Standard operating procedure Bruker Vertex 70 VNIR and TIR spectrometer

(1)

1

Standard Operating Procedure

Bruker Vertex 70 VNIR and TIR

spectrometer

version 1.1, November 2016

Contact

Chris Hecker (Room 5-035 -

c.a.hecker@utwente.nl

)

(2)

2 Version history

Version 1.1 2016nov02

first public release after reworking several sections and including feedback from beta test Version 1.0 2016jun24

pre-release for module 13 beta testing

For comments and suggestions for improvement, please contact Chris and Evelien

c.a.hecker@utwente.nl

(3)

3

Introduction

When operating the Bruker Vertex the most important thing to remember is that you want to keep the system stable. Hence, the variables that you encounter in the measurements are caused by variables in your samples and not by variables of the system itself. This SOP is designed to help you achieve a stable system that allows you obtain reliable measurements. This document should be perceived as a

guideline, if you would like to deviate from the SOP approach for your experiment please discuss options with Chris Hecker.

Before you start

List of contacts:

Dr. Chris Hecker, Assistant Professor, Department of Earth System Analysis room 5-035

phone +31534874356

c.a.hecker@utwente.nl

Evelien Rost, PhD, Department of Earth System Analysis room 5-038

e.rost@utwente.nl

Dr. Caroline Lievens, Head Geo-Science Laboratory room 4-035

phone +3153489344

c.lievens@utwente.nl

Watse Siderius, Education and Research Technician room 0-011

phone +31534874540

w.s.siderius@utwente.nl

The Bruker Vertex is generally left on, only when it is not used for a period of a few weeks or more will it be turned off. This means that, normally, you do not have to switch on the measuring device. To check if the device is on, see if lights on top (humidity/laser/status) are on. If not, the switch to turn the Bruker Vertex on is at the back of the apparatus.

The lights for the humidity and the laser should be green. The status light can be red if the detector is not yet properly cooled.

If humidity light is red, contact Watse Siderius for replacement of silica cylinder. If laser light is red, contact Chris Hecker, do not proceed with measurements!

(5)

5

Start-up of system

To login on the computer: Username: spectrometer Password: asdasd

After you logged in on the computer you will see the Bruker monitor program (fig. 1)

Figure 1 Bruker monitor program Here, you can see:

• Active source(s)

o Note, however that the computer cannot determine if the external source is turned on, it can only detect if the mirror is positioned in such a way that the beam can reach the detector

• Selected detector • Selected beam path • Selected beam splitter

Additionally the program also shows if the computer is connected to the Bruker Vertex and if the information is up to date (refresh check mark should blink).

External source

After you checked that the Bruker Vertex is turned on and working properly, the external source should be turned on.

• Turn on A • Turn on B (back) • Turn on C

It is very important that you first turn on the cooling (A) of the external source! Forgetting to do so can trigger the overheating protector switch

(6)

6 If you turn on B or C before A is switched on, the external source is internally switched off. To correct for this, the external source needs to be opened which takes ~30 min. If this is happens contact Chris Hecker.

To check if you turned on the external source correctly, you can feel if it’s slightly warm after ~15 min or you check the signal (see Operate OPUS)

Cooling of the detector

The procedure to retrieve the nitrogen is as follows: • Take nitrogen cup to room 0-060

• Lift nitrogen container out of kart • Take lid of nitrogen container

• Pour nitrogen in cup and place lid on cup • Place nitrogen container back in kart To cool the detector liquid nitrogen is poured in the detector (fig. 2)

• Take lid of detector

• Use funnel to pour liquid nitrogen in detector • At a certain point the detector will seem to be

full

• Wait until small outburst • Then fill detector completely • Take funnel out

• Put lid back on

When working with liquid nitrogen there are a couple of rules that you need to apply: a) No open shoes

b) Put on safety glasses

c) If you spill a large amount of liquid nitrogen, leave the room, alert others and contact Caroline Lievens or Watse Siderius

You can check if the detector is completely full by looking in at the top (you see the liquid), however,

(7)

7 After you filled the detector with liquid nitrogen you wait 1 hour to let the detector cool. Detector stays cool for about 6-8 hours but typically not long enough for entire work day. In that case refill the liquid nitrogen in the detector around lunch time and wait for 15 minutes.

