Try procuring antisera from any of the reputed make.

The identification should be confirmed using, for example suspected S. Typhi using Salmonella polyvalent O, O9 and dH antisera (example: Murex Biotech, England).

Accordingly check your need of antisera, depending on your interest of Salmonellae- typhoidal or non-typhoidal.

Follow Mackie and McCartney practical medical Microbiology:

Collee JG, Miles RS, Watt B. Tests for the identification of bacteria. In: Collee JG, Fraser AG, Marmion BP, Simmons A, editors, Mackie and McCartney practical medical Microbiology. 14th ed. London: Churchill Livingstone; 1996 p. 131-49

We currently serotype Salmonella strains and use sera provided by Biorad. You will get relevant informations for serotyping from this link http://www.pasteur.fr/ip/portal/action/WebdriