Upload
truongdiep
View
219
Download
0
Embed Size (px)
Citation preview
PDF-OUTPUT
AU
TH
OR
’S P
RO
OF
Journal ID: 340, Article ID: 4730, Date: 2011-09-08, Proof No: 1
« APHB 340 layout: Large v.1.3.2 file: aphb4730.tex (ELE) class: spr-twocol-v1.3 v.2011/08/17 Prn:2011/09/07; 17:42 p. 1/14»« doctopic: OriginalPaper numbering style: ContentOnly reference style: aps»
Appl Phys BDOI 10.1007/s00340-011-4730-x
1 55
2 56
3 57
4 58
5 59
6 60
7 61
8 62
9 63
10 64
11 65
12 66
13 67
14 68
15 69
16 70
17 71
18 72
19 73
20 74
21 75
22 76
23 77
24 78
25 79
26 80
27 81
28 82
29 83
30 84
31 85
32 86
33 87
34 88
35 89
36 90
37 91
38 92
39 93
40 94
41 95
42 96
43 97
44 98
45 99
46 100
47 101
48 102
49 103
50 104
51 105
52 106
53 107
54 108
Optical nanofibers and spectroscopy
R. Garcia-Fernandez · W. Alt · F. Bruse · C. Dan ·K. Karapetyan · O. Rehband · A. Stiebeiner ·U. Wiedemann · D. Meschede · A. Rauschenbeutel
Received: 4 May 2011 / Revised version: 6 July 2011© Springer-Verlag 2011
Abstract We review our recent progress in the productionand characterization of tapered optical fibers with a sub-wavelength diameter waist. Such fibers exhibit a pronouncedevanescent field and are therefore a useful tool for highlysensitive evanescent wave spectroscopy of adsorbates on thefiber waist or of the medium surrounding. We use a carefullydesigned flame pulling process that allows us to realize pre-set fiber diameter profiles. In order to determine the waistdiameter and to verify the fiber profile, we employ scanningelectron microscope measurements and a novel accurate insitu optical method based on harmonic generation. We useour fibers for linear and nonlinear absorption and fluores-cence spectroscopy of surface-adsorbed organic moleculesand investigate their agglomeration dynamics. Furthermore,we apply our spectroscopic method to quantum dots on thesurface of the fiber waist and to cesium vapor surround-ing the fiber. Finally, toward dispersive measurements, wepresent our first results on building and testing a single-fiberbimodal interferometer.
R. Garcia-Fernandez · O. Rehband · A. Stiebeiner ·A. RauschenbeutelInstitut für Physik, Johannes Gutenberg-Universität Mainz,55099 Mainz, Germany
A. Rauschenbeutele-mail: [email protected]
R. Garcia-Fernandez · A. Stiebeiner · A. RauschenbeutelVienna Center for Quantum Science and Technology,Atominstitut, TU Wien, Stadionallee 2, 1020 Wien, Austria
W. Alt · F. Bruse · C. Dan · K. Karapetyan · U. Wiedemann ·D. Meschede (�)Institut für Angewandte Physik, Universität Bonn, Wegelerstr 8,53115 Bonn, Germanye-mail: [email protected]
1 Introduction
Glass fibers are nowadays well established as the backboneof the modern communication society. In a standard glassfiber used for telecommunications, the light is guided in-side the core and is isolated from the environment by thecladding, allowing low loss and low disturbance transmis-sion over long distances [1]. In contrast, optical nanofibers[2] with a diameter smaller than the wavelength of theguided light exhibit a strong lateral confinement of theguided mode in conjunction with a pronounced evanescentfield surrounding the fiber [3]. These properties make opticalnanofibers a unique tool for efficient and controlled couplingof light with matter on or near their surface. A wide range ofnanofiber applications are emerging in fields like, e.g., op-tical sensing [4–6], nanofiber-based evanescent wave spec-troscopy [7, 8], nonlinear optics [9–12], cold atom physics[13–17], probing of nanocavities and cavity quantum elec-trodynamics (cQED) [18–21], coupling to nanocavity lasers[22], or single quantum dot spectroscopy [23]. For these ap-plications, high-quality nanofibers with a high diameter uni-formity and low surface roughness are required, renderingthe technical realization a demanding process.
A standard fabrication method for optical nanofibers isflame pulling of commercial single mode optical fibers re-sulting in tapered optical fibers (TOFs) with a nanofiberwaist [24–29]. Efficient coupling of the light into and outof the nanofiber waist requires a careful design of the ta-per profile [30, 31]. Therefore, we first describe the pro-duction of nanofibers with our computer controlled fiberpulling rig as well as their main properties in Sect. 2. In or-der to verify the accuracy and precision of our productionprocess, we present scanning electron microscope measure-ments of the TOF profile and compare the experimental re-sults with the theoretically predicted target profile. Further,
AU
TH
OR
’S P
RO
OF
Journal ID: 340, Article ID: 4730, Date: 2011-09-08, Proof No: 1, UNCORRECTED PROOF
« APHB 340 layout: Large v.1.3.2 file: aphb4730.tex (ELE) class: spr-twocol-v1.3 v.2011/08/17 Prn:2011/09/07; 17:42 p. 2/14»« doctopic: OriginalPaper numbering style: ContentOnly reference style: aps»
R. Garcia-Fernandez et al.
109 163
110 164
111 165
112 166
113 167
114 168
115 169
116 170
117 171
118 172
119 173
120 174
121 175
122 176
123 177
124 178
125 179
126 180
127 181
128 182
129 183
130 184
131 185
132 186
133 187
134 188
135 189
136 190
137 191
138 192
139 193
140 194
141 195
142 196
143 197
144 198
145 199
146 200
147 201
148 202
149 203
150 204
151 205
152 206
153 207
154 208
155 209
156 210
157 211
158 212
159 213
160 214
161 215
162 216
Fig. 1 Schematic of the fiberpulling rig. A standard glassfiber is clamped on two stackedtranslation stages, which moveand stretch the glass fiber whichis heated with the flame of astationary gas burner. Thepulling procedure is performedunder clean room conditions ina laminar flow box
in Sect. 3, we present a method for measuring the diame-ter of the nanofiber waist, which is based on second andthird harmonic generation and has an accuracy of better than2% [12].
In Sect. 4 of this paper, we focus on applications ofnanofibers for surface spectroscopy. Fiber-based evanescentwave spectroscopy (EWS) offers stability, fast response, andapplications with in situ measurements. Motivated by theseadvantages, fiber-based EWS sensors have been demon-strated [32]. It is remarkable that the sensitivity of nanofiberspectroscopy depends strongly on the fiber diameter d andthe fiber length L: At resonance, the absorption probabilityper surface adsorbed particle with absorption cross sectionσ can be roughly estimated to be σ/λ2 because, in the fiberwaist of a silica nanofiber, the light field can be confinedto an effective cross section on the order of λ2, similar to astrongly focused light beam. In contrast to the situation infree space, this confinement is, however, maintained alongthe entire fiber length L, and hence the typical depth of fo-cus of a free space beam is enhanced by 3 to 5 orders ofmagnitude. As we show in detail in Sect. 4.2, the spectro-scopic sensitivity η of the nanofiber scales as
η(λ) ≈ ηfree(λ)πdL
λ2= ηfree(λ)
πL
λ, (1)
where we assumed d = λ and ηfree is the value for a freespace beam. Therefore, the strongly confined light in thenanofiber can be considered as an extremely elongated “in-finite” focus. We demonstrate the versatility of fiber-basedEWS by a series of absorption and fluorescence measure-ments performed on systems like surface-adsorbed organicmolecules or quantum dots. Further, fiber-based EWS is alsosuited for spectroscopic measurements of atoms surround-ing the nanofiber, as shown in Sect. 5 on the example ofcesium vapor.
Finally, we use nanofibers for the realization of an in-terferometric sensor based on the interference of two co-propagating transversal modes along the TOF [33, 34]. Thefirst experimental results are presented in Sect. 6. This tech-nique is highly sensitive to changes in the refractive index inthe vicinity of the nanofiber surface and therefore opens theroute toward dispersive measurements on surface adsorbates[35].
2 Fabrication and design of nanofibers
In the following, we summarize the fabrication procedure ofour nanofibers.
