By Randall Mikkelsen

WASHINGTON, Aug 20 (Reuters) - A U.S. firm pleaded guilty to illegally providing a Chinese national with data used in developing an unmanned aerial weapons system, the Justice Department said on Wednesday.

The federal plea entered by privately held Atmospheric Glow Technologies Inc, of Knoxville, Tennessee, was the latest development in a U.S. effort to crack down on improper transfers of military technology to China.

The department said Atmospheric Glow had been working on a U.S. Air Force research contract, and admitted to illegally providing test results of the project.

The information related to advanced technology for use on drone weapons systems, according to the plea agreement that made no mention of any link to the Chinese government.

The agreement said the data was provided to Xin Dai, a Chinese national and research assistant to a University of Tennessee electrical engineering professor who had been hired to work on the project.

The professor, Reese Roth, was also indicted in the case, and pleaded not guilty. His trial is set to begin on Monday, a Justice Department spokesman said.

The indictment accuses Roth of taking to China documents related to the Air Force contract that were subject to U.S. export control restrictions. It also says Roth e-mailed restricted documents to a Chinese individual.

Roth's online resume lists an honorary professorship at the Shenzhen Campus of Tsinghua University.

Another former Atmospheric Glow employee, Daniel Sherman, pleaded guilty in April to charges of conspiring with Roth to export the data.

U.S. intelligence services in recent years have accused China of increasing its spying operations in the United States to collect military and economic secrets.

A former Boeing engineer, charged in February with stealing space shuttle secrets for China, pleaded guilty in July to unauthorized possession of defense information.

China has denied spying allegations and says the charges are an example of Cold War thinking.

Atmospheric Glow faces a maximum fine of $1 million for each of the 10 counts. (Editing by Xavier Briand)

AlertNet news is provided by