The R am an system

A s described above, the R am an effect results from the excitation o f a sam ple w ith light. The intensity o f the scattering resulting from the R am an effect is six orders o f m agnitude w eaker than the intensity o f the exciting light. T he intensity is also dependant upon the w avelength o f the exciting light, w ith a relation I ~ w here I is the R am an scattering intensity, Iq the intensity o f the exciting light and X the w avelength o f the exciting light. H ow ever, the R am an shift is independent o f the exciting w avelength. In order to perform a R am an spectroscopy experim ent, one requires several elem ents:

intense m onochrom atic exciting light (laser)

a w ay to filter the laser line after excitation o f the sam ple (notch filter or p re­ m onochrom ator)

a high resolution spectrom eter to disperse the w avelengths a very sensitive detector to collect the w eak signal

ob jectiv e. Video im aging CC D cam era T ransm itting grating N otch filter 2 N otch filter C C D D etecto r R ectangular apeiture A r L aser light path HeNe Laser light path

I

Figure 2-13 Custom built Raman system at UCL, on the left system using the H eNe laser an d

on the right system using the A r laser. The orange line indicates the path o f the H eNe laser,

the green line indicates the path o f the A C laser an d the red line indicates the path o f the scattering signal fro m the sample.

I built the m icro -R am an spectroscopy system (figure 2-13 and figure 2-14) u sin g a m ulti line air cooled Ar* laser (150M L aser physics®) and a high pow er H eN e laser (M elles G riot). T he Ar* laser provides 50 m W o f laser light at 514.5 nm and 488 nm w h ich are the tw o strongest lines. T he H eN e laser p rovides 35 m W at 633 nm. H ow ever, one m ust use a p lasm a line filter in order to cut the strong and n um erous plasm a lines from that laser. The p lasm a line filter cut ab o u t 40 to 50 % o f the laser pow er leaving only less than 20 m W o f av ailab le pow er. T he p o w er lever o f the H eN e laser is fixed, so w e vary the laser p o w er sent to the sam ple by p lacin g neutral density filters at the laser output.

W e select the laser line o f the Ar* using a Kaiser® tran sm ittin g grating (h o lo g rap h ic grating m ounted betw een tw o prism s). T he grating tran sm its the s p olarised (p o larisatio n parallel to the gratin g ) laser line at a 9 0 ° angle figure 2-15. T he grating d isp erses the w avelengths, and m akes th eir selection possible.

Argon ion laser H eN e Laser

Scattering from the sam ple _{H eN e Laser} D A C

• N otch : filter O bjective mirror rem ovable mirror _lens Spectrom eter Argon ion Laser concave mirroi Video im aging CCD camera transm ission grating Gratings rectangular aperture CCD D etector

Figure 2-14 Schem atic representation o f the custom built Raman spectroscopy system at UCL. We sh ifted the lines representing the various light paths in order to clarify the diagram.

Figure 2-15 Schem atic representation o f the Kaiser® transm ission grating. The grating sends

each wavelength at a different angle.

B roadband en h anced A1 coated m irrors (?i/10) steer the laser beam from the laser o u tp u t to the notch filter. T h o se m irrors have an average reflectance o f 88 % b etw een 400 and 700 nm . A t this stage, there is alw ays sufficient laser pow er so that it is accep tab le to have som e losses, thus red u cin g the cost o f the optical elem ents. Finally, a Kaiser® S u p erN o tch -P lu s™ holo g rap h ic filter at an in cidence angle o f less than 5 ° sends the incom ing beam into the o bjective and to the sam ple. T he notch filter acts as a m irror for the selected laser line (514.5 nm or 632.8 nm ) and

transm its ail other w avelengths. This notch filter is different from the transm ission grating as it reflects rather than transm its the specific w avelength and is not part o f a prism . A n infinitely corrected M itutoyo super long w orking distance objective SL50 w ith a 50x m agnification focuses the beam onto the sample. The w orking distance o f this objective is 21 mm. T he angle o f the notch filter is the m ost critical part o f the system as it determ ines w hether or not the laser line is steered into the objective along its optical axis.

W e collect the scattered light signal in a backscattered geom etry. Thus, the objective collecting the R am an signal is the same as that focusing the beam onto the sample. W e then filter the reflected beam and .the Rayleigh light using the Kaiser® SuperN otch-Plus™ filter previously used as a m irror. A t this stage, it is im portant to collect as m uch o f the w eak scattered signal as possible. Therefore, w e use highly efficient m irrors to steer to beam from the notch filter to the spectrom eter and detection system. The m irrors are broadband dielectric m irrors w ith a 99 % reflectivity betw een 488 and 694 nm. W e first steer the beam dow nw ard in order to reach the height o f the spectrom eter entrance slits. It is im portant to note that none o f the optics used in the system polarises the Ram an scattering from the sample. Therefore, the polarisation state o f the R am an signal is not changed.

