Vibration- Rotation Spectroscopy of HCl and DCl

Purpose: To determine the fundamental vibration frequency and bond length for H 35Cl, H37Cl,
D35Cl, and D37Cl and to compare the isotope effects to theoretically predicted values.

Introduction
Vibration spectroscopy is one of the most important tools for the accurate determination of
molecular structure. Vibration spectroscopy also plays an important role in environmental
chemistry. For example, the albedo of the earth’s atmosphere is controlled by the absorption of
infra-red light by green house gasses, which include CO2, H2O, and CH4. For normal modes that
have a changing dipole moment, the vibrational spectrum of a molecule is observable using infra-
red absorption. Such modes are said to be IR active. Raman spectroscopy is used to observe
normal modes that are not IR active. The fundamental vibration frequency of a normal mode is
dependent on the isotopes of the atoms that are involved. Therefore, vibration spectroscopy is a
very useful way to study isotope effects.
In this exercise we will determine the fundamental vibration frequency and bond length for
H Cl, H37Cl, D35Cl, and D37Cl. The shift in fundamental vibration frequency and rotational
35

constant with isotopic substitution will also be predicted and compared to the experimental
values. Because of the natural abundance of 35Cl and 37Cl, the HCl spectrum will contain peaks
for both isotopes H35Cl and H37Cl. Similarly, the DCl spectrum will contain peaks for both D 35Cl
and D37Cl. The isotopic abundance of 37Cl is 24.22%, so the peaks for the 37Cl containing
molecules are one-third the height of the 35Cl containing molecules.

Theory
The vibrational energy of a diatomic molecule using the harmonic oscillator approximation is:
E = ho (  + 1/2) (1)
where o is the fundamental vibration frequency in s-1 and  is the quantum number for the
vibration, Figure 1. The rotational energy for a diatomic molecule using the rigid-rotor
approximation is:
~
EJ = Bo hc J(J+1) (2)
~
where Bo is the rotational constant expressed in cm-1, J is the quantum number for the rotation,
and each rotational level has a degeneracy of gJ = 2J+1. The crucial connection with molecular
structure is through the relationship of o with the force constant for the bond, k:
1 k
 o = (3)
2 
m1m2
where = m +m (4)
1 2

and the dependence of the rotational constant on bond length, R o:

~ h– h–
Bo = = (5)
4πc 4πR2c
o
 2
where the moment of inertia I = Ro for a diatomic molecule.

J'=3
E
J'=2
v'=1 J'=1
J'=0
J=3
J=2
J=1
v=0 J=0

0
R
Figure 1. Rotation-vibration levels in a diatomic molecule.

Consider a transition from an initial rotation-vibrationstate given by the quantum numbers , J

to a final rotation-vibration state ', J', Figure 2. This transition is symbolized by ',J',J ,
where the order is upperlower. The fundamental vibration transition is from the  to
levels. Most molecules are found in their lowest energy vibration state,  However, the
Boltzman distribution of molecules among the rotational levels produces a wide range of initial
rotational states. Therefore, the fundamental vibration of a molecule consists of a series of
closely spaced lines that correspond to the to 1 vibrational transition and wide range of
rotational transitions, ,J',J. The selection rule for rotational transitions is that J = 1. For
example, ,1,0 , ,2,1 , ,3,2 have J = +1, and are said to be part of the R-branch.
Also, ,0,1 , ,1,2 , ,2,3 have J = -1, and are part of the P-branch. The J =0
transition is forbidden and is therefore absent, which would otherwise give the Q branch.

R Q P
J' =4
J' =3
J' =2
J ' =1
E v=1 J' =0
~
~
J =5

J=4

J =3
J=2
0 v=0 J=1
J =0

bluer ~ redder
 

Figure 2. Rotation-vibration transitions for a diatomic molecule.

The energy change for the transitions are:

~ ~
∆E',J',J = E',J' – E,J = ho (' - ) + B' hc J'(J'+1) – B hc J(J+1) (6)
If the~bond~ length of the molecule doesn’t change much on going to the higher vibrational state
then B' = B and Eq. 3 simplifies to:

~
∆E',J',J = ho ± 2B hc J (7)
where J is the rotational quantum number for the lower level. This simplification gives the
spectrum as a series of equally spaced lines with the spacing between adjacent lines, Figure 3:
~ ~ ~
∆E',J',J+1 – ∆E',J',J = 2B hc (J+1) – 2B hc J = 2B hc (8)

~
2B

bluer redder
~ 
 
Figure 3. Rotation-vibration spectrum for a transition where the bond length remains constant.

