美国留学申请套磁信的写作，没有必要非要参考一个写作模板，在套磁信中我们应该写出自己的特色，避免模板化。第一封信就是我们的正式套磁信。在信中，我们要说明写这封信的原因，我们的 专业背景，实验室背景，科研背景，我们为何对教授的研究感兴趣，我们的背景是怎样 match 到导师的研究的，我们是否有希望在他的实验室完成我们的研究生学习，不一而足。如果导师对我们感兴趣，自然会回复我们的信件。如果导师因为各种客观原因不能考虑我们的请求 时，他一般也会直接回复我们给我们明确的拒绝。否定的回复同样也很重要。很多同学发出了第一封一样的信给 N 多导师询问招生与否，结果回复寥寥。为什么?因为导师每天有自己 的工作要做，不可能花时间回复每个人的邮件。会选择性回复。如果他看到一封非常认真的邮件，知道我们的确希望来他实验室学习，即使他不能提供这种机会，他也会直接告诉你我 们原因免得耽误你时间。换句话说，一封有价值的套磁信才有被回复的价值，即便是否定的回复。
能够帮助佐证你科研实力与 match 程度的研究经历和研究能力。下面我们结合典型的套磁信范例，来看看怎么在套磁信中崭新我们的研究能力。
Dear Professor XX,
I am XXXX, an undergraduate student from the University of XXXXX. I am doing photovoltaics and solar energy.
Recently I have read your letter published on Nature about using releasable multilayer epitaxial assemblies of GaAs in the field of photovoltaics and optoelectronics. I think it is really marvellous that you bring the cost down while the efficiency of the solar cell is still above 20%. And just as you said, the performance can be further improved by the use of optimized cell designs. From the comparison of EQE and IQE curve, I believe that the efficiency could be enhanced by further reducing the reflectance of the cell. Since that GaAs is also used as the material of multijunction cell, I just cannot stop thinking that perhaps there is a chance to apply your technology to multijunction cells to achieve 40% efficiency if more layers could be grown. Anyway, that is just my imagination.
注意： 这一段要结合教授的论文或者研究方向提出， 自己的观点，问题或者研究
Frankly speaking, I am in my last year in University of XXXX and seeking opportunities to continue my study and research. Currently I am doing my thesis research on surface plasmon for light trapping in polycrystalline silicon thin-film solar cells. We have achieved over 40% Jsc enhancement by rear-locallised nanoparticles. And we are about to have an article published on that topic.
注意：这一段讲述自己做过的研究，自己的研究方向 应与教授的研究方向 一致
After I read your letter, I looked up the information on your group website. Things like MacEtch and MOCVD seem really fancy to me as I never knew them before. (I think MacEtch is a light trapping method to scatter light to get longer light travel path and hence more absorption. Please tell me if I am wrong.) So I wonder if there is any graduate position available, and more importantly, any research topic I could take based on my previous knowledge.
Thanks for your time reading my email.
这封套磁信的成功之处在于：学生前期仔细研读了教授的 paper，教授的研究方向与申请人 非常契合，学生能够清楚的阐述自己的研究内容并且提出自己的观点。所以很快就收到了教授的回复。
How is everything going on these days, Professor Santos? I am XX from XXX University, and we have had several long talks during the ICAST 2011.
I have read several of your papers on intelligence Analysis, and I find it very interesting to detect intent-driven insider threat and to determine the impacts of user’s cognitive styles on the analytic process and I have several questions about your papers on this topic, which will be described as follows.
1, In Intent-driven Insider Threat Detection in Intelligence Analyses, you carried out an experiment including eight original analysts and five malicious insiders to test your methodology. However, all of them came from your research group, which meant that they understood how it worked in advance. If this was true, I am wondering whether their knowledge about your algorithm would have a negative effect on the result of this experiment. In other words, if we can find some real cases containing several malicious insiders to test our model, it will be more reliable. However, I know it is much harder to find a suitable case that contains one or more malicious insiders who have been caught, for such records are always unavailable from the Internet or the library. What makes things worse is that even if we are lucky enough to find a real case that meets our hypothesis, we can’t guarantee that all of the required information including ALEs, ACH matrix, and research report are available and unaltered. Taking all of these disadvantages into consideration, I suppose a better algorithm should be designed so that we will detect deception and misleading report even if some important information is missing and we don’t know how much is left out. For example, if we miss part of cited evidence about Q1 in ACH matrix and part of argument about Q1, and we don’t know what percentage of information is available, we will find it hard to use the formula Deviationarg(Q1,Q2)=Simarg(Q1)-Simarg(Q2). Therefore, we have to evaluate the integrity of the collected data at first if we make use of real case.
