In mixed technology environments, from what I have read, the RTS/CTS and CTS-to-self frames are sent in such a way that older devices can understand them. These are used so that other stations will set their NAV timers with the info in the Duration/ID field so they don't try to send on the medium while another station is sending (which they may or may not be able to understand). The Duration/ID field is set to a value such that the duration of the transmission of the subsequent data and ACK will be accounted for. This is used for the hidden node problem as well.
The entire PPDU following the RTS/CTS or CTS-to-self is sent at a higher modulation that some stations may not be able to understand.
With the exception of 802.11g and DSSS-OFDM modulation, the older devices' "preamble/header [is] sent using the same modulation format as the payload" [http://www.radio-electronics.com/info/wireless/wi-fi/ieee-802-11g.php ]; however, it appears that in 802.11n, from what I have read, the PHY header seems to be always sent at a slower rate that older devices can understand. The length field is changed so that now it is not just showing the length of the rest of the PPDU, but the rest of the PPDU and the subsequent ACK. EDIT--I guess this is only the HT-Mixed format PPDU.
Is my understanding of this correct? It seems that 802.11n doesn't need to use RTS/CTS or CTS-to-self in order to reserve the medium; they can do it through the PHY header. Is RTS/CTS and CTS-to-self only used for hidden node problem in 802.11n? Is there ever a case where the 802.11n PHY header is sent at the same modulation as the payload?
Also, I have read about Dual-CTS protection, but I do not recall seeing it in the CWNA study guide. It seems that the AP will send a CTS in the non-HT format (so nearby non-HT cells will defer) and then again in the HT format. Is this only used in a pure HT environment?