Additionally you have to make sure that the mirror of external detector is aimed at the sample (i.e. black handle turned entirely to the left, fig. 2) and not the reference position.

Purging the system

o Open the N2 container by turning the handle to your left-hand side (clock-wise) o Check if there is still sufficient N2 left in container (right meter)

o Set the reduced pressure of the N2 to 5 bar (left meter) o Set the purge to 100 L/hour

Operate OPUS

Now you are ready to start operating the Bruker Vertex.

You can start the operating program OPUS 7.0. A login is required (fig. 3). User ID: Administrator

Password: OPUS (in CAPS)

Figure 3 OPUS login

When you login on OPUS a beep in the Bruker Vertex will confirm you have a successful connection. In the lower right corner of the OPUS program a green, yellow or red dot is visible. If you click on this dot you can check if all parts of the instrument are operating properly (fig. 4). If the detector is not yet properly cooled, it will be red. If other buttons are red or yellow see Problem solving/FAQ for instructions.

(8)

8 Figure 4 Instrument Status

Before you start measuring you first need to check the signal (fig. 5). • Go to Advanced data collection

• Select the tab Check Signal

Figure 5 Check signal tab Amplitude

(9)

9 Here, you can check the amplitude of the signal. If there is no signal the amplitude is ~5. When the signal reaches the detector the amplitude is significantly higher (up to 15000; depending on setting and what sample/gold is in the sample port). The amplitude is a measure for the stability of the system

Placing the sample

Before you measure the gold standard, clean the edges of the sample port with a paper towel to prevent any dust from previous samples to go onto the gold standard. Don’t stick anything into the integrating sphere. You can bring the gold standard flush in contact with the sample port by jacking up the lab jack. Make sure the gold surface is flat against the underside of the sample port with only light contact! (Fig 6) Make sure you don’t touch with your hands or scratch the surface in any way! If the gold standard appears dirty, contact Chris.

When choosing the measurement location on your sample, make sure you have a spot that has a

diameter of at least 3 cm and is as flat as possible to cover the sample port without leaking light. Leaking light causes lowered reflectance values. Use foam covered in aluminum foil to get the sample in the right position and angle on the lab jack. The laser pointer is not in the middle of the beam bundle and cannot be used for positioning of the sample. Look from several sides of the sphere to position the sample properly before jacking it up the last millimeters.

Figure 6; gold standard and sample under the sample port

Select the most flat side of the sample and when you have leakage of light always note this in your notes and the logbook (it will result in lower absolute reflectance values). When determining the most optimal position of the sample under the detector keep in mind that the sphere entrance is ~3 cm in diameter and the sampling spot about 2.5 cm.

Please write down the amplitude in the logbook at the beginning of each day (with infragold standard in the sample port)

(10)

10

Starting a measurement

Make sure that you have the gold reference and the sample ready and leave the detector entrance exposed as shortly as possible when placing gold reference or sample (so the purge of the system is not “broken” for too long)

Go to Advanced data collection (fig. 7)

• Select the Advanced tab and load the correct experiment (typically the starting point is

ITC_Sphere_MIR_4wn.xpm)

• Check settings of the experiment (file name, path, resolution, etc.) • For Results Spectrum select Reflectance

• Check all boxes under Data blocks to be saved

Figure 7 Advanced data collection

Now that you have selected the right settings you go to the button Auto RepeatMeasure (fig. 8) Check the settings, but you typically only have to change the following two settings

• Check if the Experiment Name is correct • Determine the Output Filename

(11)

11 • Determine the Output File Path

o Choose your personal output folder in the Vertex data folder • Click Continue

• The system will now instruct you when to put the gold reference under the detector followed by the sample. Follow the instructions in the pop-up window.

