intoaccount sample concentration, laserpower, and signal intensity, the ~1000 cm-

1 peak intensity is enhanced by 9.3 × 105.In addition to the increase in intensity,some of the peaks shift and change inten-sity due to the molecule to plasmon fieldinteraction. Specifically, the followingRaman to SER spectral peak shifts occur:652 to 657 cm-1, 822 to 812 cm-1, 1001 to1006 cm-1, 1386 to 1381 cm-1, 1438 to1426 cm-1, and 1572 to 1567 cm-1.

However, to obtain SERS of DPA, itmust be extracted from spores andbrought in contact with the plasmonfield. Recently, a relatively fast methodusing hot dodecylamine (DDA) has beendeveloped to break apart spores andrelease DPA (23). In addition, silverdoped in a porous glass structure (sol-gel) has been developed as a SER-activematerial for chemical separation (24).These developments were combined toperform the following three-step meas-urement. A 10-µg sample of B. cereusspores was spread over a 10-cm2 glasssurface with the aid of methanol and

allowed to dry to mimic a contaminatedsurface. A 100-µL drop of 50 mM DDAin ethanol heated to 78 °C was added tothe surface. After 1 min, approximately10 µL of the solution was drawn into a 1-mm diameter glass capillary containing a1-cm plug of silver-doped sol-gel. TheSER spectrum of DPA collected in theSER-active capillary was then measuredusing a Raman spectrometer. Figure 4cshows a representative spectrum fromone of these capillaries using 150 mW of785 nm laser excitation and a 1-minacquisition time. The primary DPApeaks at 657 cm-1, 812 cm-1, 1006 cm-1,1381 cm-1, and 1426 cm-1 are observedreadily. The DPA signal intensity is simi-lar to the 1-mg/L sample obtained inwater, and suggests that 10 ng DPA werecollected in the 10-µL sample. Assumingthat a spore contains approximately 10%DPA by weight, this sample correspondsto 100 ng of spores, or for the entire 100-µL drop of DDA, 1 µg of spores. Thissuggests that ~10% of the original sam-ple was collected. This low percentagecould be due to inefficient collection ofthe sample from the surface, incompletedegradation of the spores by DDA, orinefficient transfer of the DPA to the sil-ver particles. Nevertheless, the signal-to-noise ratio (S/N) of 125 for the 1006 cm-

1 peak in the SER spectrum suggests alimit of detection (defined as S/N = 3) ofapproximately 25 ng of B. cereus spores(2500 spores) in 100 µL DDA.

Improvements in sample collectionand/or transfer should allow detectionof spores in the hundreds-per-square-centimeter range. Finally, it is worthnoting that the sol-gel appeared tohave the desired effect of excludingother bacterial cell material from

in 5 min or less (Figure 2). Once a suspi-cious powder is measured, a simplechemical identification algorithm can beused to match the sample spectrum one-for-one to a spectrum contained in alibrary in less than 1 second (21). One ofthe advantages of Raman spectroscopy isthe fact that virtually every chemical pro-duces a unique spectrum, for example,creamer, flour, sugar, and aspirin — thepowders most often suspected as B.anthracis spores (Figure 3). It is worthnoting that the x-axis invariability of aninterferometer-based Raman spectrome-ter makes such search and match algo-rithms reliable (22).

Surface-Enhanced RamanSpectroscopy — Bacilli SporesAlthough an analyzer capable of identi-fying suspicious powders has value, it

also is important to be able to measurespores on surfaces to assess the scale ofcontamination. Such measurementswould be useful in establishing the paththat a spore-containing letter has takenfrom destination back to the source,what processing equipment handled theletter, and to what extent individualsmay have been exposed or are at risk.These measurements may also be valu-able in facility clean-up operations.

Surface-enhanced Raman spec-troscopy offers the possibility of detectingjust hundreds of spores per square cen-timeter. SERS involves the formation of ametal surface plasmon field generated byirradiation with light. The efficiency ofRaman scattering of a molecule interact-ing with this field can be enhanced by sixorders of magnitude or more. However,strict requirements of metal optical con-

Inte

nsity

(ar

bitr

ary

units

)

∆cm-1

800 1000 1200 1400 1600 1800

(a)

(b)

(c)

(d)

(e)

(f)

Figure 3. Raman spectra of (a) B. cereus spores, (b) creamer, (c) flour, (d) sugar, (e)aspirin, and (f) paper. Conditions: as in Figure 1. Note that each powder produces aunique set of Raman spectral peaks.

be used to measure numerous surfacesand map distribution of anthraxendospores in mail distribution facilitiesor other environments should anotherverified attack occur.

AcknowledgmentsThe authors would like to acknowledgethe support of the National ScienceFoundation in development of theanalysis of spores and hoax materials(DMI-0349687), Chetan Shende fordevelopment of the SERS capillaries(DMI-0215819), and the U.S. Army forthe development of the analysismethod for dipicolinic acid (DAAD13-02-C-0015, Joint Service Agent WaterMonitor Program).

References1. Centers for Disease Control and Prevention,

American Society for Microbiology, and Association

of Public Health Laboratories. Basic diagnostic test-

ing protocols for level A laboratories for the pre-

sumptive identification of Bacillus anthracis.

American Society for Microbiology, Washington,

D.C. http://www.bt.cdc.gov/Agent/Anthrax/

Anthracis20010417.pdf.

2. C.T. Sacchi, A.M. Whitney, L.W. Mayer, R. Morey, A.

Steigerwalt, A. Boras, R.S. Weyant, and T. Popovic,

Emerg. Infect. Dis., 8, 1117–1123 (2002).

3. D.M. Bell, P.E. Kozarsky, and D. Stephens, Emerg.

Infect. Dis. 8, 222–225 (2002).

4. T.V.Ingelsby et al., JAMA 287, 2236–2252 (2002).

5. See for an extensive list of hoaxes compiled by the

Associated Press:

http://www.anthraxinvestigation.com/ap.html

6. N. Tucker, ”Capitol Police Officer Convicted of Staging

Anthrax Hoax,” Washington Post, November 22, 2002.

7. Staff reporter, “Anthrax hoax rattles Edmonton post

office,” CBC News, January 3, 2003.

8. Staff reporter, “Post Office Got 20,000 Powder

Reports”, Associated Press, March 4, 2004.

9. Staff reporter, “World Bank reports further testing of

mail finds no anthrax,” Associated Press, April 23,

2002.

10. S. Farquharson, L. Grigely, V. Khitrov, W.W. Smith, J.F.

Sperry, and G. Fenerty, J. Raman Spectrosc. 35,

82–86 (2004).

11. S. Farquharson and W.W. Smith, SPIE-Int. Soc. Opt.

Eng. 5269, 9–15 (2004).

12. S. Farquharson, W.W. Smith, S. Elliott, and J.F. Sperry,

SPIE-Int. Soc. Opt. Eng. 3855, 110–116 (1999).

13. S. Farquharson, A. Gift, P. Maksymiuk, and F. Inscore,

Appl. Spectrosc. 58, 351–354 (2004).

14. W.H. Woodruff, T.G. Spiro, and C. Gilvarg, Biochem.

Biophys. Res. Commun. 58, 197 (1974).

15. F.W. Janssen, A.J. Lund, and L.E. Anderson, Science

127, 26–27 (1958).

16. W.G. Murrell, G.W. Gould, and A. Hurst, Eds., The

Bacterial Spore (Acad. Press, 1969), p. 215.

17. K. Ragkousi, P. Eichenberger, C. Van Ooij, and P.

Setlow, J. Bacteriol. 185, 2315–2329 (2003).

18. J.G. Grasselli, M.K. Snavely, and B.J. Gulkin, Chemical

Applications of Raman Spectroscopy (John Wiley

& Sons, New York, NY, 1981), Chapter 5.

19. J.C. Austin, T. Jordan, and T.G. Spiro, Biomolecular

Spectroscopy, Clark and Hester, Eds. (John Wiley &

Sons, New York, NY, 1993), p. 55.

20. P.J. Treado, M.P. Nelson, and S. Vanni, S., “Raman

chemical imaging provides rapid and non-invasive

chembiothreat detection,” Photonics East,

Philadelphia, October, 2004.

21. B.K. Lavine, C. Davidson, and A.J. Moores, Vib.

Spectrosc. 28, 83–95 (2002).

22. I.R. Lewis, N.W. Daniel Jr., and P.R. Griffiths, Appl.

Spectrosc. 51, 1854–1867 (1997).

22. P.M. Pellegrino, N.F. Fell Jr., and J.B. Gillespie, Anal.

Chim. Acta. 455, 167–177 (2002).

