What is Raman spectroscopy?

     When molecules interact with incident radiation, most of the radiation is scattered elastically (Rayleigh scattering). However, certain molecules channel some of the incident radiation into vibrational modes, resulting in the scattering of a small part of the radiation (on the order of 10-7) at different, longer wavelengths (this is called Stokes Raman scattering). Since Raman spectra are unique, they provide a sort of chemical fingerprint for unknown compounds.

1.    Abstract

Previously a low-resolution, classical spectrograph with a resolution of about 700 and a spectral range of about 200 nm was constructed.

For vibrational spectroscopy, a Raman system was built for the above spectrograph and also for use with a higher-resolution Littrow spectrograph. It used a green 532 nm laser to stimulate Stokes scattering in samples. The laser beam was reflected off a dichroic mirror at an angle of 90°. A microscope objective then focused the laser beam onto the sample and collimated the returning light. Light of the laser wavelength was filtered out by an interference filter which reduced the laser signal by

approximately 10^5. A 50 mm f.l. lens then focused the waveshifted light on the entrance slit of the spectrograph. Measured wavenumbers correlated well to those in published spectra. Difficulties included false spectral lines created by the multimode laser and flourescence in many samples.

 

2.    Objectives

1.      To construct a Raman spectrograph system using a Raman head consisting of an excitation laser, a set of filters, appropriate lenses, and a sample holder, to be attached to a spectrograph.

2.      To test the Raman system with various Raman standards and make modifications as required.

3.    Theory

3.1.                    The Raman effect

     Raman spectroscopy, like infrared spectroscopy, is a form of vibrational spectroscopy. In a Raman spectrograph system, a high-intensity beam of light is directed to a sample. The vast majority of the light is scattered elastically (Rayleigh scattering) and the reflected light has the same wavelength as the incident light. A very small portion of the scattered radiation, however (on the order of 10^-7) is shifted to a different wavelength. The scattered photons are known as Raman scattering. A few are shifted to higher energy wavelengths (anti-Stokes radiation) but most are shifted to lower frequencies (Stokes radiation) (21). Stokes scattering results when a molecule is in its ground state when it interacts with the beam of light; some of the energy from the colliding photon is channeled into the vibrational mode of the molecule. This causes the light to be absorbed and then re-emitted at a lower frequency. Anti-Stokes scattering occurs when a molecule is in a vibrationally excited state when it interacts with the radiation; the interaction can cause it to drop to the ground state and lose some of its vibrational energy to the re-emitted (higher frequency) light. (3) Stokes scattering is much more common than anti-Stokes, because at normal temperatures, electrons are most likely to be in their lowest energy state (a result of the Boltzmann energy distribution) (21). A spectrum of anti-Stokes scattering is identical in pattern, but much less intense, than a spectrum of Stokes scatterings; for this reason, only Stokes scattering is typically used in Raman spectroscopy (2).

 

 

Fig. 1 - Jablonski energy diagram (17). The Raman effect with Stokes scattering occurs when a photon striking a molecule excites an electron into a higher “virtual” energy level, and the electron decays back to a lower level, emitting a scattered photon (this is the quantum mechanical model for the Raman effect). With Anti-Stokes scattering just the opposite happens. (12) In Rayleigh scattering, the molecule absorbs a photon and scatters it at the same energy.

 

Fig. 2 - Spectrum of CCl4 demonstrating the symmetry of

Stokes and Anti-Stokes bands (12).

     Not all vibrations will be observable with Raman spectroscopy, depending on the symmetry of the molecule, but there is usually enough information present to enable a precise characterization of molecular structure (17). Unlike infrared spectroscopy, which results from a change in the dipole moment, Raman scattering results from changes in the polarizability of a molecule; thus Raman complements infrared spectroscopy. The energy of a vibrational mode depends on molecular structure and environment, including atomic mass, bond order, molecular substituents, moelcular geometry and hydrogen bonding, all of which effect the vibrational force constant that in turn dictates the vibrational energy (19). Classically, a molecule can be thought of masses attached by strings governed by Hooke’s law, where k is the vibrational force constant. Especially with infrared spectral lines, which result from changes in the dipole moment that occurs as a result of molecular vibration, the frequency of a vibrational mode may be understood in light of the fact that an incident oscillating electric field is producing alternating electric forces on a molecule, causing the dipole spacing to alternately increase and decrease (8). The force constant in a diatomic molecule can actually be classically derived, if the masses of the atoms in question are known, as a function of the fundamental frequency of vibration: I have done this below using differential equations. For more complicated systems of atoms, knowing the relationship between the observed wavenumbers in a spectrum and the force constants between the atoms can be much more difficult and require computer modeling, even for very practised chemical spectroscopists (15).

