Application

Note: 52020
Application of Automated Data Collection to
Surface-Enhanced Raman Scattering (SERS)
Timothy O. Deschaines, Ph.D., Thermo Fisher Scientific, Madison, WI, USA

Introduction Within Array Automation are several options for data

Key Words Surface-enhanced Raman scattering (SERS) is a technique collection. For each of the 12 positions on a slide, a single
that is gaining more use and popularity as an analytical spectrum or a grid of spectra can be collected. If multiple
• Array Automation
technique. The potential for SERS to open up Raman spectra per position are desired, a grid of points, up to 13
• DXR/SERS spectroscopy to new sample types and analysis methods by 13, can be collected for each position. Several parameters
Analysis Kits continues to grow. Some of the practical applications of for grid collection are available; one is step size between
SERS, including those of biological origin, involve samples points, with options ranging from 10 to 1000 microns.
• Gold Colloids
that are often numerous in nature or require multiple The overall grid size can range from 30 microns per side
• Ink Analysis instances of sampling for statistical verification. Individual to at least 5000 microns per side, while the overall grid
SERS sampling may make the SERS technique too tedious size will be limited by the size of the actual sample region,
• Laser Power
to be practically applied. Combining multiple sample which is part of the template. When a grid is collected, the
Control
slides and Raman automation enables SERS to be used user can save either each individual spectrum or create an
• MicroRNA with real world sampling scenarios with speed and reduces average spectrum for each location. For irregular samples
human error. This application note discusses the use of or samples that don’t fill the entire analysis area there are
• OMNIC Software
automated collection and analysis for SERS and introduces two options for optimal data collection: the software can
• SERS the capabilities for high throughput, multi-sample analysis either search for the strongest signal using a defined grid,
using the Thermo Scientific DXR Raman microscope. or the user can manually search the sample and focus on
• Surface-Enhanced
Two different sample sets were analyzed to illustrate each sample location before collection.
Raman Scattering
approaches to automated SERS sampling. The first sample Data from a collection can be analyzed in the Array
• TQ Analyst set consisted of microRNA, which are small RNA molecules, Automation software, in Thermo Scientific OMNIC soft-
typically in the size range of 21 to 25 nucleotides in ware, or in the Thermo Scientific TQ Analyst chemometric
length. Research has shown that microRNA are involved software. Array Automation is also directly exportable
in gene expression, cell regulation, and as potential disease to laboratory information management systems (LIMS).
markers. Thus a high throughput method is required for A large number of analysis techniques are available in
analyzing large numbers of samples to identify if certain Array Automation.
microRNA are present, and therefore a specific disease. Metrics Available for Analysis in Array Automation
The second sample set consisted of 12 samples of ink
• Area above baseline • Peak area
on paper. Forensic analysis of inks on paper is important
• Correlation • Peak area ratio
for the analysis of questioned documents to determine their
• Cluster analysis • Peak height
authenticity, to determine what changes might have been
made to the document and to help identify the ink(s) used. • Group analysis • Peak height ratio
• Height above baseline • Principal components
Automating Data Acquisition and Management • Multivariate curve resolution • Quantitative result
A critical component of this work was Thermo Scientific This note focuses primarily on the use of Array
OMNIC Array Automation software. Array Automation Automation for the collection of large data sets for multiple
is used for the automated collection and analysis of multiple SERS samples such as microRNA.
samples. Array Automation controls the movement of
the motorized stage of a DXR Raman microscope, or Experimental
the well-plate accessory of a Thermo Scientific DXR For the microRNA portion of this project, Thermo Scientific
SmartRaman spectrometer, and coordinates the stage miRIDIAN microRNA samples were used. Samples arrived