23. S. Farquharson, and P. Maksymiuk, Appl. Spectrosc.

57, 479–482 (2003).

24. L.D. Rotz, A.S. Khan, S.R. Lillibridge, S.M. Ostroff, and

J.M. Hughes, J.M., Emerg. Infect. Dis. 8, 225–230

(2002). ■

reaching the silver and interfering withthe measurement.

SummaryAnthrax remains the highest ranked bio-logical threat agent along with plague(Yersinia pestis, 25), and the need todevelop analyzers that detect and meas-ure these bioagents to minimize theirpotential harm remains. Various analyz-ers are required to address differentaspects of a biological attack rangingfrom detection to treatment. Here, wedescribed two analyzers designed for twoaspects associated with anthrax-basedattacks. The first analyzer, a portable FT-Raman spectrometer, was used to pro-vide a complete answer to the identity ofsuspicious powders that might be found

on mail sorting equipment. Raman spec-troscopy can determine if a suspiciouspowder is a bacilli spore or one of 100common substances used as hoaxes. Wealso demonstrated that longer laser exci-tation wavelengths, such as 1064 nm, arenecessary to avoid fluorescence interfer-ence by some of these common powders,as well as some envelope papers. The sec-ond analyzer, a portable FT-Raman spec-trometer was used in conjunction with aSERS-active sampling device to detectinvisible spores on surfaces.

The analysis involved three steps tobreak apart the spores, collect dipicolinicacid as a signature of bacilli, and measurethe SER spectrum. The entire processrequired just over 2 min. The single-use,disposable, sol-gel filled capillaries could

Inte

nsity

(ar

bitr

ary

units

)

-1

800 1000 1200 1400 1600 1800

(a) (x4)

(b)

(c)

Figure 4. (a) Raman spectrum of 80 g DPA in 1 L 1N KOH in a glass capillary, (b) SERSspectrum of 1 mg DPA in 1 L water in a silver-doped sol-gel filled glass capillary, and(c) ~ 1 mg of spores in a 100-�L drop of DDA collected from 10 cm2 surface. Spectralconditions: (a) 450 mW of 785 nm, 5-minute acquisition time, (b) and (c) 150 mW of785 nm, 1-min acquisition time. Top scale expanded four times.

CRC Book Chapter 10 Draft 101

Chapter 10

Identifying surfaces contaminated with Bacillus spores using surface-enhanced Raman spectroscopy

to detect extracted dipicolinic acid

Frank Inscore, Alan Gift, Paul Maksymiuk, Jay Sperry, and Stuart Farquharson

I. INTRODUCTION On September 18 and October 9, 2001, two sets of letters containing Bacillus anthracis spores passed through the United States Postal Service�s Trenton, NJ, Processing and Distribution Center.1 The first set was destined for Florida and New York, while the second set was destined for Washington, DC. The infection of 22 people by these spores resulted in 5 deaths, a media employee in Florida, two postal workers in DC, a hospital worker in New York, and a retired woman in Connecticut.2 This bioterrorism closely followed the September 11, 2001 attack on the Pentagon building and the World Trade Center towers, which added to the nation�s concern about terrorism within US borders. There was additional anxiety associated with this second attack, in that, it took a long time to positively identify the spores and to determine the extent of their distribution along the east coast, and later, within facilities. This was even true for the letter that was mailed to the Hart Senate Office (HSO) Building in DC. The powder that fell when it was opened was immediately suspected as B. anthracis, due to the previous week�s news from Florida and New York.3 Though samples were collected immediately and delivered to the Center for Disease Control and Prevention (CDC) laboratories within 24 hours,3 it still took several additional days for positive identification.1 This delay is due to the fact that spores must be germinated and grown in culture media to sufficient cell numbers such that the 16S rRNA gene unique to B. anthracis can be measured. At that time a further challenge existed in that the specificity of this gene for B. anthracis was in doubt as bacilli are highly homologous to the extent that B. anthracis, B. cereus and B. thuringiensis may belong to one species.4 Consequently, the CDC employed a combination of biological analyses of culture grown colonies and polymerase chain reactions (PCR) to differentiate bacilli from other bacteria and from each other.5,6 In the former case, presumptive B. anthracis was based on shape (1 to 1.5 by 3 to 5 µm rods), lack of motility, lack of a hemolysis on a sheep blood agar plate, susceptibility to β-lactam antibiotics and to γ-phage lysis, and staining for gram-positive bacteria.5 PCR was then used to produce millions of copies of the 16S rRNA gene so that it could be accurately sequenced and together with the biological analysis, confirm positive, unique identification of B. anthracis.6 The time consuming component of this analysis is the culture growth of cell colonies. Simply put, the fewer the initial number of spores, the longer the time to produce detectable colonies. Only samples collected from surfaces or individuals that had a high probability of being contaminated produced colonies that were evident in 24 hours. However, limiting tests to individuals within the vicinity of where the letter was opened proved insufficient, as three postal workers at the Brentwood, DC, Processing and Distribution Center became infected, two fatally.1,7 Upon notification of their hospitalization, the CDC initiated collection of several hundred environmental (mostly surface) samples and several thousand nasal swab samples from these and associated facilities and their employees. According to the team at the National Naval Medical Center, who processed nearly 4000 samples, current methods of culture growth and analysis were �extremely time-consuming and labor-intensive�.8 From these bioterrorist attacks, it became clear that considerably faster methods of analysis were required. This would expedite assessment of the extent of attack, including the path of such letters from destination back to origination. More importantly, it would minimize fatalities, since it was learned that if exposure is detected within the first few days, the majority of victims can be treated successfully using Ciprofloxacin, doxycycline and/or penicillin G procaine.9

CRC Book Chapter 10 Draft 102

Since the time of these attacks, many methods capable of rapid field analysis have been investigated to augment or replace the laboratory method of growing microorganisms in culture media.10,11 Prominent among these approaches are improved PCR,12 immunoassays,13,14 and detection of calcium dipicolinate as a biochemical signature. PCR employs primers to separate organism-specific nucleic acid sequences and polymerases to amplify the segment until it is detectable. Since the attacks, �real-time� devices have been developed that combine both of these functions in one vessel,15,16,17 and more definitive B. anthracis gene sequences have been identified. Specifically, the genes within the toxin encoding pXO1 plasmid and the capsule-encoding pXO2 plasmid are being targeted for analysis. For pXO1, the three genes that code for the protective antigen, the edema factor and the lethal factor proteins, pagA, cya, and lef, respectively, are targets, while for pXO2, the three genes that code for the protective polypeptide capsule, capA, capB, and capC, are targets.15,16 This work has resulted in the ongoing installation of real-time PCR systems at some 300 regional US Postal offices by 2006 at a cost of $600 million.18,19 Immunoassay methods are also being developed that use competitive binding of the bioagent (as an antigen) and its labeled conjugate for a limited number of antibodies. These methods can be relatively fast (~ 40 minutes) and semi-quantitative, but as yet there is no well-defined anthrax antigen that has been identified,11,13,14,15,20 and as a result, the false-positive rate is unacceptably high.21 Consequently, immunoassay development has shifted to detecting the B. anthracis proteins involved in infection. One of the most successful immunoassays (80% specific) detects the response of immunoglobulin G to the protective antigen of B. anthracis.9 But it requires at least 10 days after the onset of infection for B. anthracis to be detected, and would not substantially improve the odds of successful treatment.9 A number of other methods are focusing on the detection of calcium dipicolinate (CaDPA) or its derivatives as a B. anthracis signature since it has been reported that CaDPA represents 10 to 15% by weight of these spores (Figure 10-1).22-26 This is a valid approach, first because only 13 genera of spore-forming bacteria contain CaDPA,27 but only Bacillus and Clostridium are common (and of interest),26 and second, the most widespread, potentially interfering spores, such as pollen and mold spores, do not. Relatively fast methods have been developed to chemically extract the acid of CaDPA, dipicolinic acid (DPA),28 and then to detect it directly by mass spectrometry,29,30 fluorescence,31 or indirectly by luminescence.28,32 Although mass spectrometry provides a relatively high degree of discrimination and sensitivity, it still requires significant time due to sample preparation. Hot dodecylamine (DDA) has been used to extract DPA and form a highly luminescent complex with terbium.28 Although measurements have been performed in as little as five minutes, it was found that as many as three concentration-dependent complexes can form, each with different lifetimes. This, coupled with the fact that the Tb3+ cation produces the same luminescence spectrum, makes determinations of low spore concentrations problematic. Figure 10-1. Illustration of a Bacillus spore with major components indicated, and chemical structure of calcium dipicolinate. Deprotonated dipicolinic acid is shown within the brackets.

