On 2013 Dec 08, George McNamara commented:

Brief comments (pretty obvious I was not one of the reviewers):

• metal-dielectric coated is similar to metal enhanced fluorescence ... 164 papers in PubMed for http://www.ncbi.nlm.nih.gov/pubmed/?term="metal+enhanced+fluorescence" of which 28 mention distance http://www.ncbi.nlm.nih.gov/pubmed/?term="metal+enhanced+fluorescence"+distance (search on Dec 8, 2013). I was surprised at the lack of references to Lakowicz and colleagues, who have published extensively in this field: 109 references for metal enhanced fluorescence lakowicz

• The MEF/RDE (metal enhanced fluorescence / radiative decay engineering ... same thing, different acronyms) depends on DISTANCE and "intrinsic" quantum yield. Lakowicz MEF/RDE papers show biggest benefit for fluorophores with low intrinsic QY, like Rose bengal (and possibly flavins as in Fu). Elsayad used Alexa Fluor 488 (QY ~0.92 in water), so could have had a lot bigger enhancement factors by using Lakowicz-style fluorophores. Payoff: molecules diffusing in solution away from the silver islands would emit very little (though may generate copious oxygen radicals), so minimal background (re: Le Moal).

• for a 'spectrally" titled paper, their figures had almost free of spectral results ... their Fig 4c X-axis was 21 data points over 520-540 nm (21 data points is 1 nm interval, which does not make sense if this was Zeiss LSM710 Quasar (confocal), whose smallest interval is 1.9 nm - see last paragraph of methods). Also, would have been better to see more emission spectra range (ex. to 560 nm) for Alexa Fluor 488. .

• Fig 4cY-axis had a range of 0.95 to 1.15. Not impressive compared to Le Moal 2007 ( http://www.ncbi.nlm.nih.gov/pubmed/17172306 )

  we found the fluorescence signal emitted by a molecular cyanine 3 dye layer to be amplified by a factor ~30 when fluorophores are separated by a proper distance from the substrate. We then adapted our model to the case of homogeneously stained micrometer-sized objects and demonstrated mean signal amplification by a factor ~4. Finally, we applied our method to fluorescence imaging of dog kidney cells and verified experimentally the simulated results.

• another quantitative example: See page 12 and 14 of Fu ... Lakowicz - note Y-axes ( http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3087598/ ). They don't mention quantum yields of their fluorophores, flavin(s) are probably low (see next bullet).

• optics: Fig 1 shows objective lens from above. Text refers to 1.4 NA objective lens. No mention of coverglass on that side, or distance from that coverglass to the top of the cell (side away from the dichroic/silver islands surface). For live cells this would need to be aqueous. Staudt ... Hell 2007 ( http://www.ncbi.nlm.nih.gov/pubmed/17131355 , fig 6) a nice graph showing that an aqueous "gap" of 20 um results in collecting 20% of signal relative to at the coverglass (1.0 Airy unit confocal). So Elsayad's fig 4C would be compressed five fold.

Plan T: TIRF is also distance dependent. If their scheme is useful, I suggest could do plain glass surface with "dual view" TIRF with two emission bands (say 520-530 nm and 530-560 nm, with respect to their fig 4C and my knowing AF488's emission spectra), and simply ratio those ... See also Brunstein ... Oheim for combined TIRF + SAF benefits ( http://arxiv.org/ftp/arxiv/papers/1302/1302.1615.pdf ).

Plan V: Vutara SR-200 fluorescence nanoscope is rated as 20x20x50 nm precision localization (www.vutara.com, FPALM, also single molecule tracking of sparse non-blinking fluorophores - see also Zeiss PALm, Nikon STORM and Leica GSD, and many academic lab variants, including iPALM, ref 15 of Elsayad et al). However, this assumes a modest performance fluorophore. For easy math, a 100x more photons fluorophore would be a sqrt(100) = 10x improvement, so 2x2x5 nm, a lot better than these folks can do. [prices should be considered as "rough ballpark" for US market and without negotiations, end of year discounts, or post-warranty service contracts). Also SR-200 is around $300K vs Zeiss LSM710 is ~$600K nicely loaded (they used a 710). The 710's Quasar detector is a "standard" 32-channel PMT ... the LSM780 has a GaAsP version with ~2x higher QE, say $700K. See also Tables 1 and 2 of Ram et al 2008 ( http://www.ncbi.nlm.nih.gov/pubmed/18835896 ) ... multifocal imaging ~20 nm Z (precision localization) single QDot ... live time series, one objective lens (2 or 4 cameras).

Plan B: (or "issue B"): Bailey et al published standing wave excitation - in 1993 - that used 2 objective lenses (and 20 years earlier means earlier generation of fluorophores, objective lenses, filters, detector) that did pretty nicely ( http://www.ncbi.nlm.nih.gov/pubmed/?term=8232536 ). Too bad never offered commercially (since then carried on in 4pi, iPALM etc).

These "plans" are not the only alternatives possible and Elsayad et al do reference a few alternatives.