• If you have questions about the other settings, you can contact one of the auhors

Figure 8 Auto RepeatMeasure

Dark current measurement

For best absolute reflectance (or emissivity) results you need to measure a dark current measurement once a day, preferably at the end of the day because you leave the detector entrance open during this measurement. To perform a dark current measurement you use the same settings as with a sample measurement but after the infragold is measured you take the sample table away so that the light beam is aimed at the floor. This dark current measurement can be used in the data processing to post-process the files (see below).

Note, that if you do not want to proceed with a measurement use cancel and do not close the window with the cross button. Otherwise the program might crash or the program will start a measurement anyway.

(12)

12

Judging quality

To judge the quality of your sample measurements you need to determine if there is significant drift present between the 8 repeat measurements of the same sample (ideally, they should look identical except for the noise on top of the signal). To do this you can divide your measurements in three colors:

• The first measurement • Measurement 2-7 • The last measurement

Now, you can determine if there is significant drift in your measuring sequence. Good quality

measurements should not display significant drift. Please see the FAQs for further explanation on how to judge quality of your spectra.

Shutdown after measurements

After you successfully measured your last sample of the day, remove it from under the detector

entrance and gently jack-up the sample holder with the aluminum foil against the sample port entrance to avoid humidity and dust from entering

External source

When you are done measuring you should turn off the external source • Turn off C

• Wait two minutes • Turn off B

• Turn off A

Internal source VNIR/SWIR

If you have use the NIR source: To safeguard the lifespan of the NIR lamp, this source should always be turned off after measurements. The internal MIR source can stay on. You can use the Bruker monitor program (fig. 1) to check which source is currently on. You turn the NIR source off in the Advanced Data collection (fig 7). Use the tab “optic” to change to the internal MIR source, and the tab “check signal” to execute the change.

Purge system

Close the N2 container by turning the handle to your right-hand side.

Computer

• Copy your files from the computer to personal usb

o Including the experiment file that you use for your experiments! (fig. 8) for safeguarding • Close all programs except for the Bruker monitor program

• Shut down the computer

•

Make a copy or a picture of the logbook for safekeeping.

Note that there is no guarantee that your files are safe on the D drive, so to not forget to make a back-up on private usb at the end of each day

(13)

13

Process Data

The final output will be absolute reflectance (or emissivity) spectra in ENVI Spectral Library format. Two software packages were developed to help with the processing of the BRUKER VERTEX files in OPUS file format (from here on called <Opus files>).

I. OPUS Conversion (tkOPUS.py) in a software called HypPy (Hyperspectral Python). tkOPUS reads the binary OPUS file format and extracts the data that is required. If necessary, some averaging is done and files are written to the ENVI Spectral Library file format.

II. The Bruker Darkcurrent Correction script (in IDL) takes the Bruker spectra and applies a correction for darkcurrent as well as for the imperfect reflection behavior of the InfraGold reflection standard. The output is an ENVI spectral library with absolute reflectance (or emissivity) values.

Open OPUS converter

Install HypPy according to the instructions on <P:\ITC\GroupData\ESATools\README-HypPy.txt>. If you don’t have access to this server location AND you cannot find HypPy installed on the Bruker Vertex computer, contact Wim Bakker (w.h.bakker@utwente.nl) or Chris Hecker (c.a.hecker@utwente.nl).

• Select menu item <Conversion/Convert OPUS files  the Opus Converter interface opens (fig. 9).

Check the spectra

A) Select the OPUS files that you want to process. You can use asterisk as placeholders to select all files in a given folder. Or you can use crtl or shift to select multiple files in a folder.

o Do not forget to include your darkcurrent files

o Using long folder and file names can give problems. In that case copy the files into a simple location like shown here.