2.1 Description of the fiber pulling rig
All our experiments are performed using the nanofiber waistof tapered optical fibers (TOFs). We produce the TOFs bystretching a standard optical fiber while heating it with atraveling hydrogen/oxygen flame [36]. A schematic of thefiber pulling rig is shown in Fig. 1. It consists of two lineartranslation stages, the translator and the stretcher, which al-low sub-micrometer positioning, and is equipped with pre-cisely parallel aligned V-groove holders to which the fiberis clamped and tightened. A computer-controlled hydro-gen/oxygen burner, located between these two holders, pro-vides a clean flame. The flame heats a section of the fiberwith a length of approximately 1 mm to a temperature ofabout 1500°C, making it soft and malleable. The transla-tor moves the fiber back and forth above the flame whilethe stretcher simultaneously elongates it. Due to volumeconservation, the fiber diameter is thus gradually reduced.We use an analytic algorithm in order to calculate the tra-jectories of the translator and of the stretcher to obtain apredetermined fiber shape. This process allows us to fab-ricate TOFs with a homogeneous waist of diameters downto 100 nm and a typical length of 1–10 mm [28]. The trans-mission of light through the fiber is continuously monitoredduring the pulling process, yielding valuable information forquality control of the taper and its overall transmission. Formonochromatic light of 850 nm wavelength and a nanofiberwith a 500-nm diameter waist, our fabrication procedureyields up to 98.7% of the initial fiber transmission.
2.2 Properties of the guided light
Figure 2 shows a schematic of a TOF with three sections ex-hibiting different propagation properties: the nanofiber waist(a) surrounded by the two taper transitions (b) and the orig-inal unstretched fiber (c). The intensity distribution of thecorresponding guided mode is schematically represented byfilled curves.
AU
TH
OR
’S P
RO
OF
Journal ID: 340, Article ID: 4730, Date: 2011-09-08, Proof No: 1, UNCORRECTED PROOF
« APHB 340 layout: Large v.1.3.2 file: aphb4730.tex (ELE) class: spr-twocol-v1.3 v.2011/08/17 Prn:2011/09/07; 17:42 p. 3/14»« doctopic: OriginalPaper numbering style: ContentOnly reference style: aps»
Optical nanofibers and spectroscopy
217 271
218 272
219 273
220 274
221 275
222 276
223 277
224 278
225 279
226 280
227 281
228 282
229 283
230 284
231 285
232 286
233 287
234 288
235 289
236 290
237 291
238 292
239 293
240 294
241 295
242 296
243 297
244 298
245 299
246 300
247 301
248 302
249 303
250 304
251 305
252 306
253 307
254 308
255 309
256 310
257 311
258 312
259 313
260 314
261 315
262 316
263 317
264 318
265 319
266 320
267 321
268 322
269 323
270 324
Fig. 2 Schematic of a taperedoptical fiber (TOF) showing thelight propagation andconversion of the fundamentalmode (c) in the tapered region(b) to the nanofiber waist (a)and back again to the initialmode (c). The intensity profileof the guided mode isschematically represented by thefilled curves
In the unmodified part of the single mode TOF, thelight is guided inside the core via total internal reflectionat the core-cladding boundary. Due to the low refractive in-dex difference between core and cladding (ncore/ncladding ≈1.0035), this section is often designated as weakly guidingand Maxwell’s equations are approximately solved by lin-early polarized (LP) modes [37]. In the taper transition, bothcladding and core shrink. The decrease of the core diam-eter leads to a slight compression of the guided mode un-til the mode diameter reaches a minimum. Further reduc-tion of the core diameter leads to an expansion of the modeinto the cladding. The cladding then acts as the new guid-ing medium. In order to minimize transmission losses, thismode conversion should be adiabatic, i.e., there should beno coupling of the fundamental mode to higher transverseor radiative modes. This is achieved by maintaining shallowtaper angles of a few milliradians [30, 38].
Further decrease of the fiber diameter to the waist valuecauses the mode to be more and more laterally confined.The high refractive index difference at the cladding-air in-terface (ncladding/nair ≈ 1.5) allows a much tighter radialconfinement. Throughout the nanofiber waist, the diameterof the fiber remains constant in the sub-wavelength regime,and the guided mode exhibits tight radial confinement andstrong evanescent field. At the second taper transition, themode transformation is reversed. Note that through this pro-cess the weakly guided LP01 mode of the unstretched fiberis adiabatically transformed into the strongly guided HE11
mode of the nanofiber and back.The intensity distribution of a circularly or unpolarized
HE11 mode is plotted in Fig. 3 as a function of the distancefrom the fiber center. The intensity is maximal at the fibercenter and decreases toward the surface. Due to the high re-fractive index jump at the surface, the radial component ofthe electric field exhibits a correspondingly strong disconti-nuity, while the tangential components of the electric fieldremain continuous [3]. In the nanofiber, for visible wave-lengths the decay length of the evanescent field is typicallyon the order of few hundreds of nanometers. The strong in-tensity achieved at the fiber surface provides excellent con-ditions for high-sensitivity spectroscopic measurements andnonlinear optical experiments.
Fig. 3 Radial intensity distribution in units of 1/λ2 of the fundamentalmode HE11 guided through a nanofiber as a function of the distancefrom the fiber center in units of the wavelength. Calculated accordingto [3] for circularly or unpolarized light, a guided power of 10 nW,d/λ = 0.6 and a refractive index n = 1.46
Light propagation in a nanofiber is usually describedusing the concept of an effective refractive index neff =c/vphase = β/k0, where c is the vacuum speed of light, vphase
is the phase velocity of the mode, β is the mode propa-gation constant, and k0 = 2π/λ is the vacuum wavenum-ber [1, 37, 39, 40]. The propagation constant β is ob-tained by solving the transcendental equation arising fromthe Maxwell’s equations in conjunction with the bound-ary conditions applied to the nanofiber geometry [37]. Forthis purpose, we have created a Matlab toolbox that allowsus to automatically calculate all relevant parameters of thenanofiber guided modes [41]. As an example, Fig. 4 showsthe effective refractive indices of the lowest nanofiber modesas a function of the waist diameter. The single-mode andmultimode regime can be clearly distinguished in the fig-ure. The cutoff value at which only the fundamental HE11
mode propagates is wavelength-dependent and correspondsto d/λ ≈ 0.73 for a refractive index of nsilica = 1.45. Asthe fiber diameter decreases, a larger and larger fraction ofthe mode propagates outside of the glass, and the effec-tive refractive index of the mode decreases to unity. On theother hand, a fiber diameter increase leads to a mode guidedmostly entirely in the bulk and the effective refractive index
AU
TH
OR
’S P
RO
OF
Journal ID: 340, Article ID: 4730, Date: 2011-09-08, Proof No: 1, UNCORRECTED PROOF
« APHB 340 layout: Large v.1.3.2 file: aphb4730.tex (ELE) class: spr-twocol-v1.3 v.2011/08/17 Prn:2011/09/07; 17:42 p. 4/14»« doctopic: OriginalPaper numbering style: ContentOnly reference style: aps»
R. Garcia-Fernandez et al.
325 379
326 380
327 381
328 382
329 383
330 384
331 385
332 386
333 387
334 388
335 389
336 390
337 391
338 392
339 393
340 394
341 395
342 396
343 397
344 398
345 399
346 400
347 401
348 402
349 403
350 404
351 405
352 406
353 407
354 408
355 409
356 410
357 411
358 412
359 413
360 414
361 415
362 416
363 417
364 418
365 419
366 420
367 421
368 422
369 423
370 424
371 425
372 426
373 427
374 428
375 429
376 430
377 431
378 432
Fig. 4 Effective refractiveindex for the fundamental mode(leftmost) and a few higherorder modes as a function of thefiber diameter in units ofwavelength. The calculation wasperformed assuming a refractiveindex of nsilica = 1.45;reproduced from [12]
approaches nsilica. Throughout this paper, with the exceptionof Sect. 6, adiabatic tapers are used, i.e., to a good approx-imation the light is guided exclusively in the fundamentalHE11 mode.
3 Characterization of the diameter profile
The knowledge of the fiber diameter profile is essential fora quantitative understanding of the light propagation and ofnonlinear effects in nanofibers. In order to characterize thepredetermined diameter profile used by the fiber pulling rig,we performed scanning electron microscope (SEM) mea-surements. Although this method can offer an accuracy of3%, it is time-consuming and does not allow one to reuse theTOF because it has to be put onto a substrate and needs tobe gold-coated [12, 42]. Another method for measuring thenanofiber uniformity with a precision of 2–3% has been pro-posed [43, 44] that, however, gives no information about theabsolute fiber diameter. We have developed a nondestructiveoptical method for diameter measurements with an accuracybetter than 2%.
