Spectrometer Quick

Start Guide
Captain Steve Byron
SESSION OBJECTIVES

• Operation Principles Over View

• What the Spectrometer can and
cannot be used for
• Using the spectrometer to analyze
visual and UV samples
WHAT’S INSIDE
This meter has two light sources. Each light source has an independent function
and specific task:

• A Tungsten filament, which generates light in the visible region of the

light spectrum IE 350 – 800nm.

• Deuterium filament which generates light in the ultra-violet region of

the light spectrum IE 190 – 350nm

The visible light source is used when we want to measure spectral signatures we
can see, such as reaction due to turbidity, color, etc.

The UV light source is used when we want to measure spectral signatures that
we cannot see in the UV frequencies
Light spectrum from 1mm through to
200nm
Similar to last slide but depicted as wave length: note that IR is to the right in this
depiction
 When using the meter, light, either UV or visible is shone through a sample. If all of the light passes
through the sample, it can be said that there is either
• ZERO ABSORBANCE
or
• 100% TRANSMISSION

• Samples that are visibly discolored absorb different wavelengths of visible light corresponding to the
colour of the sample. The more color the sample has the higher the absorbance reading.
• Therefore the absorbance measured directly corresponds to the amount of color in the sample

• If the sample was off color and looked yellow, light would be absorbed mostly in the yellow
spectrum, around 550nm.
• Therefore the frequency at which the absorbance is measured directly corresponds to the color of the
sample.
For general interest purposes, the meter we are using usually measures Absorbance. Zero absorbance is the
same as 100% transmittance, but the scales are not linear.

If a sample is colored, turbid, or has some opacity it’s fairly obvious that some visible light will be absorbed.
It so happens that the amount of light absorbed has a direct relationship to the amount of color, turbidity or
opacity. Because of this relationship the amount of color can be quantified .

It is perhaps less obvious that visibly clear and bright samples will also absorb light.
They appear clear and bright because the light being absorbed is outside of the
Visible spectrum: it is being absorbed in the UV spectrum so we can not see it.

Maybe these properties can be used

to detect impurities in samples?
 Well, yes they can!

Just as samples absorb light in the visible range they also absorb light in the UV
spectrum. Different chemical groups absorb different amounts and frequencies of UV
light

Using this principle we can identify the presence of specific contaminants in a sample
The principle the meter uses is Beers Law
Important Note: Ever sat in your car and got irritated ‘coz your photochromatic
glasses won’t go dark? This is because glass is not transparent to UV light: it
blocks it, and it’s the UV light that makes the coating on the glass lenses darken.

Why is this important? If we used a glass sample cell for UV analysis it wouldn’t
work, so the meter is supplied with two sample cells, one of regular silica glass for
visible spectrum work, and one of quartz glass for UV spectrum work.

Use the right one!!

This chart shows the frequencies at which
specific chemical groups absorb UV light.
Note the frequency band is small, limited 2.5

to near UV light only H

2.0 Y A C
If a sample shows a high absorbance peak

absorbance
D r a
at 200-210 we can say with very good R o r
E
certainty that hydrocarbon is absorbing UV 1.5 O m b
s
light. This means the sample is either a C N
t a o
hydrocarbon or contains hydrocarbon as a 1.0
A V
e t n
R
contaminant. M
r i y
B c l
s
So how do we know if its 0.5 O s s
N
contaminated or that the
S
sample is a hydrocarbon? 0.0
200.0 220.0 240.0 260.0 280.0 300.0

Any guesses? wavelength (nm)

 So, what
can we use
the meter We know the meter can
‘see’ color in the visible
for?
spectrum and measure
opacity and frequency.

This is pretty handy for wall wash analysis and giving a

number to a particular turbidity or color

We also know the meter can ‘see’ various chemical

groups in the UV spectrum

This is pretty handy for a whole range of stuff!