movement with the spectral data collection of the samples. lyophilized in reaction tubes and were stored at -80°C until
Array Automation includes templates for many common needed. Sample solutions were prepared by adding RNase-free
multi-sample platforms, such as a 96 well-plate. New water to the reaction vial; the vials were shaken to ensure
templates can be easily created in the software. A template dissolution of the entire sample that may have been on the
for the 12-spot gold coated walls of the vial. The final concentration of the solutions
microscope slide used for SERS was approximately 1 microgram per microliter, as not all
analysis, which is part of the samples contained the same amount of microRNA. Each
DXR/SERS Analysis Kit, was solution was separated into several aliquots and then frozen
also created. at -80 °C until needed.
 For the collection of SERS spectra, a DXR Raman Data collection and analysis was performed using
microscope equipped with a 780 nm laser, brightfield/dark- OMNIC™ software and Array Automation software. As
field illumination, 20× microscope objective, and a motorized mentioned earlier a software template for the 12 position
microscope stage was used. The samples were analyzed gold slide was built and used for the spectral analysis of
using 1 mW of laser power. Important to the analysis was the slides.
the laser power control of the instrument, as it ensured that A set of spectra was collected for each sample spot, with
samples were not damaged by the laser during collection, the collection set up and run using the Array Automation
and that the CCD detector of the instrument was not software. Figure 1 shows a screen capture of the set-up
saturated by signal. window for Array Automation showing the template for
A DXR/SERS Analysis kit was used for the microRNA a 12-spot slide, with 6 spots selected for a grid collection.
sample preparation. The SERS samples were prepared by A square grid 650 microns on each side was collected,
mixing 2 microliters of one of the aqueous microRNA using a 50 micron step from point to point. This generated
solutions with 2 microliters of the 70 nm gold colloid 169 spectra per spot on the slide.
(also aqueous) using a micropipette. Two microliters of In the ink analysis section of this project, samples were
the resulting solution were then deposited onto one of the prepared by collecting twelve different ink sources (pens);
12 spots of the gold slide (also part of the analysis kit). eight black inks, three blue inks, and one red ink. A paper
The samples were air dried before data collection. A template was created that matched the size and dimensions
verification solution, which is part of the DXR SERS of our 12-spot microscope slide. Ink samples were deposited
analysis kit, was used in a preliminary test to verify that on specific spots by simply writing on the paper. The tem-
the combination of gold colloid particle size and laser plate was attached to a microscope slide, by taping along
wavelength could give a useful SERS response. the edges. For SERS analysis a silver colloid solution was
prepared using the standard citrate Lee & Meisel method.1
A total of 12 microliters of silver colloid were applied to
each ink spot, in a series of three microliter aliquots, with
time allowed between each application for the samples to
dry. Untreated ink samples were also prepared and analyzed
for comparison.
Spectra of the treated and untreated samples were
collected using the DXR Raman microscope, this time
equipped with a 532 nm laser, 10× objective, motorized
microscope stage, and brightfield/darkfield illumination.
The laser power used was 2 mW, with a 25 micron slit
aperture. 30 one second scans were collected per sampling
location. For the red ink sample a lower laser power was
necessary due to strong background fluorescence, thus a
0.2 mW laser power was used, coupled with 30 scans that
were each 0.2 seconds long. The 12-spot microscope slide
template in Array Automation was used, with a 13 × 13
Figure 1: Array Automation set-up window showing the template for a
step grid per sample spot with a 50 micron step between
12-spot slide, the 6 spots on the left side of the slide (A1-A3 and B1-B3)
have been selected for a multi-spectrum grid collection sampling locations, resulting in a total of 169 spectra over
a sample area of 650 microns by 650 microns, overall
more than 2000 spectra were collected for each slide.