Fig. 3 - Diagram of a triatomic molecule (H2O) as it is hit by a wave of incident electromagnetic radiation (17).

3.2.                    Derivation of the relationship between frequency and Force Constant for a Diatomic Molecule


Not IR active; raman active               IR active; not raman active

 

Fig. 4 - Diagrams contrasting symmetric (left) and anti-symmetric (right) stretches for CO2 (2).

 

     Again, for a transition to be Raman active, there must be a change in the polarizability of the molecule. But what is polarizability? Polarizability is the change of the dipole moment over distance (2), or the the ease of which an electron cloud can be distorted by an external electric field. There must be a change in polarizability during the vibration for that vibration to inelastically scatter radiation (19). In considering polarizability, it is useful to consider whether there is a change in the volume of the electron cloud during the vibration. For instance, the CO2 molecule (O=C=O) has a “symmetric stretch,” a vibrational mode in which both oxygen atoms oscillate in opposite directions away from the carbon atom; since the volume of the electron cloud around the molecule effectively grows smaller and larger, the vibration is Raman active. On the other hand, there is another vibration, an “antisymmetric stretch,” in which the oxygen atoms move in the same direction; in this vibration, the increase in volume on one side is offset by a decrease in volume on the other, so that the volume of the electron cloud doesn’t effectively change and the vibration isn’t raman active. (3) Chemists use polarizibility ellipsoids to describe polarizability; they compare the ellipsoid at equilibrium bond length to the ellipsoid for the extended and compressed symmetric motions. (2)

 

Fig. 5 -  Diagram of electron cloud around Carbon Tetrachloride (12).

 

1.    Instrumentation

     Raman systems are generally composed of several main parts, as follows.

1. The excitation source (laser).

     Lasers make modern Raman feasible because they give a coherent beam of monochromatic light and can have very high power. They must be of sufficient intensity to produce the desired and detectable amount of Raman scatter, and they should be free of extraneous bands. They should thus exhibit good wavelength stability and low background emission.

     The green 532nm laser I used in my project was a diode-pumped solid state laser. It uses an infrared laser to pump a frequency-doubling crystal.

 

 

Fig. 5 - Internal construction of 532nm laser.

http://www.repairfaq.org/sam/dpmdpss1.gif

 

The frequency of a Raman shift is independent of the laser wavelength used for excitation. This is an important concept in Raman. In Raman, unlike electronic spectroscopy, wavelengths are not normally measured in Angstroms or nanometers. Since the frequency shift, not the actual frequency, is what is important, Raman spectra are generally plots of intensity versus wavenumber:

where c is the speed of light in cm/s and v is the frequency in s-1 (12). Since it is independent of the laser wavelength, the wavenumber will be the same for a 532 nm or a 785 nm laser. The wavelength difference for a particular band, however, is not necessarily the same. For instance, a wavenumber of 1000 cm-1 for a 532 nm laser would result in a Stokes band ~30 nm away from the laser line. But the same wavenumber for a1064 nm laser would result in a Stokes band ~130 nm away from the laser line.

Fig. 5 (below). A table of wavelengths with associated wavenumbers for a laser of 532 nm is shown below. The green and orange numbers represent wavelengths for a laser and for a mercury lamp which was used for calibration.