 • We can test wall wash samples for contaminants that we can’t visually see

• We can test lab reagents such as methanol and DI water for purity

• We can test cargo samples for purity – manifolds, first foot, finals

• We can run traditional UV tests to dispute third party cargo inspector results

• We can test wash water during the cleaning program for contaminants

• We can identify specific contaminants by chemical group

Using these capabilities allows us to track cleaning progress and be more efficient
with cleaning programs. We also now have a tool that documents cleaning progress at
each stage towards a specific goal, which can be used to dispute tank failures/claims
or, better yet - sell our service expertise to our customers
 Proper Operational Usage

The Spectrometer as supplied has been programed as a tool for two main areas of usage:

• Analysis of wash water: Wash water is analyzed at various stages of tank cleaning to
determine when a specific cleaning stage has completed and its time to move onto the next
stage, or that cleaning has completed to a specific standard. This is important to understand
to improve tank cleaning efficiencies.

• Confirmation that intermediate final wall wash results are within spec

Generally speaking the spectrometer cannot be used to

analyze products that are not water white. It would not be able
to analyze, for example pure palm oil or lube oil. This is not its
purpose and isn’t useful for us.

Can you explain why the spectrometer cannot be used for

palm oils or lube oils, or similar?
 Using the Spectrometer
Sample cell usage

The correct cell must be used!

For visible light analysis regular

silica glass is used. This is a
40mm wide cell marked with a
letter ‘G’
The correct cell must be used!

For UV light analysis quartz

glass cell is used. This is a 10mm
wide cell marked with a letter
‘Q’

Using the incorrect cell will

provide meaningless or
incorrect results
 Warning!
If you see wave scan results similar to these the incorrect cell has been used! Each cell has a different
signature so incorrect cell use results will not be the same each time.
 Using the Spectrometer
Light Source

Light source selection between UV or visible is automatic. It is chosen by the meter

depending on which programs are run

It’s important to warm up the light sources before

using the meter to take readings. The light
sources take a few minutes to stabilize and will
cause readings to drift if not warmed up
 Using the Spectrometer

A few slides ago we ask ourselves a question: How do we know if the meter is measuring a sample
chemical group or contaminant chemical group? Actually we don’t, but if we zero the meter each
time we use it with a blank the meter will ‘see’ the difference between the blank and test sample,
and give a ‘result’
While this might sound like a cop out its actually a pretty handy feature. If we had off spec
methanol with 5ppm chlorides to wall wash a tank with we’d be screwed trying to do a by eye test
for chlorides. However if we used this off spec methanol to set the meter to read zero chlorides by
using it as a blank the meter would see the increase in the wall wash sample and we’d have a result.
 Using the Spectrometer

Turn the instrument on:

Choose any method from the L&I Maritime programs

menu

Press the BLUE OA/100%T button. This sets the

instrument to read Zero

Press the GREEN button several times until the screen

reading stabilizes

• The meter may show a negative reading, but we are

looking for stability so this is not a problem

The meter is ready to use!

 Visual Range

Let’s take a look at how we would run a wall wash for chlorides. This
is a visual test we can see so it is performed in the visual range of
frequencies

A wall wash is taken, tested for hydrocarbons and found to be nil,

then treated with silver nitrate to detect the presence of inorganic
chlorides. If chlorides are detected the sample turns white or milky.
Remember we need a reference sample. In the picture at right the
clear sample is pure methanol with AgNO4 and the cloudy sample
is wall wash with AgNO4.

Any guesses or estimate of how many ppm? If you were going to

call me and describe the turbidity of the sample how would you get
it across? Pretty difficult I think! As an aside,
does the
reference
sample have
to be clear?
 Inorganic Chloride

Take the wall wash and prepare the samples: You guys tell me how to prepare the
One blank sample and one test sample. sample!

Turn the instrument on

Select main menu

Select Folder 2 - L I Maritime

Programs
Select Method 1
950 Chlorides in MeOH

Note: MeOH is an abbreviation for

methanol: Why ?
The test screen for Chloride in
methanol will now be displayed

We are only going to describe

one visual test here as all analysis
procedures are similar. The only
difference is the test type
selected from the five options on
the left in the slide above
Pour the reference sample into the 40mm
glass cell

Put it in the instrument as shown and make

sure it properly seated

Press the BLUE OA / 100%T Button – why?

Remove the sample cell and pour the

reference sample back in the tube it came
from

Clean the sample cell with methanol – why?

Pour the wall wash sample in to the cell and

put it back in the instrument
Press the GREEN button.

