Multiplexed Spectral Fluorescence Detection

Multiplexed Spectral Fluorescence Detection

Dr. Shiou-jyh ‘Puck’ Ja

FLIR Systems

1024 S. Innovation Way

Stillwater, Oklahoma 74075 USA

Overview:

The concept of “multi-zone” SPCE-Fido® is first demonstrated with Zemax simulation and then verified with experiment results.  The goal of the multi-zone concept is to confine the emission from a specific reporter to a dedicated spatial partition within the annulus emission distribution so that the different emissions from multiple reporters can be spatially separated to avoid signal overlapping and cross-talking.  All the emissions from different reporters can then be completely retrieved individually without the interference from each other.

Through Zemax modeling, it was found that multi-zone concept would be possible by adjusting the imaging optics and the locations of reporter depositions.  There will not hardware change required from the original SPCE optics.  The preliminary experimental test data has been acquired and showed good agreement with the Zemax simulation data.  In a single reporter demonstration, the emission and 2D signature did present only at the specified zone with the expected explosive signature.  The details will be discussed in this disclosure.

Multi-zone SPCE-Fido®

It has been found that the SPCE light emitting from an off-centered report dot may be partially rejected by the imaging optics due to the vignetting effect.  Such effect was originally not desirable and was avoided in the past.  However, we have found an innovative way to take advantage of such effect for the multi-zone interrogation.  By carefully arranging the reporter dot off center and adjust lens elements, only a partial portion of the SPCE annulus will be captured by the detector, which leaves room to accommodate the SPCE light from other reporter dots in the array without any spectral overlapping at all in the total SPCE image.

Figure 1 shows the Zemax model of the SPCE imaging optics.  In the zoomed insert at the lower-left side of figure, the arrangement of the three reporter dots allocated along the vertical direction are shown.  At the center dot the emissions of different wavelengths are plotted by different colors, which also have different SPCE angles.  On the other contrary, all emission rays from the top and bottom dots are plotted with a same color: orange and olive green color, respectively.  The imaging lens was partially blocked since the prescription of this commercial component is proprietary.

1
The SPCE optics has been adjusted so that all the orange rays from the top reporter dot that are supposed to reach the upper portion of image plane are now rejected.  All the olive-green rays from the bottom dot are also rejected and not able to reach the bottom portion.  The rays at image plane are zoomed and plotted in another insert at the lower-right corner of the figure.   The purple, blue, cyan, and red rays from the center dot arrive both the top and bottom portion of image plane symmetrically as expected.  These color rays intercept with the orange and olive-green rays to indicate focusing.  There are no orange rays at the top portion of the image plane and no green ray at the lower portion, which indicate the vignetting effect successfully “mask” out a certain emission from certain dot so that only a partial portion of the SPCE annulus arrived at the detector.  All those rejected rays can be seen at the middle part of the imaging system where they propagating outside of imaging system.

single dot
4 dots

The Monte-Carlo simulation produced by the Zemax shows the net effect of such vignetting effect well and provides a good visual aid.  With a single 320-μm dot located at 700-μm off center, the simulated partial SPCE annulus image is shown at the top of Figure 2.  The positions of each wavelength bands (purple, blue, and green arc in the top plot) were still the same as to those from the centered dot.   However, only the lower arc of SPCE annulus was accepted by this adjusted SPCE imaging system and imaged by the detector.

Based on this single reporter result, it can be naturally postulated that four reporter dots may be accommodated to produce a full SPCE annulus image with minimized emission overlapping between the dots.  The lower plot of Figure 2 simulated with four reporter dots has clearly verified the idea.  The different emission color spectra from four emitters can be seen clearly at four different sections of the SPCE annulus.  The partitioning of the four different regions was distinct, which demonstrate the good separation of different emissions.

After the theoretical investigation, we have also conducted some experiments to verify the concept.  A single AFP dot was deposited at a spot about 700 μm off the center.  The acquired SPCE annulus image is shown at the plot (a) of Figure 3, which agrees very well with the Zemax simulation shown in Figure 2.   The intensity profiles along the SpecR direction from four different quadrants were shown in the plot (b) where the four quadrants were defined by the yellow dash lines on plot (a).  The white intensity profile corresponding to the first quadrant, where most of the SPCE light were gathered, was the only zone that has significant signal.  Other quadrants (zones) have little light due to the vignetting effect in the multi-zone configuration.

Figure 3.  SPCE-Fido® test data acquired with a single off-centered AFP dot using the multi-zone concept.
(a)
(b)
(c)
(d)
(e)
I
II
III
IV

The computed static signatures of the TNT sample from four zones are shown in the plot (c).  It also can be seen that only the first zone has significant quenching spectrum.  There was some leakage to the second zone (red trace in the plots) due to the slight orientation error.  Therefore, some small intensity profile and small quench signature can be observed in the second zone.  Such signal spillage can be easily fixed by rotating the sampling nozzle with the real-time image feedback during the data acquisition stage.

In addition, there was some short wavelength (small SpecR) component into Zone III (green trace), which produced small peak in the intensity profile plot (b) and small quenching at low SpecR region in the static quenching signature plot (c).   However, such signal was so small that it will be overcome by the dominant signal once those zone are occupied by other reporter dots.