 

Table 1. Frequency vs. Wavelength Table

Wavelength (nm)

n -1 (cm-1)

Δ-1

 

Wavelength (nm)

n-1 (cm-1)

Δ-1

532

18796.99248

0.00

 

571

17513.13485

1,283.86

534

18726.59176

70.40

 

572

17482.51748

1,314.47

535

18691.58879

105.40

 

574

17421.6028

1,375.39

536

18656.71642

140.28

 

575

17391.3043

1,405.69

537

18621.97393

175.02

 

576

17361.1111

1,435.88

538

18587.36059

209.63

 

576.97

17331.9237

1,465.07

539

18552.8757

244.12

 

578

17301.0381

1,495.95

540

18518.51852

278.47

 

579.07

17269.0694

1,527.92

541

18484.28835

312.70

 

580

17241.3793

1,555.61

542

18450.1845

346.81

 

581

17211.70

1,585.29

543

18416.20626

380.79

 

582

17182.13

1,614.86

544

18382.35294

414.64

 

583

17152.66

1,644.33

545

18348.62385

448.37

 

584

17123.29

1,673.70

546.0731

18312.56658

484.43

 

585

17094.02

1,702.98

547

18281.53565

515.46

 

586

17064.85

1,732.15

548

18248.17518

548.82

 

587

17035.78

1,761.22

549

18214.93625

582.06

 

588

17006.80

1,790.19

550

18181.81818

615.17

 

589

16977.93

1,819.06

551

18148.82033

648.17

 

590

16949.15

1,847.84

552

18115.94203

681.05

 

591

16920.47

1,876.52

553

18083.18264

713.81

 

592

16891.89

1,905.10

554

18050.54152

746.45

 

593

16863.41

1,933.59

555

18018.01802

778.97

 

594

16835.02

1,961.98

556

17985.61151

811.38

 

595

16806.72

1,990.27

557

17953.32136

843.67

 

596

16778.52

2,018.47

558

17921.14695

875.85

 

597

16750.42

2,046.57

559

17889.08766

907.90

 

598

16722.41

2,074.58

560

17857.14286

939.85

 

599

16694.49

2,102.50

561

17825.31194

971.68

 

600

16666.67

2,130.33

562

17793.59431

1,003.40

 

601

16638.94

2,158.06

563

17761.98934

1,035.00

 

602

16611.30

2,185.70

564

17730.49645

1,066.50

 

603

16583.75

2,213.24

565

17699.11504

1,097.88

 

604

16556.29

2,240.70

566

17667.84452

1,129.15

 

605

16528.93

2,268.07

567

17636.6843

1,160.31

 

606

16501.65

2,295.34

568

17605.6338

1,191.36

 

607

16474.46

2,322.53

569

17574.69244

1,222.30

 

608

16447.37

2,349.62

570

17543.85965

1,253.13

 

609

16420.36

2,376.63

 

 

     Intensity of the Raman signal is inversely proportional to the 4th power of the laser wavelength (3):

It would seem, then, that laser radiation of a high frequency would be ideal. However, since samples tend to flouresce at higher-frequency wavelengths, and since fluorescence is a much more efficient process than Raman, this is not necessarily true.  Thus, infrared lasers have become increasingly popular in Raman, despite their low frequencies. On the other hand, it has also been demonstrated that at wavelengths under 260 nm, there is essentially no fluorescence interference (3); so ultraviolet lasers have the benefits of stimulating strong Raman scattering and not causing samples to fluoresce. I chose a visible-band laser because I had one for pointing out stars. The optics involved with infrared and ultraviolet lasers and their associated filtering mechanisms seemed out of my price range.  In retrospect, the inexpensive 532nm lasers (I tried several) oscillated at 3 frequencies, making Raman analysis difficult.

2. System to irradiate sample and collect scattered light (three types of filters).

     Light from the laser is generally passed through an excitation filter in Raman systems. The excitation filter purifies the laser beam by permitting light only of a very small selected wavelength range to pass through, eliminating background emission (1). Excitation filters may be Short-Wave-Pass (only extraneous light of longer Stokes wavelengths than the laser is blocked) or Laser-Line (extraneous light of wavelengths shorter and longer than the laser is blocked) (13). In my Raman system, the laser-line filter’s cutoff was too far from the 532nm laser line to clean up the harmonics.

     After passing through an excitation filter, the laser beam is usually focused on the sample in order to increase the intensity of the electric field on the sample surface (the magnetic component of the electromagnetic radiation is generally not relevant to Raman) (12). Raman scattered light is then collected and passed through a barrier filter. The barrier filter prevents Rayleigh-scattered light of the same wavelength as the laser form passing on to the slit of the spectrograph (the Rayleigh scattering is so much more intense than the Raman that it would saturate the CCD chip).