The program is set up to display chlorides

directly in ppm as seen at right

If multiple samples are to be tested, perhaps

from several tanks, or different areas of the
same tank, repeat the process cleaning the
sample cell between each analysis. Periodically
check at each second or third analysis that the
reference sample reads zero. If not, zero the
instrument by …….. ?
 Testing for Hydrocarbons
Note: All test for hydrocarbons should be run at 25:75 per
ASTM
The test procedure is similar to that for chlorides, so
we are not going to go over it step by step. However,
we need to understand what we are looking for and at
when we refer to hydrocarbons test. FTU 25

Hydrocarbon is measured using a relative scale called

FTU, showing turbidity of a sample. FTU does NOT
show concentration of any particular hydrocarbon.
This is because not all types of hydrocarbon in the
same concentration in methanol would show the
same turbidity. FTU 5

For example 100ppm gasoil in methanol gives FTU 23,

whereas 100ppm Octene shows FTU 3, and 100ppm
pygas shows FTU 0, even though all are 100ppm
‘hydrocarbons’
ASTM says Hydrocarbon is pass or fail. FTU 0 is a pass, but still has 100ppm pygas!!! How useful is a wall
wash then??
 H/C
To reinforce the
= 5 Product Concentration in MeOH Hydrocarbon Reading #953
(25:75)
concentration issue this table
shows FTU readings for MTBE 10,000 ppm (1%) 0
various products at various Pygas 100ppm 0
concentrations. Note that a
Pygas 1000ppm 92
sample can have 10,000 ppm
MTBE and still pass Gas Oil 100ppm 23
hydrocarbons per ASTM. Gas Oil 1000ppm 410
LAB 100ppm 24
Our customers and
charterers really aren’t doing LAB 1000ppm 480
themselves any favors when Styrene 1000ppm 2
they specify this test to
determine tank cleanliness, Styrene 10000ppm (1%) 560
but they are lead astray by Octene 100ppm 3
the survey companies. Octene 1000ppm 225
Because of this it is much
better to use wash water n – paraffin 100ppm 10
analysis to look for these n – paraffin 1000ppm 200
contaminants
 UV Range

Quick recap!:
1: We have the capability
When UV light is passed through a sample different
to identify different
chemical groups that the sample consists of absorb the UV
chemical groups in a
light at different amounts and frequencies, meaning that
sample based on
each chemical group has a unique UV signature. (Actually
frequency absorbance
each chemical has a unique signature)
2: We can identify
So this means …..?
concentration based on
amount of absorbance

Methanol is a really simple chemical so let’s take a look at its UV signature

Methanol is CH3 OH. This is quite a simple chemical so has a
relatively minimal UV signature.

This is a pretty useful characteristic. Methanol is a

reasonable solvent (alcohol) for lots of substances, but
because it has a simple UV signature it’s easy for the
meter to ‘see’ any contaminants in the methanol

IE: the methanol does not interfere with the UV

signatures of contaminants

This picture of methanol is called ball and stick

model
If we use a DI water blank to reference the
instrument and then run methanol UV we
would see a curve as at right. (DI water (H2O)
has no UV signature) This is called a smooth
curve.

There is high activity in the 200-210nm range

that can only be caused by CH3 or OH
groups.

If this was a wall wash sample it would pass

for ‘Transmittance’ (Remember we measure
‘Absorbance’, but it doesn’t matter as the
sample will pass for both)
 The spectrometer has This is a typical graph of pure methanol after
software installed that
transfer to a PC in .xls format. This is the format
can transfer data to a PC
in .xls files so that each we expect to receive during a tank cleaning
graph can be sent to us program
for review, and to build a
history of a particular
cleaning operation

This Data should be

maintained on
board for a 3 year
period in the
voyage file in case
it needs to be
referred to in the
event of a claim or
dispute.
 UV Scanning Operation for Wash Water Analysis

Analysis of wash water detects contaminants that are being removed from the tank and pipeline
during cleaning. Wall wash detects contaminants left in the tank after cleaning. What’s best and why?

I think it’s fairly obvious that if we are testing wash water and we don’t detect any more
contaminants the tank must be clean, so tank cleaning could be stopped once the wash water
sample is clear of contaminants. We are of course pulling wash water samples from BOTH sides of
the manifold.