Ref 22: spectra web site reference of http://www.fluorophores.tugraz.at ... list at http://www.fluorophores.tugraz.at/substance/ ... login required. Around 90% of spectra are from PubSpectra, for example, AF488 http://www.fluorophores.tugraz.at/substance/618 "quantum yields" tab. I would have cited Carl Boswell and my PubSpectra dataset (open access, no registration, no legalese - data are not copyrightable) http://works.bepress.com/gmcnamara/9/ and http://www.ncbi.nlm.nih.gov/pubmed/?term=16969821

On 2013 Dec 08, George McNamara commented:

Brief comments (pretty obvious I was not one of the reviewers):

• metal-dielectric coated is similar to metal enhanced fluorescence ... 164 papers in PubMed for http://www.ncbi.nlm.nih.gov/pubmed/?term="metal+enhanced+fluorescence" of which 28 mention distance http://www.ncbi.nlm.nih.gov/pubmed/?term="metal+enhanced+fluorescence"+distance (search on Dec 8, 2013). I was surprised at the lack of references to Lakowicz and colleagues, who have published extensively in this field: 109 references for metal enhanced fluorescence lakowicz

• The MEF/RDE (metal enhanced fluorescence / radiative decay engineering ... same thing, different acronyms) depends on DISTANCE and "intrinsic" quantum yield. Lakowicz MEF/RDE papers show biggest benefit for fluorophores with low intrinsic QY, like Rose bengal (and possibly flavins as in Fu). Elsayad used Alexa Fluor 488 (QY ~0.92 in water), so could have had a lot bigger enhancement factors by using Lakowicz-style fluorophores. Payoff: molecules diffusing in solution away from the silver islands would emit very little (though may generate copious oxygen radicals), so minimal background (re: Le Moal).

• for a 'spectrally" titled paper, their figures had almost free of spectral results ... their Fig 4c X-axis was 21 data points over 520-540 nm (21 data points is 1 nm interval, which does not make sense if this was Zeiss LSM710 Quasar (confocal), whose smallest interval is 1.9 nm - see last paragraph of methods). Also, would have been better to see more emission spectra range (ex. to 560 nm) for Alexa Fluor 488. .

• Fig 4cY-axis had a range of 0.95 to 1.15. Not impressive compared to Le Moal 2007 ( http://www.ncbi.nlm.nih.gov/pubmed/17172306 )

  we found the fluorescence signal emitted by a molecular cyanine 3 dye layer to be amplified by a factor ~30 when fluorophores are separated by a proper distance from the substrate. We then adapted our model to the case of homogeneously stained micrometer-sized objects and demonstrated mean signal amplification by a factor ~4. Finally, we applied our method to fluorescence imaging of dog kidney cells and verified experimentally the simulated results.

• another quantitative example: See page 12 and 14 of Fu ... Lakowicz - note Y-axes ( http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3087598/ ). They don't mention quantum yields of their fluorophores, flavin(s) are probably low (see next bullet).

• optics: Fig 1 shows objective lens from above. Text refers to 1.4 NA objective lens. No mention of coverglass on that side, or distance from that coverglass to the top of the cell (side away from the dichroic/silver islands surface). For live cells this would need to be aqueous. Staudt ... Hell 2007 ( http://www.ncbi.nlm.nih.gov/pubmed/17131355 , fig 6) a nice graph showing that an aqueous "gap" of 20 um results in collecting 20% of signal relative to at the coverglass (1.0 Airy unit confocal). So Elsayad's fig 4C would be compressed five fold.

Plan T: TIRF is also distance dependent. If their scheme is useful, I suggest could do plain glass surface with "dual view" TIRF with two emission bands (say 520-530 nm and 530-560 nm, with respect to their fig 4C and my knowing AF488's emission spectra), and simply ratio those ... See also Brunstein ... Oheim for combined TIRF + SAF benefits ( http://arxiv.org/ftp/arxiv/papers/1302/1302.1615.pdf ).

Plan V: Vutara SR-200 fluorescence nanoscope is rated as 20x20x50 nm precision localization (www.vutara.com, FPALM, also single molecule tracking of sparse non-blinking fluorophores - see also Zeiss PALm, Nikon STORM and Leica GSD, and many academic lab variants, including iPALM, ref 15 of Elsayad et al). However, this assumes a modest performance fluorophore. For easy math, a 100x more photons fluorophore would be a sqrt(100) = 10x improvement, so 2x2x5 nm, a lot better than these folks can do. [prices should be considered as "rough ballpark" for US market and without negotiations, end of year discounts, or post-warranty service contracts). Also SR-200 is around $300K vs Zeiss LSM710 is ~$600K nicely loaded (they used a 710). The 710's Quasar detector is a "standard" 32-channel PMT ... the LSM780 has a GaAsP version with ~2x higher QE, say $700K. See also Tables 1 and 2 of Ram et al 2008 ( http://www.ncbi.nlm.nih.gov/pubmed/18835896 ) ... multifocal imaging ~20 nm Z (precision localization) single QDot ... live time series, one objective lens (2 or 4 cameras).

Plan B: (or "issue B"): Bailey et al published standing wave excitation - in 1993 - that used 2 objective lenses (and 20 years earlier means earlier generation of fluorophores, objective lenses, filters, detector) that did pretty nicely ( http://www.ncbi.nlm.nih.gov/pubmed/?term=8232536 ). Too bad never offered commercially (since then carried on in 4pi, iPALM etc).

These "plans" are not the only alternatives possible and Elsayad et al do reference a few alternatives.

Ref 22: spectra web site reference of http://www.fluorophores.tugraz.at ... list at http://www.fluorophores.tugraz.at/substance/ ... login required. Around 90% of spectra are from PubSpectra, for example, AF488 http://www.fluorophores.tugraz.at/substance/618 "quantum yields" tab. I would have cited Carl Boswell and my PubSpectra dataset (open access, no registration, no legalese - data are not copyrightable) http://works.bepress.com/gmcnamara/9/ and http://www.ncbi.nlm.nih.gov/pubmed/?term=16969821