You can find the datasheet for the entire Intel 4004 processor family on Wikipedia. The article is the
Intel 4004 article, and the data sheet is found in the reference section as the
MCS-4 datasheet.
The 4001 ROM has 4-bit I/O port whose direction control is mask-programmed when the program ROM is being made. The 4002 RAM has a 4-bit output port. The 4003 is a 10-bit cascadable shift register with parallel outputs and a serial input/output. The 4004 is the CPU. Addresses and data are transferred between the parts using a time multiplexed 4-bit bus. A SYNC pulse is used to synchronize the frame.
You can also find a number of original documents on
bitsavers.org, including a programmer's manual.
Edit: Added links to the Wikipedia article, the MCS-4 datasheet, and the bitsavers.org MCS-4 repository.