Referring to the three superimposed

graphs to the left, wash water wash
analyzed after 15 minutes washing
shows quite high levels of Hc. A check
after 30 minutes showed very good
improvement but still some traces. A
check after 45 minutes shows no
contaminant remaining, IE the tank is
clean and washing can be stopped.
 When running wash water analysis the spectrometer must ‘know’ the condition of the sea water
being used to wash with, hence a base line must be established using water from the tank
cleaning line to compare against. The wash water sample should be taken from the manifold
rather than the pump stack: Why should this be so?

Sometimes we see no improvement to contaminant removal after extended washing: What does
this mean?
2.0 Referring to graph at left: there’s no
60 mins cold improvement between 60 minutes
1.5 sw washing and 90 minutes washing.
absorbance

90 mins cold
This means the water has removed
1.0
sw all the contaminant it can,
indicating that the next cleaning
stage must be started. You can see
0.5 100mg/l

0.0
from this that it’s important to take
200.0 250.0 300.0 350.0 frequent samples and analysis of
wavelength (nm) the wash water to see what’s
happening with progress. There’s
no point to continue washing this
tank with water
 UV Range

Turn the instrument on:

Choose any method from the L&I Maritime programs

menu

Press the BLUE OA/100%T button. This sets the

instrument to read Zero

Press the GREEN button several times until the screen

reading stabilizes

• The meter may show a negative reading, but we are

looking for stability so this is not a problem

The meter is ready to use!

 UV Range
The correct cell must be used!!

Selected folder number 6 called ‘Wave

Scan’

Fill the 10mm quartz glass cell with

reference seawater taken from the tank
cleaning line: be sure to flush the tank
cleaning line of all rust etc. prior to
collecting the reference sample

Place the cell as shown in the far right

of the cell holder.

The correct cell must be used!!

Press the Blue OA/100%T button to
zero the spectrometer to the
seawater sample

A blank graph will be displayed as

seen at left.

Remove the 10mm cell and dispose

of the reference sample. Clean the
cell by flushing with wash water
sample, refill the cell with the wash
water sample and place it back in the
sample holder to the far right.
 acetone in water
0.5
Press the Green button which will start
the scan.
0.4
A graph will be generated showing

absorbance
peaks where contaminant is detected. Carbonyl (C=O)
That shown at right shows acetone in 0.3 Group
water, for example.
0.2
Periodic readings of the wash water will
be taken to track cleaning progress,
until all contaminate has been removed 0.1
or until the water is no longer effective.
0.0
200.0 250.0 300.0 350.0

wavelength (nm)
 Interpreting the Data

Just a quick note on determining what and how much contaminate there is in each
sample 2.5
H
Looking at the overlay on right, the wavelength Y A C
2.0
position of a peak or peaks indicates what chemical D R A

absorbance
R E O R
group the contaminate consists of. The height of the O
1.5 S M B
peak on the absorbance axis shows amount of C N
T A O
V
contaminate. 1.0
A
E T N
R M
R I Y
Knowing these two parameters, and using periodic B
O S C L
0.5
sampling to track progress allows a specific cleaning N S S
method targeted to the contaminate, and informs us S
0.0
when cleaning is complete. 200.0 220.0 240.0 260.0 280.0 300.0

wavelength (nm)
Being able to determine contaminates to this level of accuracy significantly improves tank cleaning
efficiency by allowing targeted techniques and eliminating over washing.
 Getting Help

This meter is being rolled out to all FCC owned tonnage. We expect the meter to be used for each tank
cleaning irrespective of the standards being cleaned to. Using the spectrometer for routine cleanings will
provide familiarity with its use, and build a data base of results that can be later referred to. FYG there is a
significant data base of graphs in Milbros website.

The meter is robust but needs to be taken care of. It’s expensive! We expect it to be carefully looked after
and treated with respect as a scientific instrument. There are no user serviceable parts in the meter. The
light bulbs can be expected to last many years with proper warm up and use.

If you have difficulty using the meter, experience issues or erroneous readings, or have issues transferring
data to ship PC for emailing, please contact your operator or Captain Steve Byron.

Happy testing and success in reducing tank cleaning resources!