B) Select to extract the reflectance spectra

C) Select to average spectra by basename, plot spectra and plot standard deviations of spectra o files with same name and different extensions are considered repeat measurements and

are averaged D) Run the process

The results will show the averaged sample spectra (fig. 10) as well as their standard deviations (fig. 11). Zoom in and check that the standard deviation spectra only show elevated values for CO2 and water feature but not for rest of spectrum. If a spectrum has a considerably higher stdev than the rest, this could indicate that one of the repeat measurements went wrong (e.g. not sufficient lN2 cooling in detector, see FAQs).

(14)

14 Figure 9 OPUS converter

Figure 10 OPUS converter plot window: averaged spectra Figure 11 OPUS converter plot window: standard deviation

A

B

C

(15)

15

Export data to sli

we will now export the data to an ENVI spectral library. • Change the selections in the interface to fit figure 12:

o Select ScSm (single beam sample spectra) o Select Average spectra by Basename o Select Plot Spectra

o Deselect Plot StDev of Spectra o Specify a folder at for Speclib

 Try to stick to the naming convention with a final “_ScSm.sli” since that will make life easier later in the process

o run the export of the ScSm

• Change the selection in the interface to fit figure 13: o Select ScRf (single beam reference spectra) o Select Average spectra by Basename o Select Plot Spectra

o Deselect Plot StDev of Spectra o Specify a folder at for Speclib

 Try to stick to the naming convention with a final “_ScRf.sli” since that will make life easier later in the process

o run the export of the ScRf

Figure 12 export single beam reference spectra (ScSm) to sli Figure 13 export single beam sample spectra (ScRf) to sli

(16)

16 Bruker Dark current correction

• Find the following three files and save them to your harddrive. The infragold spectrum (.sli and .hdr) should be located in the same folder as the input spectral library files that you created in the previous step, or in the root of the D: drive (D:/infragold_IRT-94-020_extended_1to20micron.sli).

o <bruker_darkcurrent_v11.sav>

o <infragold_IRT-94-020_extended_1to20micron.sli> o <infragold_IRT-94-020_extended_1to20micron.hdr>

The next step involves ENVI + IDL which you should run on your own computer or a computer in the library, since the license of the lab computer is expired.

• Open ENVI + IDL (version 5.1 or higher). In the IDL interface (fig. 14) go to the command line and type the following commands:

o Restore, dialog_pickfile() <press enter>

Browse to the location of the <bruker_darkcurrent_v11.sav> and select it. o In commandline, type: bruker_darkcurrent_v11 <press enter>

• In the next two interfaces (fig. 15 and 16), select the two spectral libraries that you produced with HypPy (.ScSm and .ScRf). If you did not use similar file name conventions as in this document, select to show all files, but make sure you select first the ScSm and then the ScRf file.

(17)

17 Figure 15 choose the ScSm file Figure 16 choose the ScRf file If you have different file naming convention, display all files (red box), but be careful that you select the correct file.

In the Bruker Darkcurrent Correction interface (Fig. 17), select

o The spectra that you want to apply a dark current correction to o The dark current measurement that should be used for the correction

Here, <empty_2013mar12> is used

o The output file name for the spectral library (a default is suggested but can be overruled). Careful: existing files will be overwritten!

o Check the IDL command line window to see that no errors occurred .

Note, that you can use the same dark current measurement (empty sample port measurement) for spectra that were acquired with the same settings; dark current measurements are typically acquired once a day. If you need to process files from several days, you select the files of a given day to be processed and the darccurrent of the same day. Then you repeat the whole process for the next day. The output files produces by IDL are spectral library files (.sli). The output format is in reflectance (absolute directional-hemispherical reflectance) or emissivity through Kirchhoff’s law of thermal radiation (E=1-R). These files can be opened with HypPy’s Spectral Library Viewer.

(18)

18 Figure 17 Bruker Darkcurrent Correction

(19)

19

Problem solving /FAQS

Figure 18 Instrument status panel (detector error is caused by no cooling yet; laser warning is due to aging laser).

Instrument status problems

Instrument status problems result in a yellow or red colored ball at the bottom right corner of the OPUS window. To open the status panel (Fig 18) click on the ball.