However, for our case, the bond length does change significantly, so that Eq. 6 must be used. The
dependence of the rotational constant on the vibration level can be expressed as a Taylor series
expansion in ( + 1/2). Keeping only the linear term gives:
~ ~ ~
B =Bo -  e ( + 1/2) (9)
~ is the vibration-rotation interaction constant in cm-1, which is also called the Coriolis
where  e
constant. Eq. 6 can be expressed entirely in wavenumbers by dividing by hc:
~ ~ ~ ~
 = ∆E',J',J/hc = o (' - ) + B' J'(J'+1) – B J(J+1) (10)
~ ~
where  is the wavenumber of the transition in cm-1 and o is the fundamental vibration
frequency also in cm-1. Substituting Eq. 9 into Eq. 10 and taking '-gives:
~ ~ ~ ~ ~
 = o + (Bo -  e ( + 1/2)) J'(J'+1) – Bo J(J+1) (11)
For the R-branch, where J=+1 giving J'=J+1, Eq. 11 simplifies to:
~ ~ ~ ~ ~ ~ ~
R = o + (2Bo – 3e) + (2Bo – 4e) J - e J2 (12)
For the P-branch, where J=-1 giving J'=J-1, Eq. 11 simplifies to:
~ ~ ~ ~ ~
P = o - (2Bo – 2e) J - e J2 (13)
Eqs. 12 and 13 can be combined by defining a new variable m, where m=J+1 for the R-branch
and m=-J for the P-branch1:
~ ~ ~ ~ ~
m = o + (2Bo – 2e) m - e m2 (14)
 5 4 3 2 1 0 1 2 3 4 5 6 J values
6 5 4 3 2 1 -1 -2 -3 -4 -5 -6 m values

R P
bluer ~ redder

 
Figure 4. Line indexing for Eq. 14

Eq. 14 is a quadratic equation, that is, it is a second order polynomial. Least squares curve
fitting works well for polynomials, just like it does for straight lines. The polynomial coefficients
are related to the parameters in Eq.14 as follows:
y = c + b x + a x2 (15)
with the independent variable x = m and where
~ ~ ~ ~
c = o b = (2Bo – 2e) a = – e (16)

3210

3110
Wavenumbers (cm-1)

3010

2910

2810

2710

2610
-12 -8 -4 0 4 8 12
m

Figure 5. Quadratic curve fitting to the peak positions for H 35Cl using m=J+1 for the R-branch
and m= -J for the P-branch.

Isotope Effects:
H35Cl, H37Cl, D35Cl, and D37Cl each have unique spectra because the substitution of different
isotopes changes the reduced mass of the molecule, Eq. 4. H atoms have the largest change in
isotopic mass of all atoms, a factor of ~2 in substituting D (i.e. 2H) for H (i.e. 1H). The study of
isotope effects, especially upon deuteration is an important tool in mechanistic organic chemistry.
From Eq. 3, theory predicts that the ratio of the fundamental vibration frequency for two different
isotopically substituted versions of the same molecule,1 and 2, is:

k(1)
~
o(1) 
~  k(2)
(17)
o(2)

If the bond strength of the two molecules is the same, the force constants will be the same and
Eq. 17 reduces to:
~
o(1) 
 (18)
~
o(2) 

For example, in going from H35Cl to D35Cl the reduced mass increases by a factor of about 2
(precisely 1.9440) so the fundamental vibration frequency should decrease by a factor of about
2. This decrease is good to keep in mind when looking for the spectrum of DCl during your lab.
Similarly, from Eq. 5 the ratio of rotational constants for two different isotopically substituted
molecules assuming that each molecule has the same bond length is:
~
Bo(1) (2)
= (19)
~
Bo(2) (1)
~ ~
In using o and Bo we have not corrected for anharmonicity or centrifugal distortion, respectively,
which will case a small deviation between the theoretical and experimental ratios.

Procedure
HCl gas is available in anhydrous form from lecture bottles. The gas cell will be filled from the
lecture bottle source using a vacuum line. The gas cell has KBr windows to allow a wide spectral
range. Remember that KBr windows are hydroscopic. Water in any form should be carefully
excluded. Sources of water include finger prints, your breath, and solvents such as alcohols and
acetone. Finger prints are permanent. Methylene chloride is a suitable solvent for cleaning cell
windows. Instructions will be supplied in the laboratory for filling the gas cell. The pressure in
the cell should be kept as low as possible to avoid collisional line broadening that will decrease
the accuracy of the wavenumber determination. At the same time the pressure in the gas cell
should be high enough to provide sufficiently good signal-to-noise for accurate determination of
the line maxima. A typical target pressure is around 50 torr.