Results and Discussion

MicroRNA Analysis
Figure 2 shows the average SERS spectra of three of the
microRNA samples that were analyzed. Each spectrum is the
result of averaging a set of spectra collected on a microRNA
SERS spot on the 12-spot slide. In Array Automation each
individual spectrum from a grid can be displayed or analyzed,
or the software can collect all the spectra for a sample spot
and create an average, in this work individual spectra were
collected. As can be seen in Figure 2 the spectra share many
similarities, but there are noticeable spectral differences
between the samples. Some spectral similarities are expected
due to the samples being made of similar nucleotides,
Figure 2: Average SERS spectra of selected microRNA samples (spectra were corrected using the however the different arrangements of the nucleotides are
fluorescence correction algorithm)
what lead to the noticeable spectral differences. This is one
of the benefits of Raman and SERS. Figure 3 shows the
results of a 12-spot analysis and how Array Automation
displays all the data points, with the squares color coded
based on an analysis algorithm. In this figure it was area
above baseline for each spectrum, the analysis method can
be changed and a new set of results will be displayed for a
quick visual analysis.

Figure 5: MicroRNA library search results for spot B4, which was correctly identified as microRNA
sample #11 (Match % 98.71)

Figure 3: Example of Array

Automation spectral collection
results, the large grid shown to Ink on Paper Analysis
the right is an expansion of the There were two goals for the ink analysis, one was to use
data for spot B3
the spectra from the ink samples to show that SERS is useful
in the analysis of inks on paper and the second goal was to
develop a method to discriminate all the inks, particularly
For one analysis a slide was prepared using duplicate
those that are very similar in composition such as one black
sets of the six samples. One set of samples was designated
ink versus another. Figure 6 shows the Array Automation
as a known set and one set was designated as an unknown
results for the analysis of the 12 samples of inks on paper
set. The sample layout is shown in Figure 4. For this data
treated with silver colloid, also analyzed was a set of
set a spectral average was generated for each sample spot.
untreated ink samples. As mentioned previously each spot
Spectra from the known samples were used to build a
contains 169 spectra, in a 13 by 13 grid. By collecting such
spectral library, using the standard OMNIC software. The
a large number of spectra a better data set can be generated
spectra from the unknown samples were then run against
because different fiber orientations, ink coverage, and other
the library of knowns. Figure 5 shows the result of one of
variances can be included in the statistical analysis. And in
the library searches. Two things of note from the figure: the
the case of the SERS analysis any “hot spots” or “dead
first is the high percentage match from the library search,
spots” can be averaged out so as not to skew the results.
and secondly, is how well the spectra of the two samples
match visually. Sample to sample reproducibility is very
important for any good analytical method.

Figure 6: Array Automation results for the silver treated ink on paper
SERS analysis

Figure 4: Showing the layout of microRNA samples, samples designated as

knowns are on the left side, and samples designated as unknowns are on the
right side, spots are labeled with the particular microRNA sample number
 In comparison to the untreated samples, the application Spectra for both sets of ink samples (SERS and regular
of the silver colloid to the ink samples showed a very strong Raman) were loaded into the TQ Analyst™ software.
increase in signal. Figure 7 illustrates the comparison of TQ Analyst is used for complex data analysis, particularly
the spectral averages of an untreated ink sample versus a large sets of data. It can be used for quantitative analysis,
SERS ink sample. The Raman signal for the untreated qualitative analysis, and large calibration sets. Data for
sample appears to be a nearly flat line when compared to ten of the ink samples were used for the chemometric
the SERS signal. All of the inks analyzed showed a similar analysis. The red ink was excluded as it required different
response, a result of a dual mechanism of fluorescence collection parameters than the other samples, and one of
reduction and signal enhancement. Figure 8 shows the the black inks was found to be the same formulation as one
average spectra for the treated and untreated red ink samples. of the other samples so it was also excluded. A qualitative
discriminant analysis method was constructed using the
spectra. For effective comparison of the SERS and regular
Raman data, the same parameters were used for both data
sets. Figure 9 shows a principal component score plot for
the untreated ink sample data, as can be seen there is not
a clear separation of the various samples. There were 750
of the 1690 spectra misclassified, which is approximately
44% of the spectra, a very significant problem. The SERS
spectra were analyzed using the same parameters, and a
plot of the resulting principal component scores is shown
in Figure 10. Now a much better separation of the samples
can be seen, though the two dimensional representation
does not adequately show the discrimination. There were
54 of the 1690 spectra misclassified, which is a much
better 3% result. The misclassified spectra are from three
black ink samples, which may have the same or very similar
formulation. More ink samples could be analyzed using
Figure 7: Spectral comparison of a Raman and SERS analysis of a black ink on paper, shown on the same sampling and analysis parameters so that an
the same intensity scale to illustrate the signal enhancement from SERS even larger and more comprehensive model could be
constructed for forensic ink analysis.