3.1 Scanning electron microscope measurement
Figure 5 shows the diameter profile of a TOF with a 320-nm diameter nanofiber waist of 1-mm length measured witha SEM. It exhibits a three-fold linear taper transition withlocal taper angles of Ω1 = 3 mrad, Ω2 = 1 mrad, andΩ3 = 5 mrad with the slope changing at the fiber diame-ters d1 ≈ 84 μm and d2 ≈ 40 μm. The solid line indicatesthe fiber diameter profile as predicted by our algorithm andthe inset shows a magnification of the submicron-diameterregion of the profile containing the waist with a length of1 mm. The nanofiber was mounted on a gold coated sub-strate for the SEM measurement (see [31] for technical de-tails). In order to determine the local fiber diameter, we suc-cessively took electron micrographs along the TOF with a
Fig. 5 Comparison between a SEM measurement (points) and the pre-dicted diameter profile (solid line) of a TOF with a nanofiber waist;reproduced from [31]
step size varying between 0.25 and 1 mm. The measureddiameter profile closely follows the predicted profile withdeviations smaller than 10%. For the two linear sectionswith local taper angles Ω1 and Ω2, they even fall below5%. From this excellent agreement, we conclude that ourproduction procedure for nanofibers indeed realizes the pre-dicted profiles to a very good approximation.
Figure 6 shows the detailed measurement of the waist re-gion of a nanofiber using a Zeiss SUPRA 55 field-emissionSEM. Since the manufacturer guaranties an accuracy of only±5%, we have recalibrated the SEM for each measurementseries and thereby achieved an improvement of the absoluteaccuracy to below ±2% (see [12] for further technical de-tails). Our fiber pulling rig produces a waist with less than±1.5% diameter variation, sufficiently uniform for the ex-periments reported here. At the same time, the pulling al-gorithm predicted a waist diameter of 393 nm while theSEM measurement yield a value of 415 nm. This accuracy ofabout 5% agrees well with the global nanofiber shape mea-surement shown in Fig. 5.
AU
TH
OR
’S P
RO
OF
Journal ID: 340, Article ID: 4730, Date: 2011-09-08, Proof No: 1, UNCORRECTED PROOF
« APHB 340 layout: Large v.1.3.2 file: aphb4730.tex (ELE) class: spr-twocol-v1.3 v.2011/08/17 Prn:2011/09/07; 17:42 p. 5/14»« doctopic: OriginalPaper numbering style: ContentOnly reference style: aps»
Optical nanofibers and spectroscopy
433 487
434 488
435 489
436 490
437 491
438 492
439 493
440 494
441 495
442 496
443 497
444 498
445 499
446 500
447 501
448 502
449 503
450 504
451 505
452 506
453 507
454 508
455 509
456 510
457 511
458 512
459 513
460 514
461 515
462 516
463 517
464 518
465 519
466 520
467 521
468 522
469 523
470 524
471 525
472 526
473 527
474 528
475 529
476 530
477 531
478 532
479 533
480 534
481 535
482 536
483 537
484 538
485 539
486 540
3.2 Optical measurement
The optical method for measuring the waist diameter isbased on second and third harmonic generation (SHG andTHG) in a bare silica nanofiber in air [45, 46]. Efficient har-monic generation requires that the phase-matching condi-tion is fulfilled, i.e., the effective refractive indices of thefundamental and harmonic waves have to be equal. Phase-matching for the two fundamental modes of both waves isdifficult to achieve due to material dispersion and the wave-length dependence of the effective indices of the modes in ananofiber (compare Fig. 4). However, the same effective re-fractive index dependence on the fiber diameter allows oneto fulfill the phase-matching condition if the harmonic lightpropagates in a higher transverse mode. If the wavelengthchanges, the effective refractive index curves shift and thephase-matching condition occurs at a different fiber diame-ter. There exists a one to one relation between the fiber di-ameter and the phase-matched wavelength for each mode.
Figure 7 shows a schematic of the experimental setup.Light from a tunable pulsed Ti:Sa laser (1 ps pulse duration,80 MHz repetition rate, 100 mW average power) is coupledinto one end of the TOF and its wavelength is scanned inorder to detect the SHG and THG peaks. A dichroic mirror atthe output of the TOF reflects the infrared fundamental light,which is measured with a power meter. The transmitted SHG
Fig. 6 SEM measurement of the nanofiber-waist region of a TOF. Theinset shows a vertically enlarged picture of the waist. Each point hasbeen measured with an accuracy of better than 7 nm; reproduced from[12]
and THG light is analyzed by a spectrometer (see [12] forfurther details).
Figure 8(a) shows the measured full response of the SHGand THG of a nanofiber. In Fig. 8(b), the phase-matchingcondition for harmonic generation is plotted. The nanofiberdiameter is determined from the position of the SHG andTHG peaks.
For the sample measured in Fig. 8, the fiber waist diam-eter is obtained as 425.7 nm with a resolution of 0.5 × 10−3
and an accuracy of <2%. Taking into account the trans-parency window of silica, our method allows to measurefiber waists with diameters down to 190 nm and up to 1 μm.
4 Applications in surface spectroscopy
This section demonstrates surface spectroscopy of differenttypes of nanofiber-adsorbed emitters. Further, we quantita-tively analyze the experimental results.
4.1 Experimental setup
Figure 9 shows the schematic experimental setup for fiber-based absorption and fluorescence spectroscopy. Depend-ing on the species under study, the particles are depositedonto the nanofiber waist using different approaches, e.g.,vapor deposition or immersion in a solution. The absorp-tion of the deposited particles is measured using a conven-tional absorption spectroscopy configuration with a tung-sten white light source and a spectrograph. Fluorescencespectra are obtained upon excitation by laser light cou-pled into the TOF after reflection from a glass plate. Thecounter-propagating fluorescence is collected after transmis-sion through the glass plate by the same spectrograph asfor the transmitted white light. This configuration avoids di-rect incidence of the excitation laser light into the spectro-graph. The residual backscattered and reflected excitationlaser light is suppressed by means of a razor edge filter,placed directly in front of the spectrograph’s entrance slit.
4.2 Theoretical considerations
In the following, we derive expressions for the measured ab-sorption and fluorescence signals to be expected from theset-up described above.
Fig. 7 Scheme of theexperimental setup for SHG andTHG measurement
AU
TH
OR
’S P
RO
OF
Journal ID: 340, Article ID: 4730, Date: 2011-09-08, Proof No: 1, UNCORRECTED PROOF
« APHB 340 layout: Large v.1.3.2 file: aphb4730.tex (ELE) class: spr-twocol-v1.3 v.2011/08/17 Prn:2011/09/07; 17:42 p. 6/14»« doctopic: OriginalPaper numbering style: ContentOnly reference style: aps»
R. Garcia-Fernandez et al.
541 595
542 596
543 597
544 598
545 599
546 600
547 601
548 602
549 603
550 604
551 605
552 606
553 607
554 608
555 609
556 610
557 611
558 612
559 613
560 614
561 615
562 616
563 617
564 618
565 619
566 620
567 621
568 622
569 623
570 624
571 625
572 626
573 627
574 628
575 629
576 630
577 631
578 632
579 633
580 634
581 635
582 636
583 637
584 638
585 639
586 640
587 641
588 642
589 643
590 644
591 645
592 646
593 647
594 648
Fig. 8 (a) The spectral response of a nanofiber for SHG (dashedline) and THG (solid line) plotted vs. the fundamental wavelength.The spectrometer response of the THG measurement is scaled up forvisibility. The four peaks correspond to phase matching of HE11(ω)
to the following modes (from left to right): HE21(2ω), HE12(3ω),
HE31(3ω), TM01(2ω). (b) Wavelength-dependent phase-matching di-ameter. Dashed lines: SHG, solid lines: THG, dotted lines: modes notobserved due to lack of effective mode overlap with the fundamentalmode. The horizontal line indicates the diameter of the investigatedsample determined by this method. Reproduced from [12]
Fig. 9 Scheme of theexperimental setup forabsorption and fluorescencespectroscopy
We consider a nanofiber with diameter d and length L
covered with n emitters, yielding a surface coverage, i.e.,number of emitters per surface area, of θ = n/πdL. Theabsorbance of light at a wavelength λ propagating in thefundamental mode of this nanofiber is given by η(λ) =−lg(Psig(λ)/Pref(λ)), where Psig(λ) and Pref(λ) are thetransmitted powers at the excitation wavelength λ in thepresence and absence of emitters, respectively. Accordingto [7], the absorbance η(λ) can be approximated as
η(λ) ≈ nσ(λ)
ln(10)Aeff(λ)= θσ (λ)
ln 10
πdL
Aeff(λ), (2)
if the absorption cross section of the emitter σ(λ) is muchsmaller than the effective area of the guided fiber modeAeff(λ) which is defined by
Aeff(λ) = Pref(λ)
Isurf(λ). (3)
Isurf(λ) denotes the intensity of the evanescent field compo-nents at the fiber surface that actually couple to the emit-ters. Isurf(λ) does account for a possible preferential orien-tation of the transition dipole of the emitter which might oc-cur, e.g., when planar molecules lie flat at the fiber surface
[47]. In this case, their transition dipole moment is orientedperpendicular to the radial electric field component of theguided fiber mode. Therefore, only the axial and azimuthalcomponents of the electric field, calculated according to [3],contribute to Isurf(λ).
