Time-course and Kinetics Measurement

NOTE: For a similar time-course monitoring experiment on the Bruker instrument, refer to this page. The following is for experiments on a Varian or Agilent NMR spectrometer under VnmrJ.

The simplest method to monitor a reaction (or in general any sample) over a time course is to set up an array experiment. Each of the array elements is an identical repeat of the same experiment. The interval between the array elements can be adjusted and the entire time-course of the reaction or time-dependent event can be tracked by monitoring the peak height, volume or other observables (such as peak positions).

The following instruction applies to the Varian instruments only. Briefly the steps are:

Test a regular, single element experiment to set proper nt and a fixed gain
array pad or nt to set up the series of experiments to build the time course
Type go to start experiment.

Test single experiment: set nt, gain, d1 (recycle delay), at (acquistion time), pad (preacquisition delay)

Set nt and gain

You should estimate or test run a 1- or 4- scan single experiment to estimate proper nt to use. Note that signal/noise ratio goes with square-root of nt.
Note that growth or drop in peak intensity over time should be taken into account so that the weakest signals can be detected with good sensitivity for analysis.
After finishing the regular 1D experiment run with autogain, type gain? at the command line. The value shown is the actual gain used. To prevent receiver or ADC overflow as the product grows over time, reduce this gain value by two units. That is, if the value is 30, enter gain=28 at the command line.
Check default d1, at, pad along with nt values and estimate the time for each element in the array. The total time for each element is δT = pad + nt*(d1+at) in second. For a slow reaction that takes hours, to avoid storing too many elements in the array, you may want to use either more scans for each element, or alternately you may use a larger pad value to wait in-between nt scans. However, note that this waiting period in second also applies to the first element. See below on how to kick-start the experiment immediately if pad is used as the array parameter.

Set up an array of parameter

In general, you can array any parameter if you just want to repeat the same experiment. Here, pad (pre-acquisition delay) is used an example. If pad=600 d1=3 nt=64, the experiment waits 600 secs, followed by 64 scans with 3 secs recycle delay between the scans. To check how long the experiment takes, type time.

If pad is arrayed, pad=600,600,600,600, ..... the experiment waits 600 secs (10 mins) before collecting a 64-scan spectrum, then wait another 10 mins, collect another 64-scan spectrum, .... until the last element of pad is done. All spectra are saved into the same experiment (as in a 2D experiment), with multiple elements. See note below on how to accurately calculate the delay time between the data points.

To set an array, from the command line:

type: array
Enter the parameter to be arrayed: pad
Enter number of array elements: 50 (or whatever desired)
Enter starting value: 600 (in sec)
Enter increment: 0 (this should be zero for any parameter you choose to array)

This sets up an array experiment with 50 element repeat of the same experiment with nt scans with 600 secs waiting in between. Type da and click Process->Text output panel to see the arrayed values.

To avoid the first element to start after a long pad wait, you may insert a brief wait to the front of the array, by entering: pad=0.1,pad. This allows the first experiment to start after 0.1 sec.

Type go to start experiment.

Check total experiment time

Type time to check total experiment time.

Other parameters to array if non-stop data collection is desired

If waiting between each experiment is not desired, you can set pad=0 and array nt and give it the same value.

From the command area, type array
Enter the parameter to be arrayed: nt
Enter number of array elements: 50 (or whatever desired)
Enter starting value: 16
Enter increment: 0

This example sets up 50 experiments in an array each with nt scans with no waiting in between the experiment except the recycle delay d1. Each experiment is saved in a separate spectrum in a single FID data file, just like the array for pad.

Type go to start experiment.

Array data processing and display

As the experiment goes, type wft f full aph to transform and auto-phase all. The 1st element is displayed. Make proper phase adjustment, expand region if necessary. Click Process->autoprocess to apply automatic processing (into a vertical stack mode with offset), but sometimes the macro doesn't work well. Alternatively, refer to the following commands for manual processing and array display:

wft (weighted Fourier Transform of all elements)
ds(1) f full (display the 1st spectrum in full window view)
wft dssh dssl (process and display all elements in horizontal view with index number)
dssh(1,50,5) (display array element from #1 to #50 with an index increment of 5, that is: #1, #6, #11, ....)
dssa(1,50,5) (as above but for vertical stack display)
vs=vs*.5 dssa (cut vertical scaling by half and display stack plot)

Printing of arrayed spectra

After processing and proper display is made on screen, modify and enter the following commands:

pl('all') pscale page (print all spectra with axis scale)
pl(1,50,5) pscale page (print spectra #1 to #50 with an index increment of 5)

Integrals can be either on or off just as in single spectrum display/printing.

Data analysis

Total delay between data points (or array elements):

The total delay between scans with pad array is D=pad+nt*(d1+at+extra) where d1, pad and at are actual vnmr parameters and extra is any extra long delays in the pulse sequence. For a simple 1D H1 spectrum, it is simply pad+nt*(d1+at). Type r1=pad+nt*(d1+at) r1? to display this value in second. However, you may want to look at the details of the experiment to make sure this is the best delay to use in your analysis.

In most cases, you want to plot the observables (peak height or volume/integral) as a function of time and perform fitting and further analysis. Peak volume or integral is preferred due to its advantage in being insensitive to drift in shims over the experiment duration. Except for simple exponentials, this should be done with your own fitting program with the text files generated below.

Peak height is simply the maximum peak height (relative or absolute) for a peak. Peak intensity is an term where there is widespread confusion about what it means and it is misused to refer to peak height in Varian manuals. It should mean peak volume per unit area. I suggest you stay with the term peak height and peak volume (or integral).

To use peak height as an observable (Y-axis) and time course as the variable (X-axis), follow these steps after processing the array:

Type wft ds(1) f full to display the first spectrum.
Expand and scale the area of interest. The area can contain one or multiple peaks of interest. Type dssh to display the arrayed spectra.
Locate which spectrum has the smallest but visible peak(s) of interest. This is often either at the beginning (for starting material) or the end (for product) of the array.
Type ds(#) where # is the index number to display the spectrum which carries close to the weakest signal you want to analyze.
Click the right side panel Threshold button and put the yellow line just below the peak(s) of interest.
Type dll fp to generate the line list and output the peak frequencies and peak height in a text file fp.out in your current experiment directory. To locate this file, from the Desktop, open your Home->vnmrsys->exp# where exp# is the experiment folder/number that contains your data (shown in top-left corner in vnmrJ).
Rename fp.out before moving on to next region
Rename fp.out to something like fp1.out
Repeat the above steps until all peaks of interest are done. Rename fp.out one by one if necessary.
Open fp*.out files by double clicking it. Make sure they have the right lists of peak index, spectrum index, frequencies and peak heights.
E-mail yourself these files or transfer them with FTP.

To use peak volume (to be continued)

H. Zhou updated Aug 2011