The Generation of DCl

DCl gas is easily generated by the reaction:

NaCl (s) + D2SO4 (l)  DCl (g) + NaDSO4 (s)

using concentrated D2SO4, which is commercially available. Concentrated D2SO4, like

concentrated sulfuric acid, is very corrosive. So be extremely careful to avoid spills. If any gets
spilled, clean up the spill immediately with lots of water. A convenient apparatus for the reaction
is shown below:

D SO4
2

Ga s Cell
Sept um

NaCl

This procedure produces a gas cell filled with a mixture of air, water vapor, and DCl. The
background gasses increase collisional line broadening. However, since the spectroscopic
constants for DCl produce bigger differences in the rotational spacing, poorer resolution is
tolerable, compared to the HCl case. KCl or NaCl windows cannot be used because of the small
amount of water vapor and D2SO4 vapor produced. Instead, CaF2 windows are used. The
absorption spectrum of CaF2 does not interfere with the DCl absorption spectrum.

Procedure: Add about 2 g of NaCl to a 25-mL round bottom flask. Use a gas cell with septa on
the inlet tubes and CaF2 cell windows. Some water and D2O is generated that would fog normal
NaCl cell windows, so we use CaF2 windows instead. Connect the gas cell as shown. Fill a
plastic syringe with about 2mL of concentrated D2SO4. When you fill the syringe, work quickly
to avoid exposure to moisture in the air, which will exchange H + for D+ ions and decrease the
purity of your D2SO4. Reseal the D2SO4 bottle quickly and store in a dessicator. Insert the
syringe needle through the septum on the 25-mL flask. Add the D 2SO4 slowly, drop-by-drop at a
rate that keeps the froth that forms from leaving the round bottom flask. Allow the gas cell to fill
until the rate of frothing decreases. Then remove the needles from the gas cell.
Make sure to wash the apparatus well, including the septa and tubing. Draw air through the
syringe needles to make sure they are dry before storing.
Cleaning the gas cell : After you are finished with the gas cell, simply remove the septa in a
fume hood and leave overnight to flush out. You don't need to wash the gas cell.

Spectra:
Acquire the spectrum of HCl using the Bruker Tensor 27 or IFS66 FT-IR. The data should be
acquired at 0.5 cm-1 resolution to insure accurate wavenumber determination. The resolution of
FT-IR instruments increases with the length of travel of the movable mirror. Using the highest
resolution setting requires the largest distance for the mirror to move. The instrument instructions
are attached.

Data Analysis:
You need to accurately record the position of each peak using the Peaks application. It is best
to work on half of the spectrum at a time. Consult the following instrument instructions.

Calculations:

Determine the wavenumbers for each line in your spectra for each molecule, H 35Cl, H37Cl,
D35Cl, and D37Cl. Determine the index, m, for each line. Use these m values to do the four
separate polynomial curve fits. The polynomial curve fitting can be done using the “Nonlinear
Least Squares Curve Fitting” Web application or an Excel spreadsheet. The Web application is
available from the course Web Home page or
http://www.colby.edu/chemistry/PChem/scripts/lsfitpl.html
Figure 5 is an example polynomial regression using Excel and the “quadest.xls” spreadsheet.
This spreadsheet is on the course Lab Web page. This spreadsheet is necessary, since most curve
fitting programs don’t provide an estimate of the fit parameter uncertainties. To do your curve
fitting using Excel, paste your peak frequencies into the “quadest.xls” spreadsheet. Nonlinear
curve fitting algorithms require an initial guess for the fit parameters. These guesses can often be
quite bad and still result in a good solution. The spreadsheet starts with the guesses a=0, b=1, and
c=0. If you don’t get reasonable fit values, you might need to try guesses that are close to the
expected values. Given below are the “quadest.xls” curve fitting results. The guesses are in the
first row. The final fit results are in the row labeled “new”.

f(x) = a*x^2 + b*x + c

guess coefficient 0 1 0
new coefficient -0.305971537 20.39985273 2885.823255
uncertainty "+-" 0.002983592 0.018157459 0.195385111