Equation (2) can be rewritten as a function of the ab-sorbance ηfree(λ) = θσ (λ)/ ln 10 of a freely propagatingbeam [7]
η(λ) = ηfree(λ)ξ(λ), (4)
where the factor ξ(λ) = πdL/Aeff gives the enhancement ofthe sensitivity of the fiber-based absorption measurement.With a length of the nanofiber L on the order of millime-ters and a diameter d in the submicron range, we thus ex-pect an increase by four orders of magnitude in the sensi-tivity in comparison with methods using freely propagatingbeams [7].
The absorbance is directly proportional to the number ofemitters n and inversely proportional to the effective areaAeff. As can be seen from the second term in (2), the quan-tity 1/Aeff determines the sensitivity for a given numberof emitters. It is plotted in Fig. 10 in units of 1/λ2 as afunction of the fiber diameter and reaches a maximum for
AU
TH
OR
’S P
RO
OF
Journal ID: 340, Article ID: 4730, Date: 2011-09-08, Proof No: 1, UNCORRECTED PROOF
« APHB 340 layout: Large v.1.3.2 file: aphb4730.tex (ELE) class: spr-twocol-v1.3 v.2011/08/17 Prn:2011/09/07; 17:42 p. 7/14»« doctopic: OriginalPaper numbering style: ContentOnly reference style: aps»
Optical nanofibers and spectroscopy
649 703
650 704
651 705
652 706
653 707
654 708
655 709
656 710
657 711
658 712
659 713
660 714
661 715
662 716
663 717
664 718
665 719
666 720
667 721
668 722
669 723
670 724
671 725
672 726
673 727
674 728
675 729
676 730
677 731
678 732
679 733
680 734
681 735
682 736
683 737
684 738
685 739
686 740
687 741
688 742
689 743
690 744
691 745
692 746
693 747
694 748
695 749
696 750
697 751
698 752
699 753
700 754
701 755
702 756
Fig. 10 Plot of 1/Aeff andd/Aeff in units of 1/λ2 and 1/λ,respectively, as a function of thefiber diameter d in units of λ.Aeff is calculated taking intoaccount all three components ofthe electric field. Assuming arefractive index of 1.46 andignoring material dispersion,both plots hold universally forany λ
dmax = 0.508 · λ. It rapidly goes to zero for small diametersbecause then the mode is only weakly bound and Aeff di-verges. The third term in (2) shows that for spectroscopicmeasurements of thin layers, the quantity d/Aeff shouldbe maximized. This quantity is also plotted in Fig. 10 inunits of 1/λ as a function of d/λ, reaching a maximum fordmax = 0.584 · λ.
In order to derive an expression for the fluorescence sig-nal, we consider an emitter at position z along the fiberwaist that absorbs the fraction σ(λexc)/Aeff(λexc) of thepower Psig(λexc, z) of the excitation wavelength λexc. Forthe fiber diameter and the wavelengths considered in our ex-periments, about 20% of the fluorescence of a single dipoleemitter on the fiber surface is expected to be coupled backinto the fundamental guided mode of the fiber, 10% in eachdirection of the fiber [48]. As a result, the fluorescence emit-ted into the fiber mode by an emitter at a position z along thefiber waist is proportional to the absorbed power:
Pfl(λ, z) = C(λ)Pabs(λexc, z), (5)
where the proportionality factor C(λ) ∝ q(λ)/Aeff(λ) in-cludes the wavelength dependent fluorescence quantumyield of the emitter q(λ) and the average fractional emissionof the emitter into the guided fiber mode. The latter is an av-erage over all possible orientations of the emitter on the fibersurface. Furthermore, it is determined by the intensity of theevanescent field at the fiber surface which is proportionalto 1/Aeff; see (3). Note that the spectral overlap betweenthe absorption and emission spectra in our system results ina partial self-absorption of the emitted fluorescence by cir-cumjacent emitters before it reaches the output of the fiber.While the high sensitivity of our method allows us to per-form measurements in a regime of low surface coverages,self-absorption is nonnegligible at higher surface coverages.Detailed theoretical considerations in order to quantify thiseffect are presented in [8].
The expressions derived for the absorbance and fluores-cence in (2) and (5), respectively, are proportional to the
wavelength dependent coupling to the fiber mode charac-terized by 1/Aeff. The absorption and fluorescence spectrapresented throughout this paper have been corrected accord-ingly.
4.3 Organic molecules: linear excitation
As a model system, we use 3,4,9,10-perylene-tetracarbox-ylic dianhydride molecules (PTCDA) because of their sta-bility under evaporation at ambient conditions, their highquantum yield and the experimental and theoretical knowl-edge concerning their spectral characteristics [49]. More-over, PTCDA molecules significantly change their spectralproperties depending on their arrangement on the surface[50]. Therefore, these organic molecules are well suited forsensitivity studies.
The measurements were performed with a 320-nm di-ameter nanofiber with a length of 1 mm fabricated from aNufern 460-HP fiber. This diameter ensures that only thefundamental mode is guided in the fiber waist for wave-lengths longer than 450 nm, thereby matching the singlemode cutoff wavelength of the unprocessed fiber. Moreover,this diameter yields the maximum sensitivity for absorptionspectroscopy in the visible domain, as discussed in Sect. 4.2.The TOF used in this experiment has a transmission of up to70% in the wavelength range between 450 and 650 nm. Themolecules are deposited on the nanofiber waist by placing acrucible with PTCDA crystals below the fiber and by heat-ing it to 325°C. By convection, sublimated molecules arecarried to the nanofiber waist where they are adsorbed. Inter-laced measurements of absorption and fluorescence spectraare performed in order to determine the respective surfacecoverage (see (2)). The fluorescence spectra are obtainedupon excitation with a freely running diode laser at a wave-length of 406 nm. The background signal induced by the ex-citation light in the fiber has been subtracted in the spectra.For further technical details, see [8].
Figure 11 displays a series of in situ fluorescence and ab-sorption spectra recorded with about 8 μW laser power and
AU
TH
OR
’S P
RO
OF
Journal ID: 340, Article ID: 4730, Date: 2011-09-08, Proof No: 1, UNCORRECTED PROOF
« APHB 340 layout: Large v.1.3.2 file: aphb4730.tex (ELE) class: spr-twocol-v1.3 v.2011/08/17 Prn:2011/09/07; 17:42 p. 8/14»« doctopic: OriginalPaper numbering style: ContentOnly reference style: aps»
R. Garcia-Fernandez et al.
757 811
758 812
759 813
760 814
761 815
762 816
763 817
764 818
765 819
766 820
767 821
768 822
769 823
770 824
771 825
772 826
773 827
774 828
775 829
776 830
777 831
778 832
779 833
780 834
781 835
782 836
783 837
784 838
785 839
786 840
787 841
788 842
789 843
790 844
791 845
792 846
793 847
794 848
795 849
796 850
797 851
798 852
799 853
800 854
801 855
802 856
803 857
804 858
805 859
806 860
807 861
808 862
809 863
810 864
Fig. 11 Fluorescence (left) andcorresponding absorption (right)spectra of surface-adsorbedPTCDA molecules duringdeposition. Five representativespectra within a range of surfacecoverages between 0.10 and1.28% of a ML are shown.Reproduced from [8]
about 3 nW white light power respectively. These powersare low enough not to saturate the molecules. For clarity,we chose five representative spectra taken during depositionof the molecules for different surface coverages. The shownabsorption spectra are the mean of two consecutive spec-tra, one recorded 420 ms before and the other 420 ms afterthe corresponding fluorescence spectrum. Within this timeinterval, the deposition rate remains roughly constant, thusenabling us to determine the surface coverages underlyingthe measured fluorescence spectra by this averaging process.Using σ = 4.1 × 10−16 cm2 [8] in (2), we infer from the ab-sorbance values ranging from 0.04 to 0.51 at the absorptionmaximum at 2.47 eV that 0.8 ×106 to 10.6×106 moleculescover the fiber waist. This corresponds to a surface cover-age of 0.8 × 1011 cm−2 to 10.6 × 1011 cm−2 or 0.10% to1.28% of a compact monolayer (ML) of flat lying PTCDAmolecules, arranged in the herringbone structure of the (102)plane of the PTCDA bulk crystal [51]. The absorption spec-tra show a clear vibronic progression and agree well withdifferential reflectance spectroscopy (DRS) spectra of sub-monolayers of PTCDA on mica [50]. The fluorescence spec-tra also show a vibronic progression as known for PTCDA insolution [52]. Note, that the absolute peak positions in both,absorption and fluorescence spectra, are slightly shifted incomparison with the literature due to the interaction of theadsorbed molecules with the fiber surface.
