Dr. Saad E. Hebboul

Appeal LP9512

April 20, 2005

April 20, 2005

Dr. Saad E. Hebboul,

Assistant Editor

Physical Review Letters

Re: LP9512

Evolution of excitons via biexcitons to an electron-hole plasma without level crossing between band edge and exciton in a quantum wire

By: Masahiro Yoshita, Yuhei Hayamizu, Hidefumi Akiyama, et al.

Dear Dr. Saad E. Hebboul,

We APPEAL for review by a divisional associate editor of PRL. We learned that our revised manuscript and cover letter of January 11, 2005, which we wrote to the assistant editor and a divisional editor, went back to the previous referees A, B, and a new referee C by the editorial decision, because the editor cannot skip the second-round review process before an appealing process. Since the second-round review was concluded on February 28, 2005, we now APPEAL for review by a divisional associate editor.

Our main reason for the APPEAL is because we believe that the referees have strong bias between our present T-wire work and the previous V-groove-wire reports mentioned in their comments. The referees A, B, and C claim that our data on our present T-wire are SIMILAR TO or have NO ADVANTAGE OVER the previous reports on V-groove wires with very large disorder. We want to ask neutral fair judgment on this case by a divisional associate editor.

We understand that the assistant editor selected the referees via a normal process. A problem is, however, that there are only limited numbers of research groups in the field of quantum-wire optics. In particular, sample sources of quantum wires are very limited in the world, and most researchers cannot study quantum wires with high quality. Therefore, if referees were selected from authors of the cited papers, or if some referee candidates declined to review and recommended another reviewer who works close to this subject, our manuscript had very high probability to go to our competitors.

Our T-shaped quantum wires have unprecedented high quality. In fact, our wires are revealing new clear experimental data, which correct or deny some vague or wrong discussions claimed by previous reports. This fact is widely accepted via our invited talks in open meetings such as APS, CLEO/QELS, MRS, and JPS. Therein, we never had criticisms on the qualities of our wires or data in comparison with previous V-groove-wire reports, which the referees made in anonymous review process. We suspect that our competitors working on lower quality wires or their close collaborators made biased comments as anonymous referees, because they felt that publication of higher quality data for higher quality samples is fatal for them. In any event, the biased referee comments which intend to exclude new data on state-of-the-art high-quality quantum wires from PRL journal are very unfair.

Furthermore, we are very unhappy about the style of the second-round reports by referees A and B. In the previous cover letter of January 11, 2005,we wrote our complete answers to all the first-round comments by referees A and B, stressing our 3 highlights. We made 5 major revisions complying with some comments by the referees. However, the referees did not respond directly to our 3 highlights and respective answers, but added or repeated their criticisms neglecting our replies for their previous comments. Therefore, the arguments never converge. We need assistance by a divisional editor to settle this situation.

We sincerely ask a divisional editor to fairly evaluateour manuscript and our points in physics described in detail in the previous cover letter of January 11, 2005.Since divisional editors such as Dr. Deveaud and Dr. Pinczuk are experimental experts in this field, they can surely judge whether the referees are our competitors or neutral researchers by their names, and can surely make a fair evaluation on our manuscript, our cover letters, and the referee comments.

In ourprevious cover letter, westressed our 3 highlights, 5 points of revisions, and our complete answers to all the questions and concerns by referees A and B. As far as we examined the second-round reports by referees A, B, and the report by referee C, our main 3 highlights were not fairly recognized by the referees due to their biases. In the following, we point out essential problems in the latest referee comments related to our main 3 highlights.

Highlight 1:on high quality of the single T wire

All the referees admit that our T-wire has unprecedented high quality, but seem to have a biased belief that the V-groove wires studied in the previous reports (ref. [12] by Guillet et al. and ref. [11] by Crottini et al.) had enough high quality to study universal 1D quantum-wire physics in carrier-density dependent evolution of PL. This must be the reason why the referees made unfair comments that our present results are similar to or simply confirming the results of the previous V-groove-wire study.