     Barrier filters may be Long-Wave-Pass (only Stokes radiation may pass through) or Notch (both Stokes and anti-Stokes radiation may pass through) (13).

     A third type of filter, known as a dichroic mirror, reflects light of the wavelength of the laser and allows Raman scattering to pass through it. It was used in my Raman system between the excitation and barrier filters at an angle of 45 degrees, to direct the laser beam to the sample.

     There are a variety of optical configurations for the Raman probe head. The filters can be arranged so that the scattered Raman is collected at an angle of 90 degrees or 180 degrees from the incident laser beam; (12) they can also be arranged so that the scattered light is collected at an oblique angle, as on the previous page.

 

 

Fig. 6 - Graph of notch excitation filter (13).

 

Fig. 7 - Graph of barrier long-wave-pass filter (13).

                

 

 

3. Lens to collect scattered light.

A lens collects the Raman scattered light coming through the dichroic mirror and focuses it to the slit of a spectrograph. Basically the job of the spectrograph is to transform the Raman light into a spectrum, which, upon being collected by the CCD camera can be analyzed by a spectral processing program. In older systems before CCD camera were available a grating was mechanically scanned and a photomultiplier tube was used as a single pixel detector. A device with an input slit, lenses or mirrors, a grating,  and an output slit is called a monochromator.

 

Research Raman Spectroscopy

 

Fig. 8 - Jobin Yvon T64000 Raman system using triple monochromator, similar to that used at OU by Dr. R. Freck (15). Such systems cost over $100,000. The CCD Camera, which is what is sticking up in the photo, is liquid nitrogen cooled. (17)

Littrow spectrograph design and CCD camera considerations

     Since I wanted my Raman spectrograph to be medium- to high-resolution, I chose to upgrade the low-resolution spectrograph I built in my project two-years ago to a higher resolution spectrograph with an entirely different design, called a Littrow spectrograph. Unlike the classical spectrograph, the littrow spectrograph only uses one lens, which decreases its size and cost and increases its stability (theoretically). This lens collimates light entering the spectrograph from a slit and sends it to the grating; after the light is dispersed, the lens focuses the spectrum onto a camera chip.

      Spectrographs output graphs of intensity versus wavelength, meaning that

they show the different wavelengths of a particular light source. The spectrograph generally incorporates a spectral dispersion device (a diffraction grating), the collimation lens, and the collection device (for me, a CCD camera).

My final drawn design for my Raman system. You can click it to see an enlarged version.

 

 

A photo of my Raman system. In this photo the Raman head is not attached to the Littrow spectrograph (diagrammed above); it is attached to the Classical spectrograph. You can click on the photo to see it enlarged.

6.    Data Analysis

 

     Raman spectra of various substances, including standard Raman chemicals like acetone and toluene and also other various materials, including ethyl alcohol, glass, diamond, acetaminophen (Tylenol), and Teflon, were successfully observed. With many materials, fluorescence was a problem, probably due to the relatively high frequency (as compared to infrared) of the laser. Spectra were calibrated with a mercury lamp, and the observed wavenumbers for various lines in the spectra correlated very well with published wavenumbers. It did prove to be a challenge to maximize the intensity of the Raman signal, especially with the higher resolution spectrograph and the very thin slit required of the higher resolution setup; but when exposure lengths were adjusted, spectra could still be successfully obtained of molecules with pronounced Raman resonances.

     After the Raman spectrograph system was completed, the main obstacle to good spectra proved to be an unexpected one: all of the Raman lines, as it became apparent as I narrowed my slit and got higher resolution spectra, did not have sharp, well-defined profiles as described in the literature but instead were complex blends of three or four peaks. At first I attributed this to some flaw of the Littrow spectrograph, which was actually producing lines with a slight ghost on the red side (perhaps because of the higher-groves per mm diffraction grating). I noticed, however, that the laser line in particular was characterized by the complex profile, and the mercury lamp calibration lines were not. When I took spectra of the laser line by itself, without any filters, I saw that it was not outputting light at one specific frequency, but rather at several, and that—since Raman frequencies are independent of the laser frequency—the Raman lines were mimicking the same spectral profile. To produce better quality Raman spectra, I would need a more expensive single-mode laser.