The mirror symmetry between absorption and fluores-cence spectra that is characteristic for numerous molecularsystems is retrieved after correction for the self-absorptioneffect. Figure 12 displays two of the spectra from Fig. 11for about 0.75% and 0.20% of a ML and the ones correctedfor self-absorption and further reemission of the reabsorbedfluorescence (see [8] for further details). Note that, althoughfor higher surface coverages like 0.75% of a ML the self-absorption changes the fluorescence spectra significantly,this effect is negligible for lower surfaces coverages like0.20% of a ML. Summarizing, our sensitivity is high enough
to perform studies for low surface coverages. At the sametime, the fact that we are able to correct for self-absorptionmakes our method applicable for a large range of surfacecoverages.
Our method yields absorption and fluorescence spectrawith a good signal to noise ratio for down to ∼106 PTCDAmolecules at room temperature. Under these conditions,thermal effects lead to strongly broadened molecular transi-tions with a FWHM of ∼850 cm−1 or ∼25 THz [53]. Whencooling the fiber to cryogenic temperatures below 4 K, thisthermal broadening could in principle be eliminated and 106
times narrower lines of a few 10 MHz should be obtainablewith organic dye molecules embedded in organic crystals[54]. Under these circumstances, our method would thusalso allow us to carry out spectroscopic studies on singlemolecules.
The high sensitivity of our method also allowed us tostudy the agglomeration dynamics of the adsorbed mole-cules on a second to minutes time scale at ambient con-ditions. Figure 13(a) shows absorption spectra measuredduring the ripening of the film after the molecular depo-sition has been stopped. We observe a continuous changeof the shape of the spectra, evolving from monomer-liketo oligomer-like. At the beginning of this process (up to23 min), the absorbance remains constant at the energies ofabout 2.34 eV and 2.67 eV. These so-called isosbestic pointsresult from the fact that, at these energies, the monomer andthe dimer phase of PTCDA have the same molar absorp-tivity. This observation thus indicates that the total num-ber of molecules remains nearly constant during the ripen-ing. Additional kinetic details of the ripening process canbe found in [7]. Note that reordering of PTCDA moleculeson a mica surface at ambient condition has already beenobserved. However, the measurement was not resolved intime [55]. The fast agglomeration process is attributed tothe presence of water at the surface that increases the mo-bility of the surface-adsorbed molecules and thus acceler-
AU
TH
OR
’S P
RO
OF
Journal ID: 340, Article ID: 4730, Date: 2011-09-08, Proof No: 1, UNCORRECTED PROOF
« APHB 340 layout: Large v.1.3.2 file: aphb4730.tex (ELE) class: spr-twocol-v1.3 v.2011/08/17 Prn:2011/09/07; 17:42 p. 9/14»« doctopic: OriginalPaper numbering style: ContentOnly reference style: aps»
Optical nanofibers and spectroscopy
865 919
866 920
867 921
868 922
869 923
870 924
871 925
872 926
873 927
874 928
875 929
876 930
877 931
878 932
879 933
880 934
881 935
882 936
883 937
884 938
885 939
886 940
887 941
888 942
889 943
890 944
891 945
892 946
893 947
894 948
895 949
896 950
897 951
898 952
899 953
900 954
901 955
902 956
903 957
904 958
905 959
906 960
907 961
908 962
909 963
910 964
911 965
912 966
913 967
914 968
915 969
916 970
917 971
918 972
Fig. 12 Influence ofself-absorption and reemissionon the measured fluorescencesignal for 0.20% (a) and 0.75%(b) of a ML. For 0.75% of aML, the signal corrected forself-absorption differssignificantly from the measuredspectrum. For 0.20% of a ML,the difference is only marginal.The correction for reemissiondoes not have a significantinfluence on the spectral shapefor neither of the two surfacecoverages. The absorbance forboth surface coverages (dashedline) is shown for comparison ofthe spectral shape
ates their reordering. In order to desorb pollutants from thefiber surface, we set up an ultrahigh vacuum (UHV) cham-ber, allowing us to work with a more well-defined surface.Figure 13(b) displays the absorption spectra measured dur-ing the agglomeration process in the UHV environment.A comparison with the analogue process at ambient con-ditions (Fig. 13(a)) shows qualitatively the same agglomer-ation dynamics. However, at UHV conditions, the processis slowed down by about two orders of magnitude, therebyshowing that the mobility of the molecules in a vacuum en-vironment is significantly reduced in comparison with am-bient conditions. Both measurements have been performedwith the same nanofiber diameter and the molecular deposi-tion has been stopped at about the same surface coverage.
4.4 Organic molecules: two-photon excitation
The high light intensity and long interaction length (“the in-finite focus”) makes nanofibers a suitable platform for non-linear optical experiments. Here, we present two-photon ex-cited fluorescence (TPEF) measurements of organic mole-cules. The theory presented in Sect. 4.2 holds for linear ex-citation and it has to be slightly modified for two-photonexcitation. In particular, an emitter at position z along thefiber waist absorbs now a fraction σ2Ph(λexc)/Aeff(λexc) of
the two-photon power P 2sig(λexc, z) of the excitation wave-
length λexc, where σ2Ph(λexc) denotes the two-photon ab-sorption cross section of the emitter.
We use the fluorescent dye Rhodamine 6G (Rh6G),which exhibits a high fluorescence quantum yield [56]. Un-like PTCDA, most organic dye molecules undergo ther-mal decomposition by evaporation at atmospheric pres-sure. In order to extend the fiber surface spectroscopy to alarger variety of organic molecules, we developed a simpleapproach—the “drip method”. The molecules are dissolvedin a spectroscopic-grade solvent and a drop of this solutionis dripped onto the nanofiber waist using a pipette. A thinfilm of the solution covers the fiber surface. Subsequently,the solvent evaporates and the molecules remain adsorbed atthe fiber surface.
With slight modifications, the experimental setup corre-sponds to the one shown in Fig. 9. For excitation, we use apulsed Ti:Sa laser at 870 nm wavelength with a pulse du-ration of 1 ps, a repetition rate of 80 MHz and a 104-foldhigher peak intensity compared to the corresponding contin-uous wave intensity. Further, in order to prevent laser powerlosses, the glass plate has been exchanged by a dichroic mir-ror that separates the infrared light of the Ti:Sa laser from theemitted visible light from the TPEF measurement. We use a
AU
TH
OR
’S P
RO
OF
Journal ID: 340, Article ID: 4730, Date: 2011-09-08, Proof No: 1, UNCORRECTED PROOF
« APHB 340 layout: Large v.1.3.2 file: aphb4730.tex (ELE) class: spr-twocol-v1.3 v.2011/08/17 Prn:2011/09/07; 17:42 p. 10/14»« doctopic: OriginalPaper numbering style: ContentOnly reference style: aps»
R. Garcia-Fernandez et al.
973 1027
974 1028
975 1029
976 1030
977 1031
978 1032
979 1033
980 1034
981 1035
982 1036
983 1037
984 1038
985 1039
986 1040
987 1041
988 1042
989 1043
990 1044
991 1045
992 1046
993 1047
994 1048
995 1049
996 1050
997 1051
998 1052
999 1053
1000 1054
1001 1055
1002 1056
1003 1057
1004 1058
1005 1059
1006 1060
1007 1061
1008 1062
1009 1063
1010 1064
1011 1065
1012 1066
1013 1067
1014 1068
1015 1069
1016 1070
1017 1071
1018 1072
1019 1073
1020 1074
1021 1075
1022 1076
1023 1077
1024 1078
1025 1079
1026 1080
450-nm diameter nanofiber with a length of 5 mm fabricatedfrom a Fibercore SM800 fiber.