Note, however, that the V-groove wires in ref. [12] by Guillet et al. and ref. [11] by Crottini et al. show strongly localized PL. The PL is spatially chopped up into short pieces, and the PL peak energy fluctuates more than 10 meV. Excitons are three-dimensionally confined by a localization energy of more than 10 meV. Biexciton binding energy (1.5mV in ref. [12]) and carrier-induced broadening of excitons and biexcitons are far smaller than the 10meV energy of structural inhomogeneity and localization. Therefore, these systems are like 0D quantum dots rather than 1D quantum wires. In fact, people often study physics of 0D quantum dots using strongly disordered quantum wells. This major problem in the previously published papers [7-12] is clearly explained in the introduction of our manuscript and in the previous cover letter, though the referees mentioned nothing on it in the second reports.

In our experiment, the PL peak of T-wire has narrow total width of 1.3 meV and a small Stokes shift less than 0.3meV. We admit that structural inhomogeneity exists, which may dominantly contribute to the broadening at the low-density limit. However, these energy values of structural inhomogeneity (<1.3meV or 0.3meV) are smaller than relevant many-body interaction energies such as a biexciton binding energy of 2.8meV and a carrier-induced width increasing gradually to 12meV. Therefore, in the intermediate- and high-excitation regions of our interest to investigate the 1D exciton-plasma crossover, the PL spectra are homogeneously broadened and inhomogeneous broadening plays only a negligible role.

High wire quality that meets this criterion on small inhomogeneity is crucially important to study 1D many-body physics of broad interest, and our present T-wire is the first sample with such high quality. Therefore, the main comment by the referee B that "our T-wire is of high structural interest but is not of broad interest" is wrong and very unfair.

We surely performed ensemble measurements with 1 micron spatial resolution on our T-wire, as the referee C mentioned. We also admit that NSOM with 0.2 micron spatial resolution used in ref. [11] by Crottini et al. is an advanced tool. However, the key difference between the two experiments is not the 1/ 5 spatial resolution, but the values of localization energy or the energy fluctuation along wires in comparison with the many-body interaction energies, which separates the systems to 1D wires or 0D dots. The comments by the referee C that "the ref. [11] measured a single quantum state in the quantum WIRE", and that "the referee sees no real advantages in this highlight" are strongly biased and unfair.

Highlight 2: on crossover of biexcitons to an e-h plasma

The second highlight of our paper is the observed crossover of biexciton PL peak to an e-h plasma peak as a result of continuous broadening. The highlight 2 is strongly linked to the highlight 1. Note that the observed broadening in the crossover regime, or the intermediate- and high-excitation-density regime, of our major interest is NOT inhomogeneous broadening, but is intrinsic homogeneous broadening. Therefore, the claim by the referee C that ensemble measurements do not detect intrinsic features of quantum wires is wrong. The argument by the referees B and C that relevant spectral features were smeared out by broadening and were invisible does not make sense, because the observed intrinsic homogeneous broadening is nothing but the relevant quantum-wire physics. It is a pity that the referees did not change their conclusions on the highlight 2 most probably due to biased judgment on the highlight 1.

In the second-round report, the referee A simply commented that the observed phenomena are the same as those observed in V-groove wires in the previous reports. The referee A did not respond to our respective answers and detailed explanation in the previous cover letter.

The referee B, in the second-round report, doubted our assignment of the high-density peak to an e-h plasma, and claimed that the peak is due to broadened X and XX peaks. However, our direct evidence is given in the wire PL intensity plot of Fig. 2 (a) explained in the 8th and 10th paragraph of our manuscript. The wire PL intensity saturates above excitation intensity of 4x10^-4 W, because all the states in the wire are filled. In other words, degenerate e-h plasma is formed. This PL intensity saturation is analogous to the saturation plateau observed in time-resolved PL in ref. [12] by Guillet et al., which was assigned to a characteristic of e-h plasma emission in ref. [12]. Thus, the doubt raised by the referee B must be completely resolved.