Core

(calcium dipicolinate)

Outer Core Wall

Cortex

Exosporium

Spore Coat

DNA

RibosomesC CO O

O ON

- -

Ca2+Inner Core

Wall

CRC Book Chapter 10 Draft 103

An alternative method, Raman spectroscopy, is attractive in that very small samples can be measured without preparation. The sample need only be placed at the focal spot of the excitation laser and measured. Moreover, the rich molecular information provided by Raman spectroscopy usually allows unequivocal identification of chemicals and biologicals. As early as 1974 the Raman spectrum of Bacillus megaterium was measured and shown to be dominated by CaDPA.33 However, the spectrum was collected using pure spores and took hours to acquire. By 1992, the improvements in Raman instrumentation and the use of resonance enhancement increased limits of detection dramatically and reduced analysis time to less than 1 hour.34 Recently, Raman spectroscopy has been used in combination with chemometrics to differentiate bacteria at the genus level35 and the species level (Enterococcus).36 And in the past year, resonance Raman spectroscopy has been used to quantify the amount of CaDPA in Bacillus spores37 and to differentiate six bacilli species.38 B. anthracis, however, was not included in the latter study. Microscopy and Raman spectroscopy have also been combined to perform measurements of single-digit numbers of Bacillus spores.39,40 However, these measurements required time consuming efforts to locate the spores in the Raman excitation beam41 and photobleaching to deplete available ground states and thereby reduce spectral obscuration by fluorescence.39 Moreover, the fragility of such instrumentation confines its use to the laboratory. Nevertheless, a process Raman analyzer has been used to measure spores captured from a mail sorting system.42 However, the field measurements were limited to fairly large, milligram samples. In related efforts several researchers have been investigating the capability of surface-enhanced Raman scattering (SERS) to measure bacteria, including bacilli and dipicolinic acid (see Efrima in this chapter for a comprehensive review). The approach and ultimately the success of these efforts not only depend on the instrumentation, but also on the specific terrorist scenario being addressed. This has significant implications for the choice of sampling. For example, detecting a plume of spores released from an airplane is very different than detecting spores in envelopes passing through a mail sorting machine. Here the focus is the detection of spores on surfaces to assess the extent of an attack. At present there are no guidelines defining the required sensitivity. However, an extensive number of surface samples were collected from the Brentwood, DC, mail Processing and Distribution Center and their analysis can be used as a guide to estimate sensitivity requirements.43 This analysis determined that the highest concentrations of spores, not surprisingly, were in the immediate vicinity of delivery barcode sorter machine number 17, which processed both letters. Analysis of dust above, within 30 meters, and 30-60 meters of this machine recorded average values of 310, 67, and 10 CFU/in2. Since the last average value included measurements that detected �zero� spores, it lacks the certainty of the other values. Consequently, and somewhat subjectively, we have chosen the middle value, 67 CFU/in2 (10 spores/cm2), as a minimum requirement for measurement sensitivity. This value should not be construed as a definition of lethality. Additional measurement requirements include the ease of sampling and speed of analysis. Based on the 2001 attacks, we consider a minimum requirement of 500 measurements per 24 hours as reasonable. Of course more than one analyzer could be used to accomplish this, but the fewer the analyzers, the lower the cost and number of operators. If one analyzer is used, then the required measurement time would be less than 3 minutes. This would include the time to collect, deliver, measure, and analyze the sample. This suggests that sampling should involve a method or device to rapidly collect the sample (e.g. a wet swab or vacuum system) and deliver it to the measurement compartment of the analyzer. It also suggests that the analyzer should be portable to minimize or eliminate sample delivery time. With these criteria in mind, we have been developing a three-step method to detect dipicolinic acid extracted from surface spores by surface-enhanced Raman spectroscopy. The first step employs hot dodecylamine to break apart the spores and release CaDPA into solution as DPA. The second step employs single-use, disposable, sol-gel filled capillaries to separate the DPA from other cell components and simultaneously deliver it to the SERS-active metal particles. The third step employs a portable Raman analyzer to measure the SERS spectrum and to identify and quantify the spores, if present. Development of this three-step method and measurements of Bacillus spores on surfaces are presented.

CRC Book Chapter 10 Draft 104

II. EXPERIMENTAL Dipicolinic acid (2,6-pyridinedicarboxylic acid), dodecylamine, and all chemicals used to produce the silver-doped sol-gels were obtained at their purest commercially available grade from Sigma-Aldrich (Milwaukee, WI) and used as received. Calcium dipicolinate was prepared from disodium dipicolinate (Na2DPA), which was prepared from DPA according to previous publications.34 Bacillus cereus, B. subtilis, and B. megaterium bacteria were grown on nutrient agar plates at 30ºC for 7 days (see Reference 34 for details). The vegetative cells were placed into distilled water and lysed by osmotic pressure. The resultant spores were collected by scraping them into distilled water and pelleting them by centrifugation at 12,100 x g for 10 minutes. The spores were washed 4 more times in distilled water by centrifugation. The spore pellet was re-suspended in distilled water and lyophilized, and scraped into glass vials for Raman spectral measurements. Approximately 1 gram each, determined to be 99% pure by microscopic observation, was produced for this study. The density of the spores varied from 0.06 to 0.11 g/mL, indicating a high amount of entrained air. An initial stock solution of 20 mg of DPA in 20 mL HPLC grade water (Fischer Scientific, Fair Lawn, NJ) was prepared for the pH study. The pH of this solution was 2.45 as verified using a pH electrode (Corning 314 pH/Temperature Plus, Corning, NY) that had been calibrated with pH 4.00, 7.00, and 10.00 buffer solutions (Fischer Scientific). For all experiments a single 2-mL glass vial coated with silver-doped sol-gel was used (Simple SERS Sample Vial, Real-Time Analyzers, Inc., Middletown, CT). The vial was never moved from the sample holder to ensure that the same portion of silver-doped sol-gel was examined. Two pH series were performed. First, 2 mL of the stock solution was added to the vial and measured. Then the 2 mL solution was returned to the stock solution and made basic using 0.1 M KOH. Prior to re-addition of the solution to the SERS-active vial, the vial was first rinsed three times with distilled water, then twice with the new solution prior to SERS measurement. This procedure was followed to obtain spectra at pH 3.55, 4.33, 4.87, 5.59, 10.69 and 11.66. Next the solution was brought to a pH of 2.00 by adding 0.1 M HNO3, and the spectrum was recorded. Again KOH was added dropwise to make the solution more basic. Spectra were obtained at pHs of 3.83, 5.10, 7.35 and 8.22. Next HNO3 was added dropwise so that spectra could be obtained at pHs of 2.19, 1.71, 1.35 and 1.17. Throughout this process, no more than 20 drops of acid or base were added, and therefore the concentration was diluted by no more than 10%. For concentration measurements, a second stock solution of DPA was prepared as above and used to prepare all lower concentration samples by serial dilution using HPLC grade water. DPA extractions were initially developed by adding pre-weighed spores into a vial containing hot DDA dissolved in ethanol. Optimization of the DDA concentration and temperature are described below. Once refined, 78 oC, 50 mM DDA was added dropwise to spore particles placed on a glass plate. After 1 min, the degraded spore sample was drawn into a SERS-active capillary for measurement. The SERS-active vials were prepared according to published procedures,44 using a silver amine precursor to provide the metal dopant and an alkoxide precursor to provide the sol-gel matrix. The silver amine precursor consisted of a 5/1 v/v ratio of 1N AgNO3 to 28% NH3OH, while the alkoxide precursor consisted of a 2/1 v/v ratio of methanol to tetramethyl orthosilicate (TMOS). The alkoxide and silver amine precursors were mixed in an 8/1 v/v ratio, then 140 µL were introduced into 2 mL glass vials, which were then spin-coated. After sol-gel formation, the incorporated silver ions were reduced with 0.03M NaBH4. The SERS-active capillaries were prepared in a similar manner with the following modifications. The alkoxide precursor employed a combination of methyltrimethoxysilane (MTMS) and TMOS in a v/v ratio of 6/1, which was mixed with the amine precursor in a v/v ratio of 1/1. Approximately 15 µL of the mixed precursors were then drawn into a 1-mm diameter glass capillary coating a 15-mm length. After sol-gel formation, the incorporated silver ions were again reduced with dilute sodium borohydride. All Raman spectroscopy measurements were performed using 785 or 1064 nm laser excitation and Fourier transform Raman spectrometers (Real-Time Analyzers, model IRA-785 and IRA-

