对科研工作者来说，追求文章高引用率是其工作的目标和源动力之一。这意味着其工作受到同行的关注和重视。然而物极必反，过于注重引用率，导致互引，钓鱼引用等成为学术界的潜规则；再加上商业化时代SCI， IF等各种指数的推波助澜，使学术越来越背离其本来面目：追求真理和质疑精神。也就是说，学术面前人人平等，任何人对任何学术问题都可质疑和证伪。
故本人不怕贻笑于方家，尝试评论一下H. Zeng & W. Cai的几篇文章。该课题组在 JPCB 2005， 109， 18260（jp052258n）和 JPCB 2007, 111, 14311 (jp0770413)中介绍了用laser ablation的方法制备ZnO/Zn的壳/核结构纳米粒子。该课题组在2008年在ACS Nano, 2008, 2, 1661（nn800353q）上发表通过侵蚀该种壳/核纳米粒子的Zn核，可以进一步制备ZnO空心球结构。我一直质疑该文章（nn800353q）的学术想法来源，是窃取本人在2007年MP3会议上得来的学术灵感。因为曾海波文章的投稿时间比本人晚八个月，且本人当时的学术想法的另一部分被该课题组李越窃取。从李的文章的逻辑和时序矛盾可以得出其想法不为该课题组所有（详见http://lk2011.blogdive.com/2011/10/19/sc1/）。以下是我对这三篇文章的一点评论，也是对两篇博文http://blog.sciencenet.cn/home.php?mod=space&uid=2425&do=blog&id=274830以及http://blog.sciencenet.cn/home.php?mod=space&uid=2425&do=blog&id=524134的回应。
首先，在两篇JPCB文中，Zeng介绍了利用激光烧蚀生成的高温高压Zn等离子体，经历绝热膨胀冷却后生成Zn cluster，再与水作用得到ZnO。而在含有表面活性剂SDS的水溶液中，能够得到ZnO/Zn的壳/核结构。作者观测到随着SDS浓度上升，ZnO相的含量逐渐减少，Zn相的含量增加。这一点从XRD，吸收光谱，PL等结果都能够反应。作者特别在jp052258n文中提到当SDS达到0.03M，ZnO的UV吸收峰（大约位于350nm）即消失，Zn的SPR峰（242nm附近）开始出现（参见 Fig5，注意样品h是在0.05M SDS中制备的）。
让人难以理解的是在nn800353q中，作为precursor的ZnO/Zn壳/核粒子是在0.05M SDS中制备的，其吸收光谱中（参见Fig2b, Fig3）仍然出现ZnO的UV吸收（是否自相矛盾？），并且作为重要证据来与酸侵蚀后的产物（即作者意图表征的“ZnO空心球”）吸收光谱相比较。
在该文里ZnO/Zn壳/核结构XRD的图谱（Fig2a）能否用来表征“ZnO纳米空心球”的产生值得怀疑。因为在SDS 0.05M条件下合成的产物仍然是混合体系，该XRD的ZnO相来源有两部分：一部分是Zn cluster与水直接生成的ZnO，另一部分是被SDS“保护”的Zn核氧化而产生的ZnO壳层。XRD光谱并不能区分这两种ZnO。也就是说，酸侵蚀后的遗留的ZnO相的归属存疑。换言之，作者不能排除ZnO/Zn壳/核结构被全部酸解，所遗留的ZnO的XRD信号来自于在水中由Zn cluster直接生成的ZnO。（由两篇JPCB的电镜照片可以看到，在水中直接生成的ZnO粒径在50 nm左右，而在ZnO/Zn壳/核结构的ZnO壳层中的晶体粒径在5 nm左右。前者体积是后者的千倍，因此，后者更容易被酸完全侵蚀。）
进一步说，随着酸侵蚀，Zn核的XRD特征峰消失，同时ZnO的晶体结构也被严重破坏，其XRD特征峰也大大削弱，与背景吸收已经很接近，即声噪比接近1（参见Fig2a的样品3），这样的结果能否用来证明“ZnO空心球”的生成？其结果在实验上的可重复性值得怀疑。(Fig2 的XRD和Fig4 的SAED都证实酸侵蚀后的产物中绝大多数是无定形态的物质，能否仍然称为“ZnO”？）
同样，酸侵蚀后由于ZnO的晶体结构被严重破坏至无定形状态，作者仍然在ZnO的晶体缺陷学和能带理论的范畴来解释酸侵蚀后的产物出现的蓝色发光增强的现象（Fig6)，显然是不可靠的。
其次，作者为了解释酸侵蚀的过程的机理，提出如下的假说：对于酒石酸(TA)这样的弱酸与ZnO/Zn壳/核作用，是质子（H+）穿过ZnO壳层结构中的grain boundary，向内进入Zn核区域与之反应而得到空心球。而对于含贵金属的弱酸如HAuCl4，作者的解释是含糊不清的。在论文的开始部分他们认为同质子相似，是AuCl4-离子通过壳层进入Zn核而与之作用。在文章后面的讨论部分，他们又提出类似于Kirkendall 效应的机理。显然，通过去除壳/核纳米结构的核部分得到空心纳米球和通过Kirkendall 效应得到空心纳米球是两种不同的方法，但是似乎作者并不加以区分。在该文中，相比较于质子，AuCl4-体积巨大，无法通过ZnO壳向内扩散。如果是Kirkendall 效应起主导作用，则应该是Zn核的原子外向扩散通过ZnO壳后与AuCl4-离子作用。作者却没有给出足够的理论或实验证据来支撑这一观点，并且这一机理也无法解释为何Au等贵金属会沉积在ZnO空心球体内。（因为按照作者的解释，贵金属纳米粒子应该生成于ZnO壳层）。
最后，在两篇JPCB文中，作者提出了液相激光烧蚀过程中的纳米粒子的产生过程的机理。但是他们的模型都是描述性的，只是将实验现象简单的用模型陈述出来而已，无法揭示这一现象背后真正的物理化学过程。比如，作者发现激光功率的增加会导致ZnO壳层厚度的增加，然而作者认为这是由于功率增加会产生更多的等离子体。一般来说，改变反应物的浓度导致反应的速度改变，而不是改变生成物的结构，。因此，我认为合理的解释是由于加大激光功率导致表面活性剂SDS与Zn cluster的界面不稳定，进而使Zn核被氧化的程度加深所致。同样，SDS浓度上升，会使ZnO/Zn纳米粒子的粒径减小，粒径的分布收窄这一现象，作者也没有利用现有的物理化学知识给予合理的解释。