The referee C claimed that the broader e-h plasma peak should coexist with the sharper biexciton peak, as was observed in V-groove wires in the previous reports. This claim must stem from the biased belief in the previous V-groove wire results: PL data of Guillet et al. and Crottini et al. in the intermediate densities show that sharp biexciton PL coexists with broad background plasma PL. The referee C proposed to add an additional fitting component of a broader background curve in addition to the two Lorentzian curves of the exciton and the S-peak. It is surly possible to increase fitting components or parameters, but our PL data in the crossover regime is already very well fit by the two Lorentzian curves. Hence, the spectral weight of the additional broader background curve is negligibly small.

Highlight 3: on the onset of the exciton band edge and band-gap renormalization (BGR)

The referee C approved that the onset of the exciton band edge and BGR is relevant, and fully agreed to the comments of the authors. On the other hand, the referees A and B did not respond directly to this point.

In the latest reports by the referees, they criticized about some of our terminologies used in the manuscript and citation. If these are indeed not appropriate, we are willing to revise. However, we feel that the referees' criticisms are extremes rather than fair comments. We want to ask suggestions by a divisional editor.

On the term "free excitons"

We used a word "free exciton" in the manuscript, quantitatively noting that there is a finite PL width of 1.3 meV and a Stokes shift less than 0.3 meV. We believe that the criticism by the referee B, "there simply is no free delocalized exciton state in a disordered quantum wire even if the disorder is arbitrarily small", is an extreme, and never makes a fair reason for rejection.

We used the expression "the free exciton state is continuous over 10 micron" on the basis of our spatially resolved PL. Obviously, "continuous" does not imply "coherent". We quantitatively explained the finite PL width and the Stokes shift in the same paragraph. Though we understand what the referee C meant, we do not think it is misleading in this context.

On the term "low-energy edge of e-h plasma PL"

The referees B and C complained about the concept of low-energy edge of e-h plasma PL and the definition of triangles in Fig. 3.

In general, e-h plasma PL denotes band-to-band transition between electrons and holes forming a plasma, where low-energy edge (onset) and high-energy edge (cut-off) of PL correspond to transitions at their band edges and Fermi edges, respectively. Because of many-body effects, scatterings, and thermalization, the onset and the cut-off can have tail or broadening. It is true that we did not give this explanation in the manuscript. However, such a general picture of e-h plasma PL is well-known, and we believe that the concept of the low-energy edge of e-h plasma PL is evident.

To interpret our experimental PL data on the basis of this picture and denote positions of the edges, there is certain ambiguity due to tail or broadening. However, the discussion of Fig. 3 is not affected by such ambiguity. Therefore, the comment by the referee B that "any conclusions drawn on this edge are obviously meaningless" is an unfair extreme statement. Indeed, in the first-round reports, any complaints about our interpretation were NOT made by the referees A and also B.

The definition of triangles in Fig. 3 is clearly given in the figure caption as "the positions where the PL intensity decreases to one twentieth of the peak intensity". We may change "one twentieth" to "one tenth", for example, but the text of the manuscript require no change. Since Fig. 3 is now plotted in log-scale after the previous suggestion by the referee B, the PL tails are emphasized. Thus, positions of low-energy edge may not be very obvious for those who are not familiar with log-plots, but this is not the case with most physicists.

On the citation of Ref. 23 on a theory by Asano and Ogawa

It is not the main issue or purpose of our manuscript to justify a theory by Asano and Ogawa on strong biexcitonic correlations in quantum wires. However, it is interesting and related to the present experiment. We believe that citation of such a paper on the cond-mat preprint server is allowed in PRL. We do not understand why the citation of a paper is not acceptable.

Finally, we wish to remark that we improved our experiments and revised our manuscript significantly in the previous resubmission to fully comply with the previous referee reports and to emphasize our 3 highlights. However, the referees did not respond directly to these essential points. This is the reason why we do not further revise our manuscript in this appeal. We are willing to revise our manuscript, if a divisional editor recommends it.

We sincerely ask fair judgment of this case by a divisional editor.