The molecules have been deposited with a 0.016% massconcentration solution of Rh6G in ethanol. The correspond-ing absorption spectrum of Rh6G adsorbed on the nanofiber
Fig. 13 Time evolution of the spectral absorption of a constant num-ber of molecules at ambient (a) and UHV (b) conditions, showing theagglomeration dynamics of the molecules on the fiber surface
is shown in Fig. 14. Using the peak absorbance value of1.47 and the absorption cross section of 6.7 × 10−16 cm2
at 2.37 eV (calculated according to [8, 23] and [57]), weestimate the number of Rh6G molecules adsorbed on thenanofiber waist to be 4 × 107. With the surface density of aML of Rh6G molecules adsorbed on silica of 5×1013 cm−2
[58], this corresponds to a surface coverage of 1.1% of aML. By exciting the molecules with an average power of20 mW of pulsed light, we detect the TPEF spectrum alsoplotted in Fig. 14. It shows the typical mirror-image relationbetween the absorbance and fluorescence spectra [59]. Theplotted fluorescence spectrum has been already corrected forthe self-absorption effect. Both absorption and fluorescencespectra, agree qualitatively well with the measurements per-formed in solution [57]. However, there is a shift in the abso-lute peak positions which can be attributed to the interactionof the molecules with the fiber surface.
The TPEF spectrum is obtained directly after startingthe excitation process since the TPEF intensity decreasesrapidly due to photobleaching of the molecules. In fact, ina separate measurement, the temporal behavior of the TPEFstrength has been recorded by replacing the spectrometerwith a photomultiplier. The result is shown in the inset ofFig. 14. Starting with an initial value of 2 × 107 adsorbedRh6G molecules, the TPEF vanishes within seconds. At thispoint, the absorbance is also zero and, therefore, no undam-aged molecules remain adsorbed at the fiber surface.
4.5 Quantum dot spectroscopy
We have further investigated whether the spectroscopicmethod used for organic molecules can be also applied toquantum dots. Quantum dots are nanometer-scale semicon-ductor particles with a high fluorescence quantum yield, amuch better resistance to photobleaching in comparison to
Fig. 14 Two-photon excitedfluorescence (solid line) andabsorption (dashed line)spectrum of 4 × 107 Rhodamine6G molecules (1.1% of a ML)adsorbed on a nanofiber. Inset:An independent measurement ofthe time evolution of the TPEFresponse, showingphotobleaching of 2 × 107
adsorbed Rhodamine 6Gmolecules (0.6% of a ML)
AU
TH
OR
’S P
RO
OF
Journal ID: 340, Article ID: 4730, Date: 2011-09-08, Proof No: 1, UNCORRECTED PROOF
« APHB 340 layout: Large v.1.3.2 file: aphb4730.tex (ELE) class: spr-twocol-v1.3 v.2011/08/17 Prn:2011/09/07; 17:42 p. 11/14»« doctopic: OriginalPaper numbering style: ContentOnly reference style: aps»
Optical nanofibers and spectroscopy
1081 1135
1082 1136
1083 1137
1084 1138
1085 1139
1086 1140
1087 1141
1088 1142
1089 1143
1090 1144
1091 1145
1092 1146
1093 1147
1094 1148
1095 1149
1096 1150
1097 1151
1098 1152
1099 1153
1100 1154
1101 1155
1102 1156
1103 1157
1104 1158
1105 1159
1106 1160
1107 1161
1108 1162
1109 1163
1110 1164
1111 1165
1112 1166
1113 1167
1114 1168
1115 1169
1116 1170
1117 1171
1118 1172
1119 1173
1120 1174
1121 1175
1122 1176
1123 1177
1124 1178
1125 1179
1126 1180
1127 1181
1128 1182
1129 1183
1130 1184
1131 1185
1132 1186
1133 1187
1134 1188
Fig. 15 Fluorescence (solidline) and absorption (dashedline) spectrum of 3 × 105 CdSequantum dots (0.3% of a ML)adsorbed on an nanofiber
organic molecules and a fluorescence emission wavelengththat is size-dependent.
The experimental setup shown in Fig. 9 has been used.For excitation, we use a continuous-wave diode laser at awavelength of 405 nm. The measurement was performedusing a 450-nm diameter nanofiber with a waist length of5 mm fabricated from a Fibercore SM800 fiber.
A solution of CdSe quantum dots (PL-QD-O-570 fromPlasmachem GmbH, core diameter 3.3 nm, total diameter8.9 nm) dissolved in heptane with a mass concentration of0.036% was deposited by the drip method on the nanofiberwaist. The absorption spectrum is shown in Fig. 15. Fromthe absorbance value of 0.043 and the absorption cross sec-tion of 8 × 10−16 cm2 at 2.22 eV (calculated from the molarextinction coefficient of 2.1 × 105 M−1 cm−1 provided bythe manufacturer), we estimate that about 3 × 105 quantumdots were adsorbed on the nanofiber waist. The correspond-ing surface coverage of 0.3% of a ML is calculated by as-suming that the spherical quantum dots arrange in one fullmonolayer in the geometrically most compact form. The flu-orescence spectrum, obtained by exciting the quantum dotswith 36 nW of laser light at 405 nm, is also shown in Fig. 15.The fluorescence maximum at 2.15 eV differs slightly fromthe manufacturer specifications of (2.175 ± 0.020) eV forquantum dots dissolved in chloroform. This discrepancycould be attributed to the interaction of the quantum dotswith the nanofiber surface. Note that, for an absorbanceas low as 0.043, the effect of self-absorption is negligi-ble.
5 Cesium vapor spectroscopy
The strong evanescent field surrounding the nanofiber waistof a TOF allows light-matter interaction not only withsurface-adsorbed species but also with media surrounding
the fiber. This feature allowed us to perform hot cesium va-por spectroscopy with a nanofiber.
The experimental setup consists of a tunable single-modediode laser, a nanofiber in a vacuum chamber, and a siliconphotodiode as a detector. The chamber contains a reservoirof liquid cesium, which is heated to 70°C, thereby providinga cesium vapor pressure of approximately 10−4 mbar duringthe experiment. The rest of the chamber is heated to 80°C inorder to prevent cesium condensation anywhere outside thereservoir. The measurement was performed with a 450-nmdiameter nanofiber with a length of 5 mm fabricated from aFibercore SM800 fiber.
Figure 16 shows the measured transmission through thenanofiber for an input laser power of 20 nW while scan-ning around the cesium D2 resonance line at a wavelengthof 852.3 nm. The data has been fitted using a function whichis the sum of three Voigt profiles corresponding to the threehyperfine transitions (F = 3 to F ′ = 2, 3, 4) of the D2 line:
A(ν) =4∑
F ′=2
aF ′∫ ∞
−∞exp(−x2/(2σ 2))
σ√
2π
× 2γ
π((ν − νF ′ − x)2 + γ 2/4)dx, (6)
where ν is the frequency, ν3 = 351.7 THz the frequency ofthe 3 → 3 transition, ν4 − ν3 = 201.3 MHz and ν3 − ν2 =151.2 MHz the separations of the hyperfine cesium lev-els [60], γ and σ the widths of the Lorentzian and Gaus-sian components of the Voigt curve. The hyperfine transi-tion strength factors for the 3 → 2, 3 → 3, 3 → 4 transi-tions, have the values a2 = 5/14, a3 = 3/8, and a4 = 15/56,respectively [60].
The fit yields the values σ = 178 MHz and γ = 130 MHz.The Gaussian width agrees well with the theoretically ex-
AU
TH
OR
’S P
RO
OF
Journal ID: 340, Article ID: 4730, Date: 2011-09-08, Proof No: 1, UNCORRECTED PROOF
« APHB 340 layout: Large v.1.3.2 file: aphb4730.tex (ELE) class: spr-twocol-v1.3 v.2011/08/17 Prn:2011/09/07; 17:42 p. 12/14»« doctopic: OriginalPaper numbering style: ContentOnly reference style: aps»
R. Garcia-Fernandez et al.
1189 1243
1190 1244
1191 1245
1192 1246
1193 1247
1194 1248
1195 1249
1196 1250
1197 1251
1198 1252
1199 1253
1200 1254
1201 1255
1202 1256
1203 1257
1204 1258
1205 1259
1206 1260
1207 1261
1208 1262
1209 1263
1210 1264
1211 1265
1212 1266
1213 1267
1214 1268
1215 1269
1216 1270
1217 1271
1218 1272
1219 1273
1220 1274
1221 1275
1222 1276
1223 1277
1224 1278
1225 1279
1226 1280
1227 1281
1228 1282
1229 1283
1230 1284
1231 1285
1232 1286
1233 1287
1234 1288
1235 1289
1236 1290
1237 1291
1238 1292
1239 1293
1240 1294
1241 1295
1242 1296
Fig. 16 Measured (points) and fitted (solid line) cesium absorptionspectra
pected value for Doppler broadening at a temperature of80°C:
σtheoretical = ν0
√kBT
mc2= 174 MHz, (7)
where ν0 = 351.73 THz is the frequency of the D2 line, kB
the Boltzmann constant, T = 353 K is the temperature, andm = 2.21 × 10−25 kg the mass of a cesium atom.