2, In On Deception Detection in Multiagent Systems, the methodology of the model is divided into three steps, namely 1calculate correlations based on past opinions, 2predict current opinions, 3identify deception. However, as you know, specialists are always working on very unique and rare tasks such as military related analysis, which is quite different from other kinds of
work. If this is the first time that these agents work on this kind of task or their past opinions on similar tasks are unavailable or unreliable due to lack of sufficient supporting materials, how can we calculate the correlation of their past opinions accurately? In addition, if we have access to their past opinions, we have to guarantee their reliability so that we can use them as training data. But it is even more difficult to prove that their past opinions are free from unknown deception. Therefore, more work has to be done to ensure that our methodology is useful in such cases.
Thank you very much for your consideration on these questions, and I’m looking forward to your reply.
PS: My paper and resume have already been attached Yours Sincerely
这篇套磁信是一封非常成功的套磁信。作者在校期间，学校举办了 IEEE 会议。刚好有一批 美国的高校的访问学者来到他们学校。我们从中挑选了一些研究方向和他相符的老师，并在前期阅读了一些这些老师的文章，让他在会议后和各个教授分别进行了交流。有几位教授当 时便对这位学生产生了浓厚的兴趣，其中这位 Dartmouth 的教授把学生的简历和 Paper 带回 了国内，并主动要求保持联系。在教授回国后，我要求这位学生马上阅读该老师更多的文章，并主动联系老师。这就是老师回国后学生写的第一封套磁信，得到了老师非常详细的回复，从此两人建立了很密切的联系。同时，该学生也得到了很多高校多位教授的回复和认可。但最终，该生还是选择了 Dartmouth 的这位教授，现已在美国就读 PHD。
Dear Professor Shalaev:
I have read your papers on optical cloaking and meta-materials. Since I read more than twenty papers on cloaking, I can see the importance of nonmagnetic meta-materials and high-order transformation. I found your information on the website of the Photonics & Spectroscopy Lab. Since I was in the last year in undergraduate, in XXX University majoring in Opto-electronics Engineering and had good background of physics and maths, I take great interest in your lab.
I wonder if you would take any Ph.D student next fall 2012, and I hope you could answer the questions puzzling me. I could see the all the simulations in your papers are two dimensional. I know that 2-D and 3-D simulation results are different. Why not use 3-D simulations for precision?
Again, thank you for your attention.
这位教授是在 University of Washington (综排 41，EE 专排 17，在西雅图) 如果单看这封套磁信，可以说也称得上是一封还算不错的套磁信。该申请者的背景不错，是 上交大，复旦，北航、华科这种 level 学校的年级第二。专业是电气工程下的光电方向，并 且有一些项目和学校重点实验室的经历。但和前两封套磁信比起来，这封信的内容就显得比较单薄了。教授每天都会受到众多申请者的套磁信，申请者的背景条件和专业度教授浏览一 遍便能知道大概。本文由太傻留学广州理工组全素洁老师撰稿，如需转载请标明出处，违者必追究法律责任。虽然这位申请者和教授的专业方向相符，也阅读了教授的文章。但很明显能看出，他的功课没有做的非常充分，也没有能够很好地在和教授的交流中体现出自己的专 业水准。所以，即使这位申请者的背景条件很好，但这封套磁信仍然没能使他在众多申请者中脱颖而出。
简单来说，套磁信有两种方式。一种是海套，一种是做足功课后比较细致的套磁。海套适合 大牛，把自己的 GT 成绩，GPA 等等优势罗列在套磁信中，对于一些排名不是很靠前的学校，单凭申请者这些优秀的硬件背景就足以吸引一些教授的目光。