Figure 9: Principal component score plot for the untreated ink on paper
results (Raman). There were 750 misclassified spectra out of 1690, 44.38%
of the spectra

Figure 8: Spectral comparison of the Raman and SERS spectra of the red ink on paper

Figure 10: Principal component score plot for the silver treated ink on paper
results (SERS). There were 54 misclassified spectra out of 1690, only 3.20%
of the spectra
Automation Conclusion In addition to these
offices, Thermo Fisher
The ability to automate data collection is critical for any In conclusion, automated data collection from the combi-
Scientific maintains
type of high throughput analysis. In the past, SERS data nation of DXR Raman microscope with SERS substrates
a network of represen-
collection has typically been limited to single samples (colloids from a DXR/SERS Kit, home-made SERS colloids,
tative organizations
requiring hands-on work of the analyst to swap samples or SERS slides) and Array Automation software add-on
throughout the world.
or analyze new areas of the same sample. The method for OMNIC moves SERS forward from a single sample
presented in this note, using Array Automation, moves analysis method to an automated high throughput analytical
SERS forward as a technique towards the ultimate goal technique that has many potential future applications. The
of a fast, reliable, reproducible method. Array Automation work shown in this note demonstrates the ability of a user
brings many benefits to the analysis process. Since multiple to prepare up to 12 samples per slide, place the slide into
samples can be analyzed on one slide, less time is needed the instrument, collect from one up to 169 spectra per Africa-Other
for sample swapping. Part of the time savings also include sample, and then analyze that data with a variety of tools. +27 11 570 1840
sample preparation, as preparing 12 samples at one time is This can all be done with one instrument and one suite of Australia
+61 3 9757 4300
much more efficient than preparing 12 individual samples. software tools. Austria
Large amounts of data can be collected and analyzed with +43 1 333 50 34 0
one process and without constant user interaction, unlike References Belgium
+32 53 73 42 41
individual samples where each data point has to be indi- 1. Lee, P.C.; Meisel, D. J. Phys. Chem. 1982, 86, 3391.
Canada
vidually collected, then collated, and finally analyzed. As +1 800 530 8447
shown in this note more than 2000 spectra per slide could Further Information China
+86 10 8419 3588
be collected with one analysis run and then analyzed by For more information on SERS please see our Technical Note 51874
Denmark
various means. “Practical Applications of Surface-Enhanced Raman Scattering (SERS)”
+45 70 23 62 60
As can be seen in the results of the two experiments, We also have recorded webinars on a variety of SERS and Raman Europe-Other
applications that may be of interest. Please see this page: +43 1 333 50 34 0
SERS can be applied to different types of analysis, using http://www.thermoscientific.com/ramanwebinars Finland / Norway /
different SERS substrates. Array Automation allows for For more information on microRNA please see the Tech Review “MicroRNAs: Sweden
the collection of a large amount of data that can be used to Review of Discovery, Biogenesis, and Research Areas” which can be found here: +46 8 556 468 00
build spectral averages, for statistical models, or other types http://www.dharmacon.com/uploadedFiles/Home/Support_Center/Technical_ France
Reviews/microrna-tech-review.pdf +33 1 60 92 48 00
of data analysis. For microRNA analysis Array Automation Germany
could be used for the development of a diagnostic method +49 6103 408 1014
where samples are tested for the presence of a specific India
+91 22 6742 9434
microRNA related to a disease, and for forensic analysis Italy
Array can be used to collect ink data to the help in the +39 02 950 591
identification of different inks used on a forged document. Japan
+81 45 453 9100
Latin America
+1 561 688 8700
Middle East
+43 1 333 50 34 0
Netherlands
+31 76 579 55 55
New Zealand
+64 9 980 6700
Russia/CIS
+43 1 333 50 34 0
South Africa
+27 11 570 1840
Spain
+34 914 845 965
Switzerland
+41 61 716 77 00
UK
+44 1442 233555
USA
+1 800 532 4752

Thermo Electron Scientific

www.thermoscientific.com Instruments LLC, Madison, WI
USA is ISO Certified.
©2010 Thermo Fisher Scientific Inc. All rights reserved. All trademarks are the property of Thermo Fisher Scientific Inc. and its subsidiaries.
Specifications, terms and pricing are subject to change. Not all products are available in all countries. Please consult your local sales representative for details.
AN52020_E 12/10M

Part of Thermo Fisher Scientific