The 2D temporal-spectral signature acquired from zone I is shown in plot (d) where the familiar TNT signature can be easily recognized.  On the other hand, the 2D signature from Zone IV has almost no signature to be recognized as expected.  Other zones with leakage have some low-level signature but again they will be masked out when those zones are occupied by other reporters.

Two-zone SPCE Data

Figure 4.  Multi-zone SPCE dots and their emission distribution: (a) locations of dots, (b) raw image (c)-(e) emission distribution analyzed with band-pass and long-pass filters.  The trace colors for channel 1 through 4 are white, green, red, and cyan respectively.
WuA
ch 2
WuA
(a)
(b)
(c)
(d)
(e)
ch 2
ch 1
ch 3
ch 4
460nm
490nm
520nm
ch1
ch3
ch1
ch4
ch4
ch1
ch2
AFP

The separation of optical emissions from different reporter dots was achieved by depositing the dot off-center by a distance between 400~700 μm.  There is a tradeoff on choosing the off-center distance.  Longer distance allows emission from each dot to be confined “more tightly”, which means that its SPCE light will only show up at its designated zone.  However, possibly smaller percentage of analyte would have a chance to interact with the reporter dot since the reporter dots are off-center and the analyte flows to all direction from the center.

In this batch of test, a 400-μm off-center distance was used.  The consequence of such short distance was that the emission from a single report would be less confined and would slightly occupy the opposite quadrant.  Majority of emission would cover one quadrant (Major Emission Channel) and small leakage in the opposite quadrant (Leakage Channel).  Since only two reporters were used in this batch of test, the leakage would show up in the spare quadrants and hence would not cause interference to each other.

The allocations of the reporter dots are shown in Figure 4(a) where AFP dot and Wu-AFP (WuA, a proprietary reporter) mixture dot were both about 400-μm away from the disk center, which can be verified by the fiducial camera on the Fujifilm Dimatix printer.  The image taken by the fiducial camera is also shown at the side.  The raw image of SPCE annulus image is shown in Figure 4(b) where four major partitions can be easily identified.  The channel 1 and 3 are the major emission and the leakage quadrants for AFP respectively.  The channel 4 and 2 are for WuA material.  From the raw image, it can be seen that channel 1 and 4 has fuller occupation in the quadrant and channel 2 and 3 only partial filling.  With longer off-center distance such partial filling in the leakage channel can be totally eliminated.

The spatial allocation of spectral content in the SPCE ring was unchanged in those major emission channels.  Therefore, the spectral-temporal SPCE signatures acquired from those major channels were similar to the ones collected in the past.  On the other hand, the spectral-spatial mapping in those leakage channels were compressed and shifted toward the shorter wavelength region.  Therefore, the spectral information from the leakage channel cannot be directly used to compare with previous data set.

In order to verify the spectral-spatial mapping relationship, band-pass and long-pass optical emission filters were used.  In the sub-figure (c), (d), and (e), the 460-nm band-pass, 490-nm band-pass, and 520-nm long-pass filters were used in the optical train. For the AFP emission, the major emission band was at channel 1.  By comparing the image in subplot (c) through (e), a clear SPCE arc at lower-left corner has apparent shifting from the inner side toward the outer edge of the annulus structure.  In the intensity versus radial distance (SpecR) diagram on the right, the 460-nm component in channel 1 was plotted as white trace with peak at SpecR=129 as it is shown in subplot (c).  Channel 3 (green trace) was the leakage channel from AFP dot, which had about half of the peak intensity and its peak was shifted to SpecR=120.  The 490-nm component in channel 1 was peaked at SpecR=145 as it is shown in subplot (d).  The leakage to channel 3 had much reduced intensity.  The AFP emission peak beyond 520-nm was further shifted to SpecR=161 in subplot (e), and its leakage to channel 3 has a smeared-out and much less apparent peak.  Hence, our discussion can focus only on the major emission channel.

For the emission from WuA dot, there was very little emission at 460-nm band in channel 4 (at the lower-right side of the annulus) as we can see in the image of the subplot (c).  We did observe some cross-channel interference from channel 1 at channel boundary, which will be addressed later in the discussion.  The WuA emission started to show up in the image of subplot (d) with 490-nm filter and became even stronger in subplot (e) with 520-nm filter.  In the diagram on the right side, the 490-nm component in channel 4 (cyan trace) has a peak at SpecR=114.  The major emission from WuA shows up at 520-nm band plot in channel 4 with SpecR=138.  The spectral-spatial relationship is summarized in Table 1.

Table 1 Allocations of spectral components from reporter dots

AFP WuA
460-nm component 129 (SpecR)
490-nm component 145 114
520-nm component 161 138

 

By comparing the spectral allocations listed in Table 1, it can be seen that for the same emission wavelength, the WuA emission has much shorter SpecR (or higher SPCE angle) than the AFP emission.  This is due to the fact that both the reporter deposition thickness and its refractive index constant play significant roles in the SPCE peak angular distribution.  Those factors move the WuA emission in the unfavorable direction so that the 520-nm emission from WuA material overlaps with the 470-nm emission from AFP material in angular space.  Fortunately, the multi-zone interrogation invention addresses this issue by separating those two emissions apart, and hence the overlapping interference can be minimized greatly.