     Since a laser line was visible in most of the spectra obtained, I wondered what the attenuation of the laser line by the barrier filter was. It was found the ratio of the laser line intensity to the intensity of the blocked laser was 1.25 * 105, with a 3.5% error (over two trials). When the intensity of the Raman signal for toluene, an enthusiastic Raman scatterer, was compared to the computed, normalized full-intensity laser beam, it was found that the ratio between the intensities of the Raman Stokes scattering and the laser line was 3.7 * 10-6. This was comparable to the 10-6 or 10-7 value mentioned in the literature.

7.    Conclusions

 

     In the end, a Raman spectrograph system was successfully constructed. The process of building the Raman head was difficult, especially the optical alignment, and only through trial and error were Raman spectra finally achieved. Optical alignment of the Raman system was difficult due to the faintness of the Raman signal (it is invisible to the naked eye, so it is difficult to get the signal aligned on the slit of the spectrograph). The idea that Raman spectrographs—high or low resolution ones—might be someday constructed for very wide-ranging industrial applications and for a relatively small cost is very exciting.

     In the future I would like to create a program to quicken data analysis of spectra produced by the Raman system, and even write a program to use least-squares analysis to automatically determine the composition of compounds using a library of Raman spectra of common compounds. I would also like to examine the relationship between compound concentration and intensity of the Raman signal, which theoretically are directly proportional. This will require a carefully alignable sample holder, with a constant thickness, since the Raman signal varies depending on how much of the sample the laser light passes through. I would like to experiment with the sample holder and the laser alignment to see if the Raman signal can be maximized.

          Raman spectroscopy has an amazingly diverse array of applications in society, including analysis of synthetic polymers like rubber, identification of legal and illegal drugs (forensics) and identification of explosives, determination of unsaturation in food oils and fats, determination of defects in semiconductors; in-situ measurements of rocks and minerals during planetary missions, real-time detection / diagnosis of cancer, and analysis of artwork. These applications pose a seemingly unlimited number of potential project ideas, and emphasize the significance that amateur Raman spectroscopy could have on the entire field.

8.    Acknowledgements

     I would like to thank Chroma Corporation for courteously donating me a set of Raman filters, including a dichroic mirror, barrier filter, and excitation filter, for use in my research; without these filters, my project probably would have been virtually impossible. I would also like to thank Alan Holmes and the Santa Barbara Instrument Corporation (SBIG) for letting me borrow an ST-10 CCD camera for future use with my project. I would like to thank the IAPPP (International Amateur Photoelectric Photometry) and the AAS (American Astronomical Society) for granting me and my previous partner Sarah Howell the Richard D. Lines award last year, which helped to finance my project this year. I give thanks also Dr. Roger Freck of the chemistry department at the University of Oklahoma, for taking the time to graciously explain the fundamentals of the Raman effect to me, especially the classical and quantum explanations of how an incoming laser beam affects molecules, and letting me see his two Raman spectrographs. I would also like to thank Mr. Jeffrey Baughman for helping me with all of the advice he offered me concerning my project and his immense help with regards to the forms and other technicalities required. I would most like to thank my parents for their support, specifically my dad, who taught me how to use AutoCAD to design my spectrograph and instructed me in the difficult process of machining the parts for my spectrograph.

9.    Bibliography

 

1.      Abramowitz, M., & Davidson, M. (2004). Molecular Expressions Microscopy Primer: Specialized Microscopy Techniques: Flourescence Filters. Retrieved November 10, 2005, from http://microscope.fsu.edu/primer/techniques/fluorescence/filters.html.

2.      Anderson, Larry (2000). Raman Spectroscopy. Retrieved November 23, 2005, from http://carbon.cudenver.edu/public/chemistry/classes/chem4538/raman.htm.

3.      Beaucage, G. Infrared Spectroscopy and Raman Scattering (1998). Retrieved January 2, 2006 from http://www.eng.uc.edu/~gbeaucag/Classes/Analysis/Chapter5.html.

4.      Bowdoin. Orientation and Intermolecular Forces (n.d.). Retrieved December 23, 2006, from http://academic.bowdoin.edu/courses/f01/chem225/chem225b/Polarity/Orientation.shtml.

5.      Brief description of raman scattering. http//www.omegafilters.com/index.php?page = omegatext/prod_raman_scatter.