The only icon that generally turns red or yellow is the detector button, indicating that the external MCT detector is not yet sufficiently cooled.

The following situations indicate that some maintenance is necessary in the future, therefore if you encounter these situations please note it in the logbook and inform Chris.

• If the “passed” icon is yellow (“passed” stands for an instrument quality check) • If the source icon is yellow and you are measuring with the external source • If the laser button is yellow

In all above mentioned situations you can proceed with your measurements.

• Passed icon is red

• Source icon is red

• Laser icon is red

• Interferometer icon is red or yellow

• Electronic icon is red or yellow

• Automation icon is red or yellow

In all other situations (see below) you should not proceed with your measurements and you should contact Chris.

(20)

20

FAQs

What should I write down in the logbook?

• The date

• Amplitude at start • Your name

• The name of experiment that you use • Purge yes/no; when started; liters per hour • Instrument status if yellow or red

• Liquid Nitrogen (re)fills with time • Sample names and starting time

• Possible light leakage, bad contact or any other observation that could explain problems that you may find with your data afterwards.

Additionally, any information that is useful to future users of your data or for troubleshooting data problems later on, should be noted in the logbook. It may be advisable to also make a personal logbook that includes personal notes and ideas. Make a scan / photo of the official logbook and keep it with your data for safekeeping.

How long does one measurement take?

If you use the standard number of cycles (8) your measurement (e.g. gold reference and sample) will take ~30 minutes. Including the cooldown and empty sample port, you can measure about 10 samples per day.

Can I stop a measurement?

You can stop a measurement by clicking abort or stop measurement. Note, however, that a small file will still be created and if you did not change the file name your previous measurement might be (partially) overwritten by this file. Before you continue, you may also want to cleanup the interrupted file in the folder that holds your data, so you don’t get confused later.

Why is the system purged?

The system is purged to prevent energy loss of the signal from atmospheric interference (water and CO2). If you do not purge the system ~16% of the signal is lost due to presence of atmospheric molecules in the air. Through purging these molecules are continuously removed from the system.

What should I do when:

• There is no signal

Check if the beam path is correct: Go to Advanced data collection (fig. 19) o Select the Beam Path tab

o Check if beam path is like in figure 19 o If not, change path to match

(21)

21 Figure 19 Checking the beam path

Check if the external source is turned on:

o To check if you turned on the external source correctly, you can feel if it’s slightly warm. o If it’s not warm, check if A, B and C are turned on (next to C is a switch “Man/Auto” which

should be off).

o Follow instructions of SOP to turn on external source.

o If it is still not working, contact Chris or Watse to reset the overheating safety switch inside the external source

• There is a lot of noise in the signal Check if purging still works properly: o Check if the N2 container is opened

o Check if there is still N2 left in the container

o Check if the pressure of the N2 is ok (should be ~5 bar) o Check if the purge is still 100 L/hour

Check if there is a signal:

o Go to Advanced data collection o Select the tab Check Signal o If no signal, see above

(22)

22

How do I know that the detector is still properly cooled?

If the temperature of the detector is rising, this will show in your resulting spectrum. During the course of one measurement the complete spectrum will drift as a result of this. However, the temperature monitoring of the detector in OPUS will not yet indicate a problem. When OPUS indicates that the detector is not properly cooled you are already too late for a refill, because the spectrum of your measurements have already been influenced.

I see drift in the results of my sample (8 measurements), where does this come from?

Local drift in the spectrum (around the water bands and general albedo) suggests that the moisture content of your sample is changes during the course of your measurement. Drift of the complete spectrum suggests that the temperature of the detector is changing, hence you need to refill the liquid nitrogen in the detector and re-measure the last sample.

What happens when the input file in Advanced Data Collection is different than the input file loaded in the Auto RepeatMeasure?

When this happens, the file input in Auto RepeatMeasure is used. Hence this input defines the experiment set-up.

