There have been few comprehensive studies tracking the extent of open access archiving. A 2009 survey published in Information Research by the same team put the proportion of open access papers in 2006 at around 19 per cent.

More copies are being deposited in subject-specific open access archives than on the home pages of authors or their departments as was the case in the past.

"A majority of international publishers actually allow the posting of some versions of published articles, sometimes after a delay, in such repositories," Björk told SciDev.Net.

"Funding organisations are increasingly making it a condition of their funding that a copy is also available open access," Björk said. When around a third of all research articles become open access, libraries will seriously question paying for journals, he added.

"I would expect that the overall percentage would increase by 1–1.5 per year based on the number of open access journals in existence," he said, adding that this might increase further in the future.

"Professor Björk’s data is extremely important but it is not that sensitive," said Stevan Harnad, 'open access scientometrics' researcher from the UK-based Southampton University. "What was missing from the analysis was from when papers were made free to view."

"The growth rate in green and gold open access is not the same — green is only at around 15 per cent," he said. "Gold open access is growing fairly steadily year to year while green publishing is accelerating."
The survey was published last month (23 June) in PLoS ONE.