Our sample in this experiment is the 630nm emission colloidal core/shell CdSe/ZnS QDs with core diameter equal to 10 nm and thickness of shell equal to 1 nm. In this experiment we can find that both the vibration mode of CdSe and ZnS appears in our Raman spectrum. Our data shown in Fig3-2、3 agree with the Raman spectrum obtained by Ref1[See Fig3-1]. The peaks presented in Fig3-2、3 are the CdSe LO mode at ~210cm-1, CdSe 2LO mode at ~420 cm-1 and ZnS LO mode at~300 cm-1. By the way, the data shown in Fig3-2, 3 are coming from the same QDs observed at different days for double check.
As we had just mentioned that the experimental data shown below are observed at different days, the data shown in Fig3-2 is observed on Oct-24/2005. In this
experiment, the pressure is added up to 34.11 GPA, and there were only three pressure
observation presented before the occurrence of phase transition. The other experiment data shown in Fig3-3 is obtained Nov-08/2005 and this data is decomposed into three parts, one is for the process of adding pressure until the peaks of CdSe LO and 2LO disappears (phase transition) [Fig3-3], the other is to add pressure up to the highest one(36.72GPA)after phase transition[Fig3-4], and the final one is for the decreasing pressure from the highest pressure down to 3.44GPA which is shown in Fig3-5. The data shown in Fig3-6 is the whole Raman spectrum in this experiment；include the loading pressure and unloading pressure process.
From our experiment data shown in Fig3-2 and Fig3-3 , one can easily see that：in the loading process the LO mode, 2LO mode of CdSe and the LO mode of ZnS are all shift to high frequency at first.（blue shift）
When the pressure approaches nearly 7GPA, we find that both LO peak and 2LO peak of CdSe start to disappear and the ZnS LO mode remains exist up to 35GPA as shown in Fig3-2 and Fig3-3. From this information we deduce that the core CdSe begins phase transition at about 7GPA, compared to the pressure-induced phase transition in bulk CdSe at about 3 GPA. We label this phase transition in QDs as the “Phase Transition I”. Due to the core-shell structure of QDs, we can reasonably assume that this retarded pressure-induced phase transition in core comes from the screening effect of ZnS shell.
By synchronal measuring the PL spectrum and Raman spectrum, we can find some interesting information by comparing these two data. The most important
phenomenon is that：the PL spectrum can’t be explored after phase transition had been found in QDs from Raman spectrum. This gives us information that the core in QD may become metal phase for two reasons：
1. When QD is under phase transition, the LO peak and 2LO peak of CdSe start to disappear in Raman spectrum; this implies PHI to occur in core of QDs.
2. QD is not under fluorescence after PHI; this verifies that the metal phase of core happens in QD.
From the band diagram of bulk CdSe with rocksalt structure as shown in Fig3-7, one can see that the conduction band overlaps the valance band; this verifies the
occurrence of the metal phase of CdSe.（the band diagram is calculated by ab-initio）
We believe the structure difference caused by phase transition in QDs is just the same as it in bulk CdSe whose structure is from zinc-blende to rocksalt. In other words we believe after phase transition the structure of core in QDs is the rocksalt structure the same as the bulk behavior.
After phase transition І, the LO and 2LO mode of CdSe disappears and a new
undetermined peak appears nearly ~156 cm-1, this new peak is very interesting due to the fix position of this peak as pressure varies! After increasing the pressure so that it
is larger than ~7GPA, the peak is always presented and the position of it unchanged until pressure goes up to 34GPA. Even if the pressure reduced below 7GPA that come to the same situation. (See the data shown in Fig3-5, this peak is also present even the pressure is reduce to be lower)
The LO signal of ZnS is very weak in 1108 data, it is almost unrecognizable before phase transition І, but after phase transition І it can be recognized even if it is still weak and nearly disappears at 36.72GPA（See Fig3-4）. At first, this pressure might be regarded as another phase transition, however by comparing with the 1024 data
（Fig3-2）, we see the signal of ZnS is always distinguishable up to pressure equal to 34.11GPA. For this reason it should not be another phase transition and may come from the defect of sample.
While in the unloading process, the LO mode of ZnS starts to be red shift along the original path of loading process.（see Fig3-5） When the pressure reaches 3.44GPA, the LO mode of CdSe appears again and its peak position corresponds to the data of loading process, however, the CdSe 2LO mode does not appear again. This
phenomenon makes us to believe that CdSe might not reduce back to the zinc-blende structure again. Thus the phase transition І is an irreversible one.
Fig3-1：Raman spectrum of CdSe/ZnS QD’s with a shell thickness 3.4 ML excited by a 476.5-nm line of
an Ar12 laser.
Fig3-2：Raman spectrum of CdSe/ZnS QDs in di-water under different pressure.（gotten in 1024）
Fig3-3： Raman spectrum of CdSe/ZnS QDs in di-water. The pressure is added up to phase
transition.（gotten in 1108）
Fig3-4：Raman spectrum of CdSe/ZnS QDs in di-water.（gotten in 1108）
Pressure is added above phase transition up to 36.72GPA.
Fig3-5：Raman spectrum of CdSe/ZnS QDs in di-water.（gotten in 1108）
Reduce pressure from 36.72GPA to 3.44GPA.
Fig3-6：Whole Raman spectrum of CdSe/ZnS QDs in di-water.（gotten in 1108）
Fig3-7：Band diagram of CdSe with rocksalt structure calculated by CASTEP.