The inferred Lorentzian width is significantly larger thanthe natural line-width γlifetime = 1/(2πτ) = 5.2 MHz, whereτ is the D2 lifetime of 30.4 ns. This may be explained by thefact that the thermal atoms pass the evanescent field of thefiber within a typical transit time of about 1 ns.
These measurements show that nanofibers can also beused for high-resolution spectroscopic studies of gases andliquids surrounding the fiber waist. A similar experiment ofnanofiber-based vapor spectroscopy has been done with ru-bidium vapor [61].
6 Toward dispersive spectroscopy
Dispersive spectroscopy is interesting because it avoids de-structive effects like, e.g., molecule photobleaching.A Mach–Zehnder interferometer (MZI) is the natural can-didate for this purpose. With the nanofibers, we realize theMZI by modal interference. The interference is achieved us-ing the single-mode fiber pigtails of the TOF as the interfer-ometer input and output, the down-taper as a beam splitterand the up-taper as a recombiner. In comparison with thenanofiber described above, the taper slope here is steeper,i.e. nonadiabatic, in order to induce coupling between thefundamental and higher order modes. If the entire nanofiberis cylindrically symmetric around its axis, light fields can becoupled only between the modes with the same azimuthal
Fig. 17 Contour plot of the transversal intensity distribution of theHE11 and HE12 modes. The dashed line indicates the nanofiber sur-face. The calculation was performed for a 1.3-μm diameter nanofiberand linearly polarized light in the x-direction at a wavelength of850 nm
symmetry, e.g., from the HE11 mode to HE12, HE13, etc. Atthe same time, a waist diameter can be chosen such that theHE13 and the higher HE1n modes are cut off (see Fig. 4),while the two lowest modes HE11 and HE12 are guidedalong the waist. Since the output pigtail is a single-modefiber, the HE12 mode is not guided there. The output powerthus depends on the interference in the up-taper, determinedby the relative phase accumulated by the HE11 and HE12
modes as they propagate along the waist: light can be eithercoupled back into the HE11 mode and guided out, or to theHE12 mode and be lost by coupling to radiative modes ofthe up-taper. Therefore, the output power is given by:
Pout ∼ sin(ϕ11 − ϕ12) with ϕ1n = β1nL, (8)
where β = 2πneff/λ is the propagation constant of a mode,L is the propagation distance, ϕ1n is the phase accumulatedby the HE1n mode, and n = 1 or 2.
The amount of light outside the fiber body is differentfor the HE11 and HE12 modes. For example, for a waist di-ameter of 1.3 μm and a wavelength of 850 nm, 2% of theoptical power of the HE11 mode and 31% of the opticalpower of the HE12 mode is propagating in the evanescentfield. Figure 17 shows the calculated intensity distributionfor both modes. As a result, a dispersive medium aroundthe fiber will influence the phase accumulated by the HE12
mode more strongly than that for the HE11 mode, and thuscreate a measurable phase difference.
We have implemented and tested such an interferometer.For this purpose, we use a Fibercore SM800 fiber with a purefused silica cladding, which is single-mode for wavelengthsabove 800 nm. The pulling rig was programmed to producea taper slope of 16 mrad and a waist of 1.3 μm diameter and3 mm length. After pulling, the interferometer was tested inthe pulling rig by coupling a single-frequency diode laser at852 nm into the input pigtail of the fiber. The output wasdetected with a photodiode, and the stretcher stage of thepulling rig was used to elastically stretch the fiber by a fewten micrometers.
AU
TH
OR
’S P
RO
OF
Journal ID: 340, Article ID: 4730, Date: 2011-09-08, Proof No: 1, UNCORRECTED PROOF
« APHB 340 layout: Large v.1.3.2 file: aphb4730.tex (ELE) class: spr-twocol-v1.3 v.2011/08/17 Prn:2011/09/07; 17:42 p. 13/14»« doctopic: OriginalPaper numbering style: ContentOnly reference style: aps»
Optical nanofibers and spectroscopy
1297 1351
1298 1352
1299 1353
1300 1354
1301 1355
1302 1356
1303 1357
1304 1358
1305 1359
1306 1360
1307 1361
1308 1362
1309 1363
1310 1364
1311 1365
1312 1366
1313 1367
1314 1368
1315 1369
1316 1370
1317 1371
1318 1372
1319 1373
1320 1374
1321 1375
1322 1376
1323 1377
1324 1378
1325 1379
1326 1380
1327 1381
1328 1382
1329 1383
1330 1384
1331 1385
1332 1386
1333 1387
1334 1388
1335 1389
1336 1390
1337 1391
1338 1392
1339 1393
1340 1394
1341 1395
1342 1396
1343 1397
1344 1398
1345 1399
1346 1400
1347 1401
1348 1402
1349 1403
1350 1404
Fig. 18 Interference fringes due to stretching of the fiber. The elonga-tion of the waist leads to an additional phase shift between the HE11and the HE12 modes. The slight modulation of the sinusoidal signalmay indicate a weak presence of further modes possibly excited due tonanofiber asymmetries
Figure 18 shows the resulting interference signal betweenthe HE11 and the HE12 modes [34]. The changes in the phasedifference are expected to be given by:
�(ϕ11 − ϕ12) = (β11 − β12)�L + �(β11 − β12)L, (9)
where the term �(β11 − β12) takes into account the influ-ence of both the Poisson and photoelastic effect on the prop-agation constants. The Poisson effect leads to the stretch-induced change of the fiber diameter. This, according toFig. 4, results in the change of the propagation constantsof both of the propagating modes. The photoelastic effectexplains the change of the material refractive index understrain. This change, in turn, is also influencing the propaga-tion constants [41].
The measurement outcome demonstrates the realizationof a single-nanofiber-based interferometer with good inter-ference contrast. A possible application of this device is tomeasure variations of the refractive index of the surroundingmedium. The latter will differently change the propagationconstants β, and thus lead to interference fringes which canthen be traced.
7 Outlook
In this review, we have shown applications of fiber-based ab-sorption and fluorescence spectroscopy. Our spectroscopictechnique could be useful for biosensing, extending the pos-sible applications from label-free detection by absorptionmeasurements [62] to label-based detection using fluores-cence measurements [63]. The fact that we are able to si-multaneously record absorption and fluorescence spectra fora given surface coverage, in conjunction with the stable and
reproducible fluorescence collection efficiency, makes ourmethod ideal for analyzing and quantifying dynamic pro-cesses at or near the nanofiber surface. Excitation and fluo-rescence collection via a single fiber mode yields an entirelyfiber-based method that can be used for spectroscopic stud-ies at remote locations.
We have also demonstrated our first results on the imple-mentation of a nanofiber-based interferometer. The next goalis to perform dispersive detection on emitters adsorbed at thefiber waist and to compare the results with the absorptionmethod. Further, the dispersive method should allow us todetect species that do not absorb in the available wavelengthrange or to carry out measurements on sensitive moleculeswhile avoiding photobleaching.
Acknowledgements This work was supported by the DeutscheForschungsgemeinschaft DFG (Research Unit 557), the Volkswa-gen Foundation (Lichtenberg Professorship), the European ScienceFoundation (European Young Investigator Award), and the EuropeanComission (STREP CHIMONO).
References
1. A.W. Snyder, J.D. Love, Optical Waveguide Theory (Chapman &Hall, New York, 1983)
2. G. Brambilla, J. Opt. A 12, 043001 (2010)3. F. Le Kien, J.Q. Liang, K. Hakuta, V.I. Balykin, Opt. Commun.
242, 445 (2004)4. J. Villatoro, D. Monzón-Hernández, Opt. Express 13, 5087 (2005)5. P. Polynkin, A. Polynkin, N. Peyghambarian, M. Mansuripur, Opt.
Lett. 30, 1273 (2005)6. L. Zhang, F. Gu, J. Lou, X. Yin, L. Tong, Opt. Express 16, 13349
(2008)7. F. Warken, E. Vetsch, D. Meschede, M. Sokolowski, A. Rauschen-
beutel, Opt. Express 15, 11952 (2007)8. A. Stiebeiner, O. Rehband, R. Garcia-Fernandez, A. Rauschen-
beutel, Opt. Express 17, 21704 (2009)9. T.A. Birks, W.J. Wadsworth, P.St.J. Russell, Opt. Lett. 25, 1415
(2000)10. V. Grubsky, A. Savchenko, Opt. Express 13, 6798 (2005)11. R.R. Gattass, G.T. Svacha, L. Tong, E. Mazur, Opt. Express 14,
9408 (2006)12. U. Wiedemann, K. Karapetyan, C. Dan, D. Pritzkau, W. Alt, S.