What do all plot abbreviations mean (fig 20)?

• Refl = reflectance; ratio of the SC and the IFG

• SC = Single scan sample (s) and reference (r); no physical value, only relative detector voltage • PH = phase; relates to IFG van Fourier Transform (FT)

• IFG = Interferogram sample (s) and reference (r); original measurements • History; shows acquisition and processing history

(23)

23 Figure 20 Examples of data blocks that are saved in each measurement file

Why do I need to perform a gold reference measurement?

The spectrum that is measured when measuring a sample is composed out of spectral features from e.g. the source, the apparatus itself, the beam splitter, the atmosphere within the apparatus and the sample. However, you are only interested in the spectral features originating from the sample. Therefore, to eliminate all the other factors influencing the spectrum you perform a gold reference (or “background”) measurement that composes of all the spectral features minus the features originating from the sample. Sample measurement and background measurement are ratioed to get the reflectance spectrum of the sample

Why do I need to perform a dark current measurement?

The spectrum that is measured when measuring a dark current is essentially the base signal of the system. Hence, it is the signal that the system registers even if there is no energy (e.g. light) from the sample entering the system. The dark current is measured once a day to allow for the correction of results. Additionally, frequent monitoring of the dark current is also essential to keep track of any fluctuations in the system.

(24)

24 No you do not need to darken the room, because you are measuring in the thermal range of the

spectrum. Therefore, natural light will not greatly influence a dark current measurement, since it does not provide a lot of energy in this spectral range.

Why do you measure a sample in multiple cycles whilst the gold reference is only measured once?

The reflectance of the gold reference is almost 100%, whilst the reflectance of, for instance, rocks varies between 0 – 30 % with an average of 10%. Therefore, a lot of energy is lost and the noise is much higher than in the background measurements. To suppress this noise and increase the signal-to-noise ratio, the sample is measured multiple times and the results averaged. Having several measurements per sample also allows to see if the conditions (e.g. sample moisture content; atmosphere; cooling state of detector … ) changes during the 8 repeat measurements.

When I want to load my ScSm and ScRf files in the ENVI darkcurrent removal program does not recognize them.

This is probably because you forgot to type .sli behind the files you created in the OPUS converter of HypPy.

I can’t find my darkcurrent fil in the spectra that I can select in ENVI for the darkcurrent correction.

Did you include the darkcurrent file when you used to OPUS converter in HypPy?

How do I identify problems with the data that I measured (quality control)?

Below are a few examples of what you may see in the data. If you have 8 repeat measurements, make 1 to 6 the same color and 0 and 7 a separate color. That way you can identify how things changed through time.

Figure 21 shows an example of a drying sample. This case it is a sandy soil with soil moisture. The soil moisture evaporates during the measurement (also due to the dry purge air that flows over the sample) and gradually changes the reflectance values from the first measurement (blue) to the last

measurement (red). The reflectance changes gradually rather than starting to change at a certain repeat measurement. There is not much you can do about this. You could try to not use purge gas but then you have noise in the water bands. The red circle shows the CO2 features which often change during

measurements. Not much you can do about this.

Figure 22 shows an example of a warming detector. The first few measurement (blue and underlying green) are identical and as they should be. When the detector cooling is non-optimal anymore, the results start to drift (e.g. yellow, orange red). Throw away the measurements of this sample and start over.

Figure 23 shows spectra that are plainly wrong (red). Since all repeats are the same, the problem is most likely with the gold standard measurement. These problems can be identified with some experience and certain expectations of how the spectra should look like. In this case the reflectance peaks in the left half of the graph looked very different from a typical spectrum (green) and indicated a problem. If in doubt, quickly start re-measure a sample with a single iteration. While the noise levels are quite high, the values and patterns are already visible for double-checking the results.

(25)

25 Figure 21 Drying sample (CO2 “issue in red circle)

(26)

26 Figure 23 weird results

Standard operating procedure Bruker Vertex 70 VNIR and TIR spectrometer