In order to discrim inate the signal from the sam ple from Ram an scattering or lum inescence from the diam ond, w indow s o f our high pressure cell, we use a confocal spatial filtering system . A 100 mm focal length achrom atic doublet lens focuses the beam onto a rectangular aperture. The continuously variable rectangular aperture spatially selects the desired beam diam eter o f the signal. Then a second 100 mm focal length achrom atic doublet lens collim ates the beam. Typically, the aperture is set to select a 5 x 5 |xm area on the sam ple.

Then, the beam passes through a second K aiser SuperN otch™ filter in order to discrim inate further against the incident laser line (usually 514.5 nm). Experience show ed that “ leakage” can occur along the light path or from stray diffuse or specular reflections w ith the laboratory, w hich could result in sw am ping the w eak R am an signal incident on the detector. It is particularly necessary to use o f the second notch filter for studies o f w eakly scattering m etallic sam ples such as the transition m etal nitrides studied here. In the case o f such sam ples, there is little penetration o f the incident beam into the sam ple. Therefore, the sam ple reflects a large portion o f the elastically scattered light along the m ain beam path. Thus, there is m uch m ore laser light to discrim inate from the w eak Ram an signal. W e do not use any secondary notch filter w ith the H eN e laser excitation, as the first notch filter appears to rem ove enough o f the laser light for useful spectroscopy.

Finally, a 31.8 m m focal length achrom atic doublet lens focuses the beam onto the entrance slit o f the spectrom eter. W e carefully selected the focal length o f that lens in order to provide

m axim um coverage o f the diffraction grating inside the spectro m eter, thus p ro viding m axim um resolution. T he focal length o f the spectrom eter is / / = 500 m m . T he size o f the g ratin g s is a square o f 68 m m x 68 mm. T he radius o f the beam is c/ ~ 2 m m . W ith these data, on e can calculate the optim um focal length for the lens [f]) using sim ple geom etry.

/ = X /', = — X 5 0 0 = 2 9 .4 mm

& 5x68

jW

A slightly longer focal length ensures the p reservation o f the totality o f the signal since it reduces the size o f the beam at the spectrom eter grating. T h erefo re, w e selected a 31.8 m m focal length lens. 0 .9 - 0 .7 - 0.6- 0.2- 0.0 400 450 500 550 600 650 700

Wavelength (nm)

Figure 2-16 A bsorption curve o j the Kaiser® SuperNotch^^' filte rs used in the Ram an system.

In blue f o r the fir s t filter, in red f o r the second filte r and in black f o r the com bination o f both

fd ters.

F igure 2-16 show s the absorption curve o f a typical notch filter. Such notch filters p ro v id es an extrem ely high th ro u g h p u t com bined w ith rejection o f the incident laser light w hen co m p ared to a traditional triple spectrom eter system . H ow ever, the quality o f the notch filter d eterm in es the low relative w av en u m b er (R e m '') detection lim it o f the system . In our ex p erien ce, it is very d ifficu lt to obtain interpretable R am an signals at less than 100 Rem ' w ith a notch filter system . T hat point is the m ajor inconvenience o f the notch filter system , co m p ared to a d o u b le or triple g ratin g sp ectro m eter system . H ow ever, it allow s for the first tim e the co llectio n , in situ in the diam o n d anvil cell, o f spectra o f very w eakly scattering solids such as m etals. In o u r system ,

using tw o notch filters, it is possible to detect R am an shifts as close as 45 c m '’ o f the laser line, w ith sam ples th at are o f good optical quality; i.e. that do not reflect too m uch o f the laser line. In that case, w e fix the position o f one o f the notch filter and slightly rotate the o th er one o f f its optim um angle in o rd er to adjust the low -frequency cut-off, figure 2-17.

0.9 - w 0.7 0) o c a 0.6 c/5 I 0-4 ^ 0.3 0.2 0.0 -300 -250 -200 -150 -100 -50 0 50 100 150 200 250 300 Raman Shift ( c m ’ )

Figure 2-17 Schem atic representation o f the notch filters optimisation. By rotating the notch

filter, the edge o f m inim um in the transm ission curve m oves either tow ards higher or low er

waven umbers.

T he sp ectro m eter used is an A cton R esearch S p ectraP ro ® 500i sp ectro m eter w ith a focal distance o f 500 m m , an aperture ratio X/6.5 w ith a im aging C z e rn y -T u rn er arran g em en t using aspheric m irrors. The spectrom eter includes a choice o f three in terchangeable diffraction gratin g s (600, 1200, 2400 g rooves/cm ) m ounted on a ro tateab le tuiTet. T he chosen range o f gratin g s allow s interplay betw een the desired spectral resolution and the spectral range as show n in table 2-2. T he detecto r is a b ack-illum inated silicon C C D d etecto r Princeton Instrum ents Spec 1 0 :1 0 0 8 o f 1340 x 100 pixels w ith a pixel size o f 20 x 20 pm .

600 g .c m '’ 1200 g .c m '’ 2400 g.cm '

S pectral R ange (c m '') 2500 1400 800

R esolution (cm ’) ~ 3 ^ ~ 1.0 - 0 . 5

Table 2-2 S pectral range an d resolution o f the spectrom eter f o r each grating centred a t 520

2.3.

High pressure synthesis techniques