Multi-zone SPCE-Fido® Test Results

AFP data

Figure 5.  Multi-zone SPCE detection data of TNT vapor with AFP (ch 1) and WuA (ch 4) dots.
(a)
(b)
(c)
(d)

The two-zone SPCE detection data of TNT vapor is shown in Figure 6.  The intensity profiles from four channels are plotted in plot (b) with white, red, green, and cyan traces representing channel 1 through 4.  The channle 1 (white) and channel 4 (cyan) were the major emission channels form the AFP and WuA dots.  The channel 3 (green) and channel 2 (red) was the leakage channels of the AFP and WuA dots.

The static spectral quench signatures shown at plot (a) clearly indicate strong quench by TNT vapor on channel 1 (AFP major) and 3(AFP leakage) as expected.  The deepest recorded quench was about 6.8% .  The spectral quench profile of ch. 3 (AFP leakage), however, has apparent distortion with deminished quench at the longer SprcR regeion.  Hence, the information on leakage channel is disregarded at this moment.

On the contrast, the WuA major emission channel (channel 4) has very little quench since WuA has much less sensitivity to TNT vapor comparing to AFP material.  The strong contrast on the quench levels beween ch. 1 and 4 proves the validity of optical isolation between channels.

The temporal-spectral maps of AFP (ch. 1) and WuA (ch4) are shown as plot (c) and (d).  The familiar TNT signature can be observed on the plot (c) which represents the AFP major channel.  On the other hand, there is only very weak signature (deepest quench was only 3%) at the WuA major channel shown in plot (d).

 PETN data

The PETN detection data are shown in Figure 6.  Test samples were produced by liquid dosing on the Teflon swipe and then desorbed using the Desorber.  The PETN static quenching signatures in plot (a) indicate strong quenches at WuA major emission channel (ch. 4, cyan) and leakage channel (ch. 2, red) and much less quenches at AFP channels.  Note that there might be some cross-talking from ch. 4 to ch. 3 (green, AFP leakage) and hence contributed some spurious quench in the AFP leakage channel.  The deepest quench at WuA channel was about 21% in plot (d), which was 4 times more sensitivity than the 5% quench presented at the AFP channel in plot (c).  The large quench level discrepancy again proves the validity of multi-zone concept.

Figure 6.  Multi-zone SPCE detection data of PETN (2 ng) with AFP (ch 1) and WuA (ch 4) dots.
PETN
(a)
(b)
(c)
(d)

The temporal-spectral maps from AFP major channel in plot (c) and WuA major channel in plot (d) are now both available without any overlapping which demonstrates the additional identification capability of the multi-zone configuration.  The PETN-with-AFP signature was an initial quench dip at about SpecR=110 and then shifted to SpecR=150, which was also consistent with the previous PETN signature acquired with single-channel AFP dot system.  The PETN-with-WuA signature has fewer features as we have learned from previous data.  It has an initial quench dip at about SpecR=125 and then slowly shifted to SpecR=140 at later time.  Again, such relatively monotonic spectral variation was consistent with previous data using single-channel WuA dot setup.

RDX data

Figure 7.  Multi-zone SPCE detection data of RDX (2 ng) vapor with AFP (ch 1) and WuA (ch 4) dots.
RDX
(a)
(b)
(c)
(d)

The RDX samples were prepared in the same way and the detection data are shown in Figure 7.  The static quench signatures shown in plot (a) is similar to the PETN curves shown in Figure 6 (a).   The deepest RDX quench in WuA channel was about 17% comparing to the 7% shown in the AFP channel.  The RDX sample was always tested after PETN sample, which may explain the slightly lower sensitivity on the WuA channel since the sensitivity of sensing material degrades slightly after each test.

The 2D temporal-spectral signature on AFP channel in plot (c) shows the familiar RDX-with-AFP signature, which was distinctly different from the PETN-with-AFP signature in Figure 6 (c).  The 2D signature on WuA channel shown in subplot (d) has less spectral variation but it still provided some feature to be distinct from the PETN data.  By comparing the plot (d) in both Figure 6 and Figure 7, the initial RDX-with-WuA quench dip appeared at shorter SpecR region comparing to PETN-with-WuA quench dip (SpecR=110 in Figure 7 vs. SpecR=125 in Figure 6).  The vertical dashed yellow lines are the absolute wavelength markers: they are 460-nm, 490-nm, and 520-nm markers from the AFP channel and 490-nm and 520-nm markers from the WuA channel.  The RDX 2D signature in WuA channel leaned more toward to the 490-nm lines in the beginning than PETN signature, and PETN-with-WuA 2D signature leaned more toward the 520-nm line.  Such distinction in WuA channel was repeatable.

We have successfully demonstrated multi-zone SPCE interrogation technique using up to three reporters in an explosive detection test conducted in Fort A. P. Hill.

Multiplexed Spectral Fluorescence Detection