由于在这三篇文章里报道的工作是作者一系列后续论文的基础，前提的不可靠带来了对其后期工作质量的不少疑问。
Comment on Some Work of Zeng & Cai
Prof. Cai’s group introduced the preparation of Zn/ZnO core/shell nanoparticle (CSNP) via laser ablation method in their two publications in J. Phys. Chem. B, i.e. JPCB 2005, 109, 18260 (jp052258n) and JPCB 2007, 111, 14311 (jp077413). And in another paper (nn800353q) published in 2008 in ACS Nano they further developed a technique to synthesize ZnO hollow nanoparticles (HNP) by etching the obtained Zn/ZnO CSNP, which was employed as a sacrificial template, in weak acid aqueous solution. I have been questioning the origin of the idea presented in this paper (nn800353q) in the last years, for the authors are suspected to have stolen the ideas that I conceived in the MP3 conference in Beijing in 2007 and put them into this paper. The detailed story and proofs you may find in my email exchange with editors of several journals.
In the following section, I would like to comment on the above-mentioned work of Cai group，in order to prove that even with the stolen ideas, the group is unable to realize them into laboratory work with reasonable quality.
First, in the two JPCB papers, Zeng reported that in the laser ablation process, zinc plasma with high temperature and high pressure was generated by laser ablation firstly, which turned into zinc clusters after a subsequent ultrasonic and adiabatic expansion. Thereafter two competitive reactions may take place simultaneously in the aqueous solution. When the formed zinc clusters encounter water molecules, zinc oxide nanoparticles come into being; while in the aqueous solution containing surfactant such as SDS, Zn nanoparticles are produced at first, owing to the protecting and capping effects of SDS molecules. Because of highly reductive potential of the Zn nanoparticles, their surface is oxidized into ZnO, resulting in the fabrication of Zn/ZnO core/shell nanostructure. It is observed that with the increase of SDS concentration, the amount of ZnO phase drops significantly, whereas the content of Zn phase increases accordingly. The finding is reflected by the results of XRD together with absorption spectra as well as photoluminescence (PL). Moreover, the authors pointed out in paper jp052258n with emphasis “when the SDS concentration is about 0.03 M, the peak at 350 nm vanishes, and another peak at 242 nm appears and increases with further increase of the SDS concentration”. (See Fig.5; note that sample h is prepared in 0.05M SDS). Here, the peak at 350 nm is attributed to the UV absorption of ZnO and the peak at 242 nm is due to the Zn SPR absorption.