CRC Book Chapter 10 Draft 105

1064, East Hartford, CT).45 For pure Na2DPA, CaDPA, and the spore samples 1064 nm excitation was used, for pure DPA and DPA solutions both 785 and 1064 nm laser excitation were used, while for all DPA SERS measurements, solutions or extractions, 785 nm laser excitation was used. Fiber optics were used to deliver the excitation beam to the sample probe and the scattered radiation to the interferometer (2 m lengths of 200 and 365 µm core diameter, respectively, Spectran, Avon, CT). For 1064 nm excitation, a 24 mm diameter f/0.7 aspheric lens focused the beam to a 600 µm spot on the sample and to collect the scattered radiation back along the same axis. An f/2 achromat was used to collimate laser beam exiting the source fiber optic, while a 4 mm prism was used to direct the beam through an f/0.7 aspheric lens that focused the beam to a 600 micron spot on the sample. The scattered radiation was collected back along the same optical axis, while a second f/2 lens focused the beam into the collection fiber optic. A short pass filter was placed in the excitation beam path to block the silicon Raman scattering (RS) generated in the source fiber from reflecting off sampling optics and reaching the detector. A long pass filter was placed in the collection beam path to block the sample Rayleigh scattering from reaching the detector. For 785 nm excitation, a similar optic probe was used, except a dichroic filter was used to reflect the laser light to sample and pass the Raman scattered radiation to the collection fiber. In this case the beam was focused to a 300 µm spot on the sample. Also, appropriate short and long pass filters were used for this wavelength. All spectra presented were collected using 8 cm-1 resolution. In the case of Raman spectral measurements of spores, the samples were placed on a glass slide with the probe aimed downward. In the case of SERS-active vials or capillaries, the samples were mounting horizontally on an XY positioning stage (Conix Research, Springfield, OR), so that the probe aimed upwards and the focal point of the aspheric lens was just inside the vial or capillary. For the SERS concentration and extraction measurements, nine spectra were recorded along the length of the capillary with 1 mm spacing. As a practical approach to minimizing the variability associated with the SERS activity as a function of sample position, the three high and three low intensity spectra were discarded, while the three median spectra were averaged and reported. Relative standard deviations for all concentrations are reported as percent standard deviation in Table 10-2.

III. RESULTS AND DISCUSSION The present application begins with a Raman spectral analysis of Bacillus spores with regards to contributions from calcium dipicolinate. The primary CaDPA peaks occur at 659, 821, 1014, 1391, 1446, 1573, 3062, and 3080 cm-1 in the spore spectrum (Figure 10-2), and can be assigned to a CC ring bend, a CH out-of-plane bend, the symmetric pyridine ring stretch, an OCO symmetric stretch, a symmetric ring CH bend, an asymmetric OCO stretch, and the CH symmetric and asymmetric stretches, respectively (Table 10-1).42 The remaining peaks can be assigned to protein modes associated with the peptidoglycan cell wall, such as amino acids and peptide linkages (amide modes).33,34,46,47,48 The former include peaks with little intensity at 821, 855, 900 cm-1, which are assigned to several CC bending modes, as well as the phenylalanine modes that appear at 1003 and 1598 cm-1. The latter include the amide I peak at 1666 cm-1, which is primarily a C=O stretch, and amide III combination peaks at 937, 1241, and 1318 cm-1, which are various CC and CN stretching combinations (peak positions are given for B. cereus). In several cases, protein and CaDPA vibrational modes occur at or close to the same frequency, such as the 821 and 1446 cm-1 peaks. Next, the amount of CaDPA available in a spore that could be measured as DPA was considered. Although it is often stated that bacilli spores contain 10-15% calcium dipicolinate by weight,22-26 this value has been reported as low as 1%.49 Since this amount will be used to calculate the number of spores measured, it is important to have as accurate a number as possible. For this reason, the Raman spectra of Bacillus subtilis, B. megaterium, and B. cereus were acquired (Figure 10-3). In fact it was found that the most obvious differences between the spectra for the three bacilli are the CaDPA peaks. In particular, the 1014 cm-1 peak noticeably changes

CRC Book Chapter 10 Draft 106

intensity, especially when compared to the neighboring phenylalanine peak at 1003 cm-1. If it can be assumed that the composition of these bacilli is very similar, then it may be assumed that the

A B

A B C D

In

tens

ity (a

rbitr

ary

units

)

1000 1500 2000 2500 3000 Wavenumber (∆cm-1)

Figure 10-2. RS spectra of A) Bacillus cereus spores and B) calcium dipicolinate. Conditions: 500 mW of 1064 nm at the sample, 5-min acquisition time.

In

tens

ity (a

rbitr

ary

units

)

800 1000 1200 1400 1600 1800 Wavenumber (∆cm-1) Figure 10-3. RS spectra of A) B. subtilis, B) B. megaterium, C) B. cereus, and D) CaDPA. Conditions as in Fig. 10-2.

CRC Book Chapter 10 Draft 107

relative phenylalanine concentration is nearly constant and its Raman peak can be used as an internal intensity standard. (The amide I peak at 1666 cm-1 could also be used.) Using the ratio of the CaDPA and phenylalanine peak heights suggests then that the salt concentrations for B. megaterium and B. cereus are 1.85 and 2.05 times that of B. subtilis. In the latter case, a recent study using resonance Raman spectroscopy by one of us (Sperry) of the same sample concluded that the CaDPA peak intensity corresponded to 6-7 weight percent.37 This suggests that the CaDPA weight percent for the B. megaterium and B. cereus spore samples are 11-13 and 12.5-14.5 wt%, respectively, or in the case of DPA, 9-11 and 10-12 wt% (based on MW). It should be noted that the differences between these bacilli do not imply that the CaDPA concentrations are species specific. It is more likely that experimental conditions during the original growth of the bacteria, such as time, temperature, or available nutrients, influenced the extent of sporulation. Consequently, any calculations of the number of spores based on DPA content should assume a range of at least 5-13 weight percent of the spores. As a practical matter 10±5% will be used here. For comparison purposes, the spectra of CaDPA and DPA are shown in Figure 10-4, along with Na2DPA, while the observed spectral peaks with vibrational mode assignments are listed in Table 10-1. The assignments for both CaDPA and DPA, based on literature,50,51 were used to assign the peaks observed for Na2DPA. Both DPA and Na2DPA contain unique peaks with significant intensity at 760 and 1730 cm-1, respectively. Since neither peak is observed in the spectrum of CaDPA, it can be concluded that this sample does not contain either chemical as an impurity.

Next, dipicolinic acid was analyzed by SERS. The assignment of SERS peaks to vibrational modes is less straightforward than for RS peaks due to the metal-to-molecule surface interactions that shift and enhance various vibrational modes to different extents (see Otto Chapter 1). Furthermore, it is usually found that RS spectra of analytes in solution more closely match the SERS spectra than in the solid-state. However, it is usually beneficial to acquire and examine both when making assignments. Since DPA dissolves in water only sparingly, 1N KOH was used to dissolve 80 mg/mL. The RS spectrum of the solution phase is largely the same as the solid phase except for some minor changes in peak frequencies, intensities, and widths (Table 10-1).

A B C

In

tens

ity (a

rbitr

ary

units

)

1000 1500 2000 2500 3000 Wavenumber (∆cm-1)

Figure 10-4. RS spectra of A) dipicolinic acid B) disodium dipicolinate, and C) calcium dipicolinate. Conditions as in Fig. 10-2.