The row labeled “uncertainty” lists the standard deviation of the fit results. Propagate the
uncertainties through to your final results in terms of force constants and bond lengths. Rather
than doing a careful propagation of errors treatment, however, you can just use significant figure
rules. For example, from the printout above, we see that the fundamental vibration frequency is
2885.82±0.20 cm-1. Therefore, we know the fundamental vibration frequency to five significant
figures. However, upon converting o to the force constant, Eq. 3, o is squared. The relative
uncertainty of the square of a number is twice the relative uncertainty of the number. So the force
~
constant will have more like 4.5 significant figures. From the printout above, we see that (2B o–
~
2e) is 20.39985±0. 0.018 cm-1. Therefore, we know the rotational constant to 0.09% or four
~
significant figures. However, upon converting Bo to the equilibrium bond length, Eq. 5, a square
root is taken. The relative uncertainty of the square root of a number is half the relative
uncertainty of the number. So the bond length will be known to 0.045%, which is better than four
significant figures but not quite five.
Note that the units for the reduced mass, , are kg molecule-1. Mono-isotopic masses should be
used instead of the isotope-averaged atomic masses that are listed in the periodic table. Mono-
isotopic masses are listed in the data section of your text or in standard handbooks (Lange’s or
the CRC). The units for the bond force constant, k, in Eq. 4 are N m -1. A reasonable range for the
force constants for stable diatomic molecules is 100-2000 N m -1 (for older units 1-20 mdynes/Å).
~
The units in Eq. 5 for the bond length, Ro, are in m. Note also that Bo should be in units of m-1.
The conversion from cm-1 can be easily done using 1m = 100 cm, or
~ ~ 100 cm
Bo(m-1) = Bo(cm-1) 1m

Report:

1. Report the polynomial fit coefficients and the uncertainties for each of the four molecules.
Include the curve fit plots. Report the fundamental vibration frequencies, moments of inertia,
reduced masses, and bond lengths for H35Cl, H37Cl, D35Cl, and D37Cl. Report the force constant
for H35Cl. Make sure to use the proper number of significant figures, but do adjust for the
propagation of errors using squares of numbers or square roots as discussed above. List your
significant figures using the “underlining” convention, for example 123.456. In other words, give
a couple extra unsignificant digits, to make checking your calculations easier and in case you get
the number of significant figures incorrectly.
2. Compare your values for the bond length to the literature values for H 35Cl (only). Your value
is not corrected for centrifugal distortion; even so, a very small deviation from the literature value
is expected.
~ ~ ~ ~
3. Calculate the theoretical ratios o(1)/o(2) and Bo(1)/Bo(2) using Eqs. 18 and 19 and
compare to the experimental ratios for each pair of (D35Cl/H35Cl), (H37Cl/H35Cl), (D37Cl/D35Cl).
How well does theory and experiment compare?
4. Finally, answer the following question: how would an increase in temperature change the
spectrum of H35Cl?

Note: If you compare your value of the force constant to the literature value, you will find a
rather large error (but <5%). This is because we didn’t correct for anharmonicity. The observed
~
frequency, o, is related to the corrected fundamental vibration frequency, ~
e, through
~
o = ~ ~
e – 2 e e

with all values in wavenumbers. The anharmonicity, ~ ee, for H35Cl is 52.05 cm-1 and for D35Cl
is 27.1825 cm-1.2 To compare to literature values of the force constant, first calculate ~
e and then
use Eq. 3 to calculate the force constant.

References:
1. D. P. Shoemaker, C. W. Garland, J. W. Nibler, Experiments in Physical Chemistry, 5th. ed.,
McGraw Hill, New York, N. Y., 1989. Experiment 38.

2. K.P. Huber and G. Herzberg, Molecular Spectra and Molecular Structure, v. 4, Constants of
diatomic molecules, Van Nostrand, Toronto, Canada, 1945.

Bruker FT-IR Instructions: Opus Operation Software

To take a spectrum, click on the Single-spectrum Advanced Measurement Icon:

In the Measurement dialog, click on the Advanced tab. Enter the file name for your spectrum.
Choose the Resolution and Sample Scan number. Higher resolution (smaller value) and more
scans require more time. For low pressure gases choose 0.5 cm-1, for ambient pressure gases 1.0
cm-1 is sufficient, for liquids and solids 4.0 cm-1 is sufficient. Four scans is often sufficient for
transmission studies. The normal wavenumber range is 4000 to 400 cm -1. For the HCl and DCl
experiment, saving the data from 4000 to 1000 cm-1 creates a smaller data file.

FT-IR spectra are best displayed as absorbance versus wavenumber. However, for use with gas
cells, Transmittance works best. The spectrum can always be converted from transmittance to
absorbance by the software if necessary. Check to make sure the Background, Single Channel,
and Transmittance Data blocks are to be saved (or Absorbance if a liquid or solid). These
choices allow flexibility in data manipulation and allow you to see the background absorption.
Click on the Basic tab at the top of the Measurement dialog window. Make sure no sample is in
the IR-beam. Click on the Background button to take the background scan. The progress of the
scan is shown in green at the bottom of the spectrum display window. When the background is
complete, place the sample in the IR-beam and click the Sample button. After the scan is
complete, click the Close button.
Your spectrum should be displayed. You can control the axis ranges to focus on regions of the
spectrum.