Irsen, D. Meschede, Opt. Express 18, 7693 (2010)13. G. Sagué, E. Vetsch, W. Alt, D. Meschede, A. Rauschenbeutel,
Phys. Rev. Lett. 99, 163602 (2007)14. K.P. Nayak, P.N. Melentiev, M. Morinaga, F. Le Kien, V.I. Ba-
lykin, K. Hakuta, Opt. Express 15, 5431 (2007)15. K.P. Nayak, K. Hakuta, New J. Phys. 10, 053003 (2008)16. M.J. Morrissey, K. Deasy, Y.Q. Wu, S. Chakrabarti, S.N. Chor-
maic, Rev. Sci. Instrum. 80, 053102 (2009)17. E. Vetsch, D. Reitz, G. Sague, R. Schmidt, S.T. Dawkins,
A. Rauschenbeutel, Phys. Rev. Lett. 104, 203603 (2010)18. S.M. Spillane, T.J. Kippenberg, O.J. Painter, K.J. Vahala, Phys.
Rev. Lett. 91, 043902 (2003)19. K. Srinivasan, P.E. Barclay, M. Borselli, O. Painter, Phys. Rev. B
70, 081306(R) (2004)20. P.E. Barclay, K. Srinivasan, O. Painter, B. Lev, H. Mabuchi, Appl.
Phys. Lett. 89, 131108 (2006)21. K. Srinivasan, O. Painter, Phys. Rev. A 75, 023814 (2007)
AU
TH
OR
’S P
RO
OF
Journal ID: 340, Article ID: 4730, Date: 2011-09-08, Proof No: 1, UNCORRECTED PROOF
« APHB 340 layout: Large v.1.3.2 file: aphb4730.tex (ELE) class: spr-twocol-v1.3 v.2011/08/17 Prn:2011/09/07; 17:42 p. 14/14»« doctopic: OriginalPaper numbering style: ContentOnly reference style: aps»
R. Garcia-Fernandez et al.
1405 1459
1406 1460
1407 1461
1408 1462
1409 1463
1410 1464
1411 1465
1412 1466
1413 1467
1414 1468
1415 1469
1416 1470
1417 1471
1418 1472
1419 1473
1420 1474
1421 1475
1422 1476
1423 1477
1424 1478
1425 1479
1426 1480
1427 1481
1428 1482
1429 1483
1430 1484
1431 1485
1432 1486
1433 1487
1434 1488
1435 1489
1436 1490
1437 1491
1438 1492
1439 1493
1440 1494
1441 1495
1442 1496
1443 1497
1444 1498
1445 1499
1446 1500
1447 1501
1448 1502
1449 1503
1450 1504
1451 1505
1452 1506
1453 1507
1454 1508
1455 1509
1456 1510
1457 1511
1458 1512
22. B. Min, S. Kim, K. Okamoto, L. Yang, A. Scherer, H. Atwater,K. Vahala, Appl. Phys. Lett. 89, 191124 (2006)
23. K. Srinivasan, O. Painter, A. Stintz, S. Krishna, Appl. Phys. Lett.91, 091102 (2007)
24. L. Tong, R.R. Gattass, J.B. Ashcom, S. He, J. Lou, M. Shen, I.Maxwell, E. Mazur, Nature 426, 816 (2003)
25. G. Brambilla, V. Finazzi, D.J. Richardson, Opt. Express 12, 2258(2004)
26. A.M. Clohessy, N. Healy, D.F. Murphy, C.D. Hussey, Electron.Lett. 41, 954 (2005)
27. M. Sumetsky, Y. Dulashko, P. Domachuk, B.J. Eggleton, Opt. Lett.32, 754 (2007)
28. F. Warken, A. Rauschenbeutel, T. Bartholomus, Photon. Spectra42(3), 73 (2008)
29. S. Pricking, H. Giessen, Opt. Express 18, 3426 (2010)30. J.D. Love, W.M. Henry, Electron. Lett. 22, 912 (1986)31. A. Stiebeiner, R. Garcia-Fernandez, A. Rauschenbeutel, Opt. Ex-
press 18, 22677 (2010)32. R.A. Potyrailo, S.E. Hobbs, G.M. Hieftje, Fresenius J. Anal.
Chem. 362, 349 (1998)33. J. Bures, R. Ghosh, J. Opt. Soc. Am. A 16, 1992 (1999)34. F. Orucevic, V. Lefevre-Seguin, J. Hare, Opt. Express 15, 13624
(2007)35. J. Lou, L. Tong, Z. Ye, Opt. Express 13, 2135 (2005)36. T.A. Birks, Y.W. Li, J. Lightwave Technol. 10, 432 (1992)37. A. Yariv, Optical Electronics (Saunders College Publishing,
Philadelphia, 1991)38. J.D. Love, W.M. Henry, W.J. Stewart, R.J. Black, S. Lacroix,
F. Gonthier, IEE Proc. Optoelectron. 138, 343 (1991)39. J.A. Buck, Fundamentals of Opticals Fibers (Wiley, Hoboken,
2004)40. D. Meschede, Optics, Light and Lasers (Wiley-VCH, Berlin,
2006)41. K. Karapetyan, U. Wiedemann, C. Dan, D. Pritzkau, F. Bruse,
W. Alt, D. Meschede, Optical Fibre Toolbox for MATLAB.http://www.mathworks.com/matlabcentral/fileexchange/27819(2010)
42. G. Brambilla, D.N. Payne, Nano Lett. 9, 831 (2009)
43. T.A. Birks, J.C. Knight, T.E. Dimmick, IEEE Photonics Technol.Lett. 12, 182 (2000)
44. M. Sumetsky, Y. Dulashko, J.M. Fini, A. Hale, J.W. Nicholson,Opt. Lett. 31, 2393 (2006)
45. V. Grubsky, J. Feinberg, Opt. Commun. 274, 447 (2007)46. J. Laegsgaard, J. Opt. Soc. Am. B 27, 1317 (2010)47. T. Ogawa, K. Kuwamoto, S. Isoda, T. Kobayashi, N. Karl, Acta
Crystallogr., B Struct. Crystallogr. Cryst. Chem. 55, 123 (1999)48. F. Le Kien, S.D. Gupta, V.I. Balykin, K. Hakuta, Phys. Rev. A 72,
032509 (2005)49. S.R. Forrest, Chem. Rev. 97, 1793 (1997)50. H. Proehl, Th. Dienel, R. Nitsche, T. Fritz, Phys. Rev. Lett. 93,
097403 (2004)51. S.R. Forrest, Y. Zhang, Phys. Rev. B 49, 11297 (1994)52. V. Bulovic, P.E. Burrows, S.R. Forrest, J.A. Cronin, M.E. Thomp-
son, Chem. Phys. 210, 1 (1996)53. R. Scholz, A.Yu. Kobitski, T.U. Kampen, M. Schreiber, D.R.T.
Zahn, G. Jungnickel, M. Elstner, M. Sternberg, Th. Frauenheim,Phys. Rev. B 61, 13659 (2000)
54. W.P. Ambrose, Th. Basché, W.E. Moerner, J. Chem. Phys. 95,7150 (1991)
55. H. Proehl, R. Nitsche, Th. Dienel, K. Leo, T. Fritz, Phys. Rev. B71, 165207 (2005)
56. R.F. Kubin, A.N. Fletcher, J. Lumin. 27, 455 (1982)57. H. Du, R.-C.A. Fuh, J. Li, L.A. Corkan, J.S. Lindsey, J. Pho-
tochem. Photobiol. A, Chem. 68, 141 (1998)58. T.F. Heinz, C.K. Chen, D. Ricard, Y.R. Shen, Phys. Rev. Lett. 48,
478 (1982)59. J.B. Birks, D.J. Dyson, Proc. R. Soc. Lond. A 275, 135 (1963)60. D.A. Steck, Cesium D line data. Available online at
http://steck.us/alkalidata, rev. 2.1.4 (2010)61. S.M. Hendrickson, T.B. Pittman, J.D. Franson, J. Opt. Soc. Am. B
26, 267 (2009)62. X. Fan, I.M. White, S.I. Shopoua, H. Zhu, J.D. Suter, Y. Sun, Anal.
Chim. Acta 620, 8 (2008)63. M.D. Marazuela, M.C. Moreno-Bondi, Anal. Bioanal. Chem. 372,
664 (2002)