It is not understandable that in paper nn800353q the Zn/ZnO CSNP, prepared in 0.05M SDS and employed as a precursor, still shows evident UV absorption around 350 nm (Fig.2b, Fig.3), which is utilized as a chief proof for confirming the generation of ZnO hollow nanoparticle by comparing the absorption spectra between Zn/ZnO CSNP and ZnO HNP. Obviously, the results of the UV absorption of ZnO contradict with each other in the two papers (nn800353q & jp052258n).
Also, in nn800353q it is questionable whether the XRD pattern shown in Fig.2a can be used for determination of the crystal structure of the ZnO hollow nanoparticles. We know that, in the same preparation conditions as sample h (in 0.05M SDS) the obtained product is still a mixture. And the characteristic reflection of ZnO in the XRD pattern may originate from two sources. One is the ZnO generated by Zn cluster directly in water (we can name it ZnOw); the other is the ZnO in the CSNP shell, formed by the oxidation of the Zn core, named as ZnOS herein. Unfortunately, the XRD pattern itself cannot provide information to differentiate the two kinds of ZnO crystals, so the assignment of the ZnO phase after the acid etching of the Zn/ZnO CSNP is still unclear. In other words, the authors cannot rule out the possibility that the ZnOS is completely decomposed and the remaining signal in the XRD pattern arises from the existence of ZnOw. From the TEM images presented in the two JPCB papers, we can see that the particle size of ZnOw is around 50 nm, while the diameter of ZnOS is only 5 nm. The volume of the former is thousand time large as the latter. Undoubtedly, the ZnOS particles are much more vulnerable to acid etching.
Furthermore, with the acid etching going on, the characteristic peaks of Zn core in XRD patterns disappear gradually. And the crystal structure of ZnO is damaged severely as well, leading to the considerable weakening of the intensity of the ZnO peaks, which is very close to the background reflection, indicating that S/N is nearly 1. Thus the reproducibility of the experiments is doubtful. It is also worth noting that both the results of XRD in Fig.2 and SAED in Fig.4 show that the product after the acidic etching of Zn/ZnO CSNP, referred as ZnO HNP by the authors, consists of mostly amorphous materials due to the destruction of ZnO crystal structure. Hence it is clearly unreliable that the authors interpret on the phenomenon of the enhancement of ZnO HNP blue emission, still in the scope of ZnO crystal defects and bandgap theories (see Fig.6).
Next, in order to explain the mechanism on the process of the acid etching of the Zn/ZnO CSNP, the authors propose such a hypothesis. When weak acid like tartaric acid interacts with Zn/ZnO CSNP, it is the proton (H+) that penetrates the ZnO shell via the grain boundaries and enters the Zn core area, leading to the elimination of the core and the formation of the hollow nanosphere. For acids containing noble metals such as HAuCl4, the authors’ interpretation is ambiguous. They proposed at the beginning of the paper that “noble metal ions would also diffuse into or through the nanoshell-layers during the etching process, and thus noble metal ultrafine particles (or clusters) could be deposited in the shell-layers of the final HNPs owing to the redox reaction between active and noble metal.” And at the end of the paper, they assumed the reaction is “to some extent, similar to the nanoscale Kirkendall effect.” Clearly, the generation of hollow nanosphere via decomposing the core part of a CSNP or applying a Kirkendall-effect-based reaction on a CSNP is two different approaches. In the specific case as above-mentioned, in comparison with proton the anion AuCl4- has a huge size which hampers its diffusion ability, so that it cannot pass through the shell layers. If the Kirkendall effect is dominant, the redox reaction should take place in the ZnO shell, owing to the outward diffusion of Zn atoms (inward diffusion of vacancies). However, TEM images as presented in Fig.9, Fig 10 and Fig.11 show that many noble metal nanoparticles deposit inside the hollow spheres, which is not interpretable by the mechanism as the authors put forward.
Finally, in the two JPCB papers, the authors propound a mechanism for the fabrication of Zn/ZnO nanoparticles by laser ablation. However they simply described the experimental results via an illustrative scheme, without unveiling the true physical and chemical processes behind the phenomena. For instance, the authors observed that decreasing the laser power causes the reduction of the shell thickness of the Zn/ZnO CSNP. They attributed the fact to less zinc clusters and particles generated by the decrease of laser power. Normally, the change in the concentration of reactant (Zn particles in this case) may not alter the structure of the resultant but the rate of the reaction. I hence suggest that less laser power causes the increase in the stability of the interface between the surfactant and Zn particle, leading to less extent of oxidation on the surface of Zn particle, and consequently resulting in the thinning of the shell. Similarly, the observation that both the size and size distribution decrease with the increase of SDS concentration is not explained by the authors on the basis of the available knowledge of physical chemistry.