To set the Y-axis to full scale, click on the Full Scale Y icon:

To zoom in to a region of the spectrum using the mouse, click on the Zoom In icon:
Left click on the spectrum on the upper left of the zoom region. Double-click the left mouse
button on the lower right of the zoom region to expand the X- and Y-scales. You can also access
the spectrum scaling options by right clicking on the white background. Zoom in is available
under the Zoom option and returning the spectrum to full scale X and Y is the Show Everything
(XY) option under Scale All Spectra. Left click on the white background to exit the Zoom In
mode.

The full spectrum is always displayed in the spectrum bar at the bottom of the screen, as shown
below. Position the mouse at the left side of the white region in the bar. The mouse cursor
changes to a  cursor, which can then be used to drag the left boundary for the displayed region.
The right side of the white region controls the low wavenumber boundary in a similar way.
Dragging the mouse in the middle of the white region translates the displayed region. Position the
mouse along the bottom of the white region. The mouse cursor changes to a . Dragging the
mouse then expands and contracts the Y-axis scale.

To read data points on the spectrum, click on the Cursor (follow data) icon:
Record the wavenumber range that contains the peaks of interest. The follow data cursor is also
accessible by right clicking on the white spectrum background.

Peak Picking:

First do Background subtraction by clicking on the Baseline Correct icon:

In the Baseline Correction dialog window, make sure that your spectrum file is highlighted (even
if it the only file displayed). If your spectrum isn’t listed in the Files to correct list, you must drag
the file from the Opus Browser using the mouse. Select the TR file (or ABS if you selected that
data acquisition mode).

Then click on the Correct button.

Next click on the Peak Picking icon:

In the peak picking dialog window, make sure that your spectrum file is highlighted (even if it
the only file displayed, or drag the file from the Opus Browser).

Click on the Frequency Range tab. Enter he range of wavenumbers for peak picking. To pick
peaks over the range of the data that is currently displayed, click on Get Display Limits. (If you
don’t remember the wavenumber range you can click the Interactive button.)
Click on the Select Files tab at the top of the Peak Picking dialog window. From the Select Files
tab, click on Interactive button to set the Sensitivity. A new window is displayed.
The peaks that have been selected for peak
labeling are indicated with a short vertical
dash. Adjust the sensitivity to select the range
of peaks that you wish to label. If some noise
peaks are selected, they may be removed
during the next stage. Click the Store button to
save the picked peaks and return to spectral
display. If peaks are labeled that you don’t
wish to be labeled, highlight the label with a
left click of the mouse. A right click of the
mouse brings up a context sensitive menu that
allows you to delete the peak label.

In the spectrum, the peaks may overlap the labels, making them difficult to read. The full
spectrum bar at the bottom of the screen can be used to adjust the Y-axis scale. Move the mouse
to the bottom of the white area in the full spectrum bar. The cursor changes shape to a .
Dragging with the left mouse button then expands and contracts the Y-axis scale.

 Drag to adjust the Y-axis

To print the labeled spectrum, click on the printer icon.
To list the peaks in a table, double click on the Peaks button in the Opus browser in the upper
left of the main window. The peaks are then listed. You can select and copy the listing into Word
Pad, which can then be saved as a text file that is readable using Excel. To return to the spectral
display, click on the tabs at the bottom of the Table display.
To remove the Peaks listing altogether, click right on the Report-Display entry in the Opus
browser and select Close Window.

If all the required peaks are not listed by Peak Picking, use the Add Annotation tool as follows.
First, position the cursor on the spectrum curve at the wavenumber that you want to label. Click
right for a context sensitive menu and choose Add Annotation.

The label may be repositioned by dragging with the left mouse button.

Getting Ready for the Next Spectrum:

More than one spectrum can be active at a time. The displayed spectrum is controlled by
clicking on the corresponding TR or ABS buttons in the Opus browser. However, it is often less
confusing to remove the old spectrum before taking a new spectrum. In addition, if you are
finished for the day, you should remove your spectra and peak tables to prepare for the next
student. To remove your spectrum, move the mouse over your spectrum file name in the Opus
Browser (not the “Display – “ line that lists the current Opus window system set-up file). Right
click and select the Unload File Option.