CRC Book Chapter 10 Draft 108

Table 10-1. Tentative Raman vibrational mode assignments for dipicolinates. B. Cereus CaDPA Na2DPA DPA

solid DPA solution

SERS Tentative Assignments

403 413 405 433 425 458 CC ring benda 478 499 489 C-CO2 stra 575 573 567 659 661 650 646 652 657 CC ring benda 696 760 (795) HO-C=O in-plane defb 805 801 821 820 814 822 812 CC strd, CaDPA CH out-of-

plane defb 855 857 856 853 858 CC strd 900 897 CC strd 925 923 CaDPA 937 CC str + amide IIId 1003 phe sym ring str.c,d 1014 1015 1003 997 1001 1006 sym ring breathc 1029 1077 1086 1079 1085 1087 trigonal ring breathingb 1150 1150 1147 1153 1154 1157 CH benda 1199 1185 1179 1191 1184 1227 1230 1241 amide III (b)c,d 1274 1256 1271 CC strb 1293 1296 1285 1318 CH2 bend, amide IIId 1324 C=O str 1377sh 1383 CH benda 1391 1398 1393 1386 1381 OCO sym str 1446 1447 1437 1445 1438 (1426) CH2 bend,d CaDPA ring CH

benda or CC strb 1466 1464 1461 1466 ring CC str 1573 1568 1569 1575 1572 1567 OCO asym str 1583 1583 1589 (1590) CC ring str 1598sh phe sym ring str.c,d 1643 1634 1643 carboxylatee 1666 amide Ic,d 1704/30 C=O str (doublet)

2879sh CH3 sym strf 2934 CH2 antisym strf 2968sh CH3 antisym strf 3019 3021 3062 3060 3070 aromatic CH sym strf 3080 3088 3084 3098 CH antisym strf 3137 3150 CH strf 3302 amide NH strc,d

a is from Ref. 51, b is from Ref. 50, c is from Ref. 33 d is from Ref. 47, e is from Ref. 34, f is from Refs 46 and 48. Notably, the 760 cm-1 in the solid phase is completely absent in the solution phase, while a new peak at 1386 cm-1 appears in the solution phase. The former peak is likely associated with carboxylic acid groups (e.g. HO-C=O deformation), while the latter peak is likely associated with deprotonated carboxylic acid groups (e.g. O-C-O stretch.) The latter assignment is consistent with

CRC Book Chapter 10 Draft 109

a sample pH of 10 due to the 1N KOH. The former assignment is supported by the fact that the peak does not disappear when DPA is dissolved in the aprotic solvents dimethylsulfoxide or N,N-dimethylformamide. The SERS spectrum of 1 g/L DPA in water is more like the solution than solid phase as shown in Figure 10-5. The quality of this SERS spectrum is considerably better than the first reported SERS spectrum of dipicolinic acid obtained on a silver electrode in an electrolytic cell.52 In fact, not only are most of the peaks of the solution phase RS spectrum observed in the SERS spectrum, but peaks shift no more than 10 cm-1 and change little in relative intensity. These similarities suggest a weak molecule to silver surface interaction. The RS to SERS shifts of the major peaks are: 652 to 657 cm-1, 822 to 812 cm-1, 1001 to 1006 cm-1, 1386 to 1381 cm-1, 1438 to 1426 cm-1, 1572 to 1567 cm-1. The SERS peaks are assigned according to CaDPA above and literature as follows:33,34 the 1006 cm-1 peak is assigned to the symmetric ring stretch, the 1381 cm-1 peak to the O-C-O symmetric stretch, the 1426 cm-1 peak to the symmetric ring C-H bend, and the 1567 cm-1 peak to the asymmetric O-C-O stretch. The greatest difference between the RS and SERS spectra is the appearance of a new band in the latter at 795 cm-1 (see below).

Figure 10-5. A) RS spectrum of 80 mg DPA in 1 mL 1N KOH in a glass capillary. B) SERS spectrum of 1 mg DPA in 1 mL water in a silver-doped sol-gel filled glass capillary. Spectral conditions: A) 450 mW of 785 nm, 5-minute acquisition time and B) 150 mW of 785 nm, 1-minute acquisition time; both 8 cm-1 resolution. Next the pH dependence of both the measurement and analyte was considered. This could be significant if an acid or a base is used to digest spores and extract the CaDPA. It is widely known that the pH of the solution can have an effect on the SERS signal,53,54 particularly in the case of metal colloids where pH affects the extent of aggregation,53 which in turn affects the plasmon field and the Raman signal enhancement. Other SERS-active media are more tolerable to pH changes, such as metal coated spheres and posts, or silver-doped sol-gels, as used here. Although these sol-gels may not be affected by pH, the analyte is a diprotic acid and the neutral and ionic forms of DPA, DPA-, or DPA=, must be considered. These species may interact with the silver quite

A B

In

tens

ity (a

rbitr

ary

units

)

400 600 800 1000 1200 1400 1600 1800 Wavenumber (∆cm-1)

CRC Book Chapter 10 Draft 110

differently and consequently influence the amount each vibrational mode is enhanced. For example, it might be expected that DPA= will interact more strongly with electropositive silver increasing the chemical component of the SERS mechanism (see Otto Chapter 1). Furthermore, added enhancement might be expected for the vibrational modes of the deprotonated carboxylic acid groups that participate in this interaction, or for modes that are favorably aligned perpendicular to the surface due to this interaction. The relative concentrations of DPA, DPA-, and DPA= can be determined at any pH as long as the pKas are known and the initial concentration. According to Lange's Handbook of Chemistry, the pKas are 2.16 and 6.92, and the deprotonation reactions are: DPA !" DPA- + H+ pK1a = 2.16 Reaction 1 DPA- !" DPA= + H+ pK2a = 6.92 Reaction 2 The relative concentrations can then be determined by expressing [DPA] and [DPA=] in terms of [DPA-] using Reactions 1 and 2, and summing all three to equal the total starting concentration, here 1 g/L, viz: [DPA] + [DPA-] + [DPA=] = 1 g/L Equation 1 substituting from Reactions 1 and 2: ([H+][DPA-])/K1a + [DPA-] + (K2a[DPA-])/[H+] = 1 g/L Equation 2 rearranging: [DPA-] = 1 g/L /(1+[H+]/K1a + K2a/[H+]) Equation 3 As shown in Figure 10-6, at pH less than pK1a DPA dominates, at pH between the pKas DPA- dominates, and above pK2a DPA= dominates. Figure 10-7 shows SERS spectra of DPA for pH 4.87, 5.59, 7.35, 8.22, 10.69, and 11.66 with spectra of the 800 cm-1 region for pH 1.35, 1.71, 2.19 and 3.83 (inset). Overall there is only a modest decrease in intensity for most of the peaks as a function of pH. For example, the 1006 cm-1 peak assigned to the pyridine ring stretching mode decreases by ~7% from pH 2 to 11. The greatest changes observed, yet still modest, are in the peak intensities at 795, 812, 1567, and 1590 cm-1 between pH 1.3 and 5.5. These peaks change intensity as pairs. The 795 cm-1 peak loses intensity as the pH becomes basic, while the 812 cm-1 peak gains a little intensity. Similarly, the 1567 cm-1 peak loses intensity as the pH becomes basic, while the 1590 cm-1 peak gains intensity. The intensities of the former pair are plotted as a function of pH in Figure 10-6. The peak heights were divided by the peak height of the 1006 cm-1 peak at each pH and then scaled with the lowest value set to 0 and the highest to 1 g/L. As can be seen the 795 cm-1 peak tracts the DPA concentration, while the 812 cm-1 peak tracts the DPA- concentration. The former peak is likely associated with carboxylic acid groups, just as in the case of the 760 cm-1 peak in the solid phase RS spectrum of DPA. However, a 35 cm-1 shift is somewhat inconsistent with a weak analyte-to-surface interaction. It is also apparent in Figure 10-6 that the concentrations of DPA and DPA- based on the 795 and 812 cm-1 peak intensities are shifted to the basic side of the predicted curves. This shift may be due to the silver surface influencing the carboxylic acid dissociation energy. Or the peaks may contain contributions from the DPA= species. Although clarifying this point will require further measurements, the most important conclusions from this data is that the SERS intensity for most of the prominent DPA peaks change little as a function of pH, and that the silver-doped sol-gels do not appear to influence the measurement to any significance.

CRC Book Chapter 10 Draft 111

Figure 10-6. DPA and its anion concentrations as a function of pH (lines). A) The 1006 cm-1 peak intensity is shown as measured, but scaled to a 0 to 1 g/L concentration range. B) The 795 and 812 cm-1 band intensities are normalized to the 1006 cm-1 peak intensity and then scaled. These two peaks appear to represent DPA and DPA=, respectively, but both with some DPA- character.

0

0.001

0.002

0.003

0.004

0.005

0.006

0.007

0.008

0.009

0.01

0 2 4 6 8 10 12 14pH

Con

cent

ratio

n, m

M

DPADPA-DPA=1006795/1006812/1006

DPA DPA=DPA-

1.0 0.7 1.0 0.8 0.6 0.4 0.2 0

Con

cent

ratio

n (g

/L)

A B

■ 1006

pH

1.35 1.71 2.19 3.83

pH

4.87 5.59 7.35 8.22 10.69 11.66

In

tens

ity (a

rbitr

ary

units

)

400 600 800 1000 1200 1400 1600 1800 Wavenumber (∆cm-1)

Figure 10-7. SERS spectra of 1 mg/mL dipicolinic acid as a function of pH. The spectral intensities have been normalized to the 812 cm-1 band. Inset: Expanded view of low wavenumber region. Spectral conditions: 100 mW 785 nm, 100 scans (44-sec).

CRC Book Chapter 10 Draft 112

Next, the response of the SERS intensity for DPA as a function of concentration was examined. A preliminary calibration curve was prepared by measuring 100, 50, 20, 10, 5, 2, 1, 0.5, 0.2, 0.1, 0.05, 0.02, and 0.01 mg/L samples. Figure 10-8 shows SERS spectra for 100, 1, 0.1 and 0.01 mg/L samples measured using 100 mW of 785 nm and 1-min acquisition time. It can be seen that even at 10 µg/L the signal-to-noise ratio is quite good. The SERS intensity was taken as the peak height at 1006 cm-1 minus the value at 950 cm-1 as the baseline. For each concentration, a different capillary was used. Spectra were measured at nine points along the length of each capillary and the median values are plotted in Figure 10-9. It is obvious that the response is not linear, in that the peak heights change from 0.2 to 1.5, while the concentration changes over 4 orders of magnitude. This Langmuir isotherm response is typical for SERS substrates where signal intensity is a function of available silver surface area.55

Figure 10-8. SERS spectra of DPA in water at A) 100, B) 1, and C) 0.01 mg/L (100 pg in 10 µL sample) using the SERS-active capillaries, 100 mW of 785 nm and 1-min acquisition time. All of these values were also used to estimate limits of detection (LOD), defined as the concentration that produces a signal three times as intense as the baseline noise. The signal was taken as the height of the 1006 cm-1 peak, while the noise was the relative standard deviation of baseline noise measured between 50 and 150 cm-1. The LODs are for 1-min measurements using 100 mW of 785 nm laser excitation and 8 cm-1 resolution. As Table 10-2 indicates, the lower the measured concentration, in general, the lower the predicted LOD. Note that the 10 µg/L (0.01mg/L) sample suggests that 0.7 µg/L can be measured (S/N equaled 33 for the 1006 cm-1 peak). This is consistent with the fact that attempted measurements of 1 µg/L samples did yield spectra, but not in every case. It is also worth noting that only 10 µL samples were used to generate the spectra, or in the case of the 10 µg/L sample, the equivalent of 100 pg of DPA. Finally, an enhancement factor for DPA can be estimated by comparing the measurement conditions and signal intensities for the 10 µg/L SERS and 80 g/L RS. The spectra are plotted on the same scale in Figure 10-10. The 1006 cm-1 peak heights are nearly identical at 0.20 and 0.173 (arbitrary units), while the laser power at the sample was somewhat different at 150 and 450 mw

A B

C

In

tens

ity (a

rbitr

ary

units

)

600 800 1000 1200 1400 1600 Wavenumber (∆cm-1)

CRC Book Chapter 10 Draft 113

Figure 10-9. Plot of SERS intensity of 1006 cm-1 band of DPA as a function of concentration using 100 mW of 785 nm. Line connects average value at each concentration. Inset includes 10 and 100 mg/L data. Table 10-2. Estimated limits of detection in terms of mg DPA per L water and corresponding spores per 0.1mL DDA.

Conc (mg/L) spores/0.1mL Sig (ave) Std Dev RSD (%) Noise S/N LOD factor LOD (mg/L) LOD (spores/0.1mL)0.01 1229 0.14 0.12 86.68 0.0033 41.4 13.8 7.24E-04 890.02 2457 0.25 0.05 20.00 0.0042 59.5 19.8 1.01E-03 1240.05 6143 0.31 0.08 25.99 0.0043 72.9 24.3 2.06E-03 2530.1 12,285 0.40 0.10 25.00 0.0047 85.1 28.4 3.53E-03 4330.2 24,570 0.50 0.15 30.00 0.005 100.0 33.3 6.00E-03 7370.5 61,425 0.56 0.14 25.17 0.006 92.8 30.9 1.62E-02 19861 122,850 0.74 0.15 20.24 0.0067 110.9 37.0 2.70E-02 33222 245,700 0.83 0.06 7.78 0.008 103.3 34.4 5.81E-02 71335 614,251 0.93 0.16 16.90 0.0067 138.8 46.3 1.08E-01 1327610 1,228,501 1.02 0.17 16.75 0.0096 106.3 35.4 2.82E-01 34687100 12,285,012 1.38 0.16 11.50 0.0122 113.1 37.7 2.65E+00 325820

~1 (exp 1)* 100,000 0.70 0.14 20.11 0.0055 126.7 42.2 2.37E-02 2368~0.015 (exp 2) 1500 0.20 0.10 52.20 0.0081 24.7 8.2 1.82E-03 182 * Approximate concentrations for surface measurements, see below.

for the SERS and RS, respectively. In both cases, 1-mm capillaries were used to hold the samples, as well as the same sample optics. Taking the concentration into account yields an estimated enhancement factor of 2.4x107. It is difficult to determine the precise number of molecules in the field of view for the sol-gel, and this number may represent better than average enhancement, i.e. better than 106, or it may reflect the ability of the sol-gel to concentrate the sample. In either case, the measurement of 10 µg/L suggests that 10 ng of spores in a 100 µL solution of a digesting chemical can be measured; assuming all of the CaDPA was made available as DPA (10%). Recent estimates suggest that this mass corresponds to 1000 spores.2

0

0.2

0.4

0.6

0.8

1

1.2

0 1 2 3 4 5 6DPA Concentration (mg/L)

1006

Ban

d In

tens

ity

0

0.2

0.4

0.6

0.8

1

1.2

1.4

1.6

0 20 40 60 80 100 120DPA Concentration (mg/L)

1006

Ban

d In

tens

ity

CRC Book Chapter 10 Draft 114

Figure 10-10. A) SERS and B) RS spectra of DPA plotted on the same scale, but offset. Conditions: A:B 10-5:80 g/L and 150:450 mw of 785 nm.

Next, methods were developed to rapidly extract CaDPA as DPA from Bacillus spores. Initially, DPA was obtained from B. cereus spores following the procedure of Pellegrino et al.28 Specifically, a 2 mg sample was placed in 2 mL of 5 mM dodecylamine in ethanol that was heated and maintained at 78 oC for 40 minutes. Approximately 10 µL of this solution was drawn into a SERS-active capillary and measured. Since SERS spectra of DPA were readily observed, shorter heating periods, higher DDA concentrations and smaller spore masses, were examined. In due course it was found that the spores could be broken apart in 1 minute using 78 oC 50 mM DDA. Two series of experiments were performed using µg spore masses. In the first series, three 100 µg samples of B. cereus were weighed and then each diced into ~ 10 equal parts, producing 30 measurable particles consisting of ~10 µg or 1 million spores each (Figure 10-11). To each particle 100 µL of 78 oC 50 mM DDA in ethanol was added. After 1 minute, approximately 10 µL of the solution was drawn into a SERS-active capillary and measured. Figure 10-12 shows a representative spectrum from one of these capillaries using a 1-min acquisition time. The primary DPA peaks at 657 cm-1, 812 cm-1, 1006 cm-1, 1381 cm-1, and 1426 cm-1 are easily observed, even in the case of a 2-sec scan. Furthermore, an attempted measurement of 50 mM DDA (without sample) did not produce a spectrum that might interfere with the measurement (Figure 10-12C). The amount of DPA that was extracted can be estimated to be between 0.5 and 5 mg/L by comparing the 0.7 signal intensity of the 1006 cm-1 band to that measured for DPA in water (see Table 10-2). In fact this intensity is closest to that obtained for the 1 mg/L samples. This value can be used to estimate the number of spores in the 100 µL DDA sample. Assuming, as stated above, that a spore contains approximately 10% DPA by weight, and that 100 spores have a mass of ~1 ng, then this corresponds to 100,000 spores per 100 µL DDA or~10% of the spores in the prepared particles. This low percentage could be due to incomplete degradation of the spores by DDA, inefficient collection of the sample from the surface, inefficient transfer of the DPA to the silver particles, or saturation of the silver particle surface. Based on the latter possibility, methods were explored to produce significantly lower surface concentrations.

A B

In

tens

ity (a

rbitr

ary

units

)

600 800 1000 1200 1400 1600 Wavenumber (∆cm-1)

CRC Book Chapter 10 Draft 115

Figure 10-11. Sample preparation includes A) three initial 100 µg B. cereus spore samples, B) addition of 100 µL 78 oC 50 mM DDA to ~10 µg portion, C) drawing 10 µL into SERS-active capillary, and D) mounting capillary in Raman analyzer sample compartment.

Figure 10-12. SERS spectra of DPA extracted from ~10 µg B. cereus particle using 100 µL of 50 mM hot DDA acquired in A) 1 minute and B) 2 seconds. C) Attempted SERS spectrum of 50 mM hot DDA in ethanol using silver-doped sol-gel coated glass capillary acquired in 1 minute. Spectral conditions: 150 mW of 785 nm, 8 cm-1 resolution. In the second series of experiments, anhydrous ether was used to disperse spores on a surface to the point of being invisible to the unaided eye. In this series of experiments a 100 µg B. cereus

A B C

A B

C D

In

tens

ity (a

rbitr

ary

units

)

600 800 1000 1200 1400 1600 Wavenumber (∆cm-1)

CRC Book Chapter 10 Draft 116

sample was divided into three near equivalent particles. To each particle, 600 µL of ether was added and allowed to dry. The dispersed spores and ether produced a solvent ring ~5 cm in diameter with the majority of the spores at the edge. A non-cotton swab was used to collect the residual spores in the center, ~10 cm2, of this area. The swab was added to a vial containing 100 µL of 78 oC 50 mM DDA. After 1-min, ~10 µL of this solution was extracted into a SERS-active capillary and measured as before (Figure 10-13). The peaks in the SERS spectrum, acquired in 1-min, are ~ 1/4th the intensity of those in the previous experiment, and the 1006 cm-1 peak height of 0.20 suggests a collected sample of 15 µg/L based on the concentration curve (Figure 10-9, Table 10-2). And according to the assumptions above, this intensity corresponds to ~1500 B. cereus spores in 100 µL DDA. This clearly suggests that nearly all of the dispersed spores were carried to the edge of the solvent ring. Also, the error in this measurement, based on the ±50% DPA weight percent in spores (see above), and the ±30% concentration (see Table 10-2), is as much as ±1000 spores. The S/N of 25 for this measurement was also used to estimate an LOD of 1.8 ng/mL or 180 B. cereus spores in 100 µL DDA. This LOD is twice that estimated using pure DPA samples of similar concentration, which may indicate that only half of the DPA was successfully extracted from the spores and measured. However, considering the large uncertainly in the initial number of spores swabbed, this can not be firmly concluded.

Figure 10-13. SERS spectrum of representative 10 µg spore sample (0.1 µg DPA) per 100 µL DDA (from Figure 10-12A) compared to B) SERS spectrum of surface swab. Relative intensities suggest ~1.5 ng DPA per 100 µL DDA (see Figure 10-9, Table 10-2 and text) . Spectral conditions for B are 100 mW of 785 nm and 1-min acquisition time. Finally, the ability to assess surface contamination is considered. The stated goal of 1 measurement per 3 minutes has been met, since, as described these measurements required 1 minute to break apart the spores with DDA and release DPA and 1 minute to acquire the SERS spectrum. Sample manipulation, which included adding the DDA, drawing it into the SERS-active capillary, and placement of the capillary into the sample holder, required less than 30 seconds. The above measurement suggests that 150 spores/cm2 were measured, with an LOD of

A B

In

tens

ity (a

rbitr

ary

units

)

600 800 1000 1200 1400 1600 Wavenumber (∆cm-1)

CRC Book Chapter 10 Draft 117

~18 spores/cm2. This compares favorably to the stated goal of 10 spores/cm2. Unfortunately, the actual number of spores on the sampled surface has not been verified by an independent measurement.

IV. CONCLUSIONS We have demonstrated that by combining rapid extraction of dipicolinic acid from Bacillus cereus spores with chemical identification by surface-enhanced Raman spectroscopy, bacilli spores on a surface can be identified in as little as 3 minutes. This includes the time required to add hot dodecylamine to the spores, the time required to collect the DPA into the SERS-active capillary, and the time required to acquire the DPA SERS spectrum. Based on a concentration curve for DPA in water and the assumption that the B. cereus spores contained 10 wt % DPA, and each spore has a mass of 10-11g, we estimate that as few as 150 spores/cm2 were measured. However, it is cautioned that the accuracy of this measurement has not been verified by independent means, and the precision is also low with an error in the measurement as high as 50% of the stated value. Nevertheless, improvements in accuracy and precision should allow the use of SERS-active capillaries to measure surfaces and map distribution of anthrax endospores in mail distribution facilities or other environments should another verified attack occur.

V. ACKNOWLEDGEMENTS The authors are grateful to Chetan Shende and the National Science Foundation for the development of the SERS capillaries (DMI-0215819). The authors would also like to acknowledge the support of the U.S. Army for the development of the analysis method (DAAD13-02-C-0015, Joint Service Agent Water Monitor Program).

VI. REFERENCES 1. Jernigan, J.A., Stephens, D.S., Ashford, D.A, et al. Bioterrorism-Related Inhalational Anthrax: The First 10 Cases

Reported in the United States, Emerg. Infect. Dis., 6, 933, 2001. 2. Inglesby, T.V., Henderson, D.A., and Bartlett, J.G., Anthrax as a biological weapon: updated recommendations for

management, JAMA, 287, 2236, 2002. 3. Hsu, V.P., Lukacs, S.L., Handzel, T., et al., Opening a Bacillus anthracis-containing envelope, Capitol hill,

Washington, D.C.: The Public Health Response, Emerg. Infect. Dis., 8, 1039, 2002. 4. Helgason, E., Økstad, O.A., Caugant, D.A., Johansen, H.A., Fouet, A., Mock, M., Hegna, I., and Kolstø, A.B.,

Bacillus anthracis, Bacillus cereus, and Bacillus thuringiensis-one species on the basis of genetic evidence, Appl. Environ. Microbiol., 66, 2627, 2000.

5. Centers for Disease Control and Prevention, American Society for Microbiology, and Association of Public Health Laboratories. Basic diagnostic testing protocols for level A laboratories for the presumptive identification of Bacillus anthracis. American Society for Microbiology, Washington, D.C. http://www.bt.cdc.gov/Agent/Anthrax/Anthracis20010417.pdf

6. Sacchi, C.T., Whitney, A.M., Mayer, L.W., Morey, R., Steigerwalt, A., Boras, A., Weyant, R.S., and Popovic, T., Sequencing of 16S rRNA Gene: A Rapid Tool for Identification of Bacillus anthracis, Emerg. Infect. Dis., 8, 1117, 2002.

7. Sanderson, W.T., Hein, M.J., and Taylor, L., Surface Sampling Methods for Bacillus anthracis Spore Contamination, Emerg. Infect. Dis., 8, 1145, 2002.

8. Kiratisin, P., Large-scale screening of nasal swabs for Bacillus anthracis: Descriptive summary and discussion of the National Institute of Health�s experience, J. Clin. Microbio., 40, 3012, 2002.

9. Bell, D.M., Kozarsky, P.E., and Stephens, D., Clinical issues in the prophylaxis, diagnosis, and treatment of anthrax, Emerg. Infect. Dis., 8, 222, 2002.

10. Pasechnik, V.A., Shone, C.C., and Hambleton, P., Purification of bacterial exotoxins. The case of botulinum, tetanus, anthrax, pertussis and cholera toxins. Bioseparations, 3, 267, 1992-3.

11. Jackson, P.J., Hugh-Jones, M.E., Adair, D.M., Green, G., Hill, K.K., Kuske, C.R., Grinberg, L.M., Abramova, F.A., and Keim, P., PCR analysis of tissue samples from the 1979 Sverdlovsk anthrax victims: The presence of multiple Bacillus anthracis strains in different victims, Proc. Natl. Acad. Sci., 95, 1224, 1998.

12. Glick, B.R., and Pasternak, J.J., Molecular biology: Principles and Applications of Recombinant DNA, ASM Press., Wash. D.C., 1994.

CRC Book Chapter 10 Draft 118

13. Gatto-Menking, D.L., Yu, H., Bruno, J.G., Goode, M.T., Miller, M., and Zulich, A.W., Sensitive detection of

biotoxoids and bacterial spores using an immunomagnetic electrochemiluminescence sensor, Biosens. Bioelectron., 10, 501, 1995.

14. Quinlan, J.J., and Foegeding, P.M., J. Rapid Methods Automation Microbiol., 6, 1, 1998. 15. Bell, C.A., Uhl, J. R., and Cockerill, F. R., Direct Detection of Bacillus anthracis using a Real-Time PCR Method,

ASM 101st General Meeting, Orlando, May 2001. 16. Bell, C.A., Uhl, J.R., Hadfield, T.L., David, J.C., Meyer, R.F., Smith, T.F., and Cockerill, F.R. III, Detection of

Bacillus Anthracis DNA by Light Cycler PCR, J. Clin. Microbiol., 40, 2897, 2002. 17. Thayer, A., Homeland Security: Postal Service Readies Defense - Team will install PCR-based systems to detect

biohazards in mail facilities, C&EN., 81, 7, 2003. 18. Shane, S., Post office unveils anthrax detector, Baltimore Sun, July 23, 2004. 19. Leingang, M., Post office installs anthrax detector, The Enquirer (Cincinnati)., Sept. 24, 2004. 20. Hindle, A.A., Hall, E.A.H., Dipicolinic acid assay revisited and appraised for spore detection, Analyst, 124, 1599,

1999. 21. Ascher, M.S., US Department of Health & Human Services (www.hhs.gov/ophp/presentations/Ascher.doc) 22. Janssen, F.W., Lund, A.J., and Anderson, L.E., Colorimetric assay for dipicolinic acid in bacterial spores, Science,

127, 26, 1958. 23. Murrell, W.G., Gould, G.W., and Hurst, A. Eds., The Bacterial Spore, Acad. Press., 215, 1969. 24. Ragkousi, K., Eichenberger, P., Van Ooij, C., and Setlow, P., Identification of a New Gene Essential for Germination

of Bacillus subtilis Spores with Ca2+-Dipicolinate, J. Bacteriol., 185, 2315, 2003. 25. Liu, H., Bergman, N.H., Thomason, B., Shallom, S., Hazen, A., Crossno, J., Rasko, D.A., Ravel, J., Read, T.D.,

Peterson, S.N., Yates, J. III, and Hanna, P.C., Formation and Composition of the Bacillus anthracis Endospore, J. Bacteriol., 186, 164, 2004.

26. Phillips, Z.E., and Strauch, M.A., Bacillus subtilis sporulation and stationary phase gene expression, Cell. Mol. Life Sci., 59, 392, 2002.

27. Berkeley, R.C.W., Ali, N., Classification and identification of endospore-forming bacteria, J. Appl. Bacteriol. Symp. Suppl. 76, 1S, 1994.

28. Pellegrino, P.M., Fell Jr., N.F., and Gillespie, J.B., Enhanced spore detection using dipicolinate extraction techniques, Anal. Chim. Acta., 455, 167, 2002.

29. Beverly, M.B., Basile, F., Voorhees, K.J., and Hadfield, T.L., A rapid approach for the detection of dipicolinic acid in bacterial spores using pyrolysis / mass spectrometry, Rapid Commun. Mass Spectrom., 10, 455, 1996.

30. Hathout, Y., Setlow, B., Cabrera-Martinex, R-M., Fenselau, C., and Setlow, P., Small, acid-soluble proteins as biomarkers in mass spectrometry analysis of Bacillus spores, Appl. Environ. Microbiol., 69, 1100, 2003.

31. Nudelman, R., Bronk, B.V., and Efrima, S., Fluorescence Emission Derived from Dipicolinate Acid, its Sodium, and its Calcium Salts, Appl. Spectrosc., 54, 445, 2000.

32 Rosen, D.L., Sharpless, C., and McBrown, L.B., Bacterial spore detection and determination by use of terbium dipicolinate photoluminescence, Anal. Chem., 69, 1082, 1997.

33 Woodruff, W.H., Spiro, T.G., and Gilvarg, C., Raman Spectroscopy In Vivo: Evidence on the Structure of Dipicolinate in Intact Spores of Bacillus Megaterium, Biochem. Biophys. Res. Commun., 58, 197, 1974.

34. Ghiamati, E., Manoharan, R.S., Nelson, W.H., and Sperry, J.F., UV Resonance Raman spectra of Bacillus spores, Appl. Spectrosc., 46, 357, 1992.

35. Naumann, D., FT-IR and FT-NIR Raman Spectroscopy in Biomedical Research, Fourier Transform Spectroscopy: 11th International Conference, J.A. de Haseth (ed.), American Institute of Physics Conference Proceedings, 430, 96, Woodbury, NY, 1998.

36. Kirschner, C., Maquelin, K., Pina, P., Ngo Thi, N.A., Choo-Smith, L.-P., Sockalingum, G.D., Sandt, C., Ami, D., Orsini, F., Doglia, S.M., Allouch, P., Mainfait, M., Puppels, G.J., and Naumann, D., Classification and Identification of Enterococci: A comparative Phenotypic, Genotypic and Vibrational Spectroscopic Study, J. Clin. Microbiol., 39, 1763, 2001.

37. Nelson, W.H., Dasari, R., Feld, M., and Sperry, J.F., Intensities of calcium dipicolinate and Bacillus subtilis spore Raman spectra excited with 244 nm light, Appl. Spectrosc., 58, 1408, 2004.

38. Lopez-Diez, E.C., and Goodacre, R., Characterization of microorganisms using UV resonance Raman spectroscopy and chemometrics, Anal. Chem., 76, 585, 2004.

39. Esposito, A.P., Talley, C.E., Huser, T., Hollars, C.W., Schaldach, C.M., and Lane, S.M., Analysis of single bacterial spores by micro-Raman spectroscopy, Appl. Spectrosc., 57, 868, 2003.

40. Treado, P.J., Nelson, M.P., and Vanni, S., Raman chemical imaging provides rapid and non-invasive chembiothreat detection, Photonics East, Philidelphia, Oct. 2004.

41. Zhao, X., Hilliard, L.R., Mechery, S.J., Wang, Y., Bagwe, R.P., Jin, S., and Tan, W., A rapid bioassay for single bacterial cell quantitation using bioconjugated nanoparticles, Proc. Natl. Acad. Sci.101, 15027, 2004.

42. Farquharson, S., Grigely, L., Khitrov, V., Smith, W.W., Sperry, J.F., and Fenerty, G., Detecting Bacillus cereus spores on a mail sorting system using Raman Spectroscopy, J. Raman Spectrosc., 35, 82, 2004.

43. Sanderson, W.T., Stoddard, R.R., Echt, A.S., Piacitelli, C.A., Kim, D., Horan, J., Davies, M.M., McCleery, R.E., Muller, P., Schnorr, T.M., Wardand, E.M., and Hales, T.R., Bacillus anthracis contamination and inhalational anthrax in a mail processing and distribution center, J. Appl. Microbiol. 96, 1048, 2004.

CRC Book Chapter 10 Draft 119

44. Farquharson, S., and Maksymiuk, P., Simultaneous chemical separation and surface-enhancement Raman spectral

detection using silver-doped sol-gels, Appl. Spectrosc., 57, 479, 2003. 45. Farquharson, S., Smith, W., Carangelo, R.C., and Brouillette, C., Industrial Raman: providing easy, immediate, cost

effective chemical analysis anywhere, SPIE-Int. Soc. Opt. Eng, 3859, 14, 1999. 46. Grasselli, J.G., Snavely, M.K., and Bulkin, B.J., Chemical Applications of Raman Spectroscopy, John Wiley & Sons

(NY, NY), 1981, Chapter 5. 47. Bandekar, J., Amide modes and protein conformation, Biochim. Biophys. Acta, 1120, 123, 1992. 48. Austin, J.C., Jordan, T., and Spiro, T.G., UVRR studies of proteins and related compounds, Biomolecular

Spectroscopy, Clark & Hester (eds.), John Wiley & Sons (NY, NY), 1993, p.55 49. Spores II. American Society for Microbiology, Washington, D.C. 153, 1961. 50. Carmona, P., Vibrational spectra and structure of crystalline dipicolinic acid and calcium dipicolinate trihydrate,

Spectrochim. Acta, A, 36, 705, 1980. 51. Hameka, H.F., Jensen, J.O., Jensen, J.L., Merrow, C.N., and Vlahacos, C.P., Theoretical studies of the fluorescence

of dipicolinic acid and its anion, J. Molec. Struct., (Theo) 365, 131, 1996. 52. Farquharson, S., Smith, W.W., Elliott, S., and Sperry, J.F., Rapid biological agent identification by surface-enhanced

Raman spectroscopy, SPIE-Int. Soc. Opt. Eng., 3855,110, 1999. 53. Laserna, J.J., Berthod, A., and Winefordner, J.D., Evaluation and optimization of experimental conditions for surface-

enhanced Raman detection of analytes in flow injection analysis, Microchem. J., 38, 125. 1988. 54. Dou, X., Jung, Y., Cao, Z., and Ozaki, Y., Surface-Enhanced Raman Scattering of Biological Molecules on Metal

Colloid II: Effects of Aggregation of Gold Colloid and Comparison of Effects of pH of Glycine Solutions between Gold and Silver Colloids, Appl. Spectrosc., 53, 1440, 1999.

55. Mullen, K., and Carron, K., Adsorption of chlorinated ethylenes at 1-octadecanethiol-modified silver surfaces, Anal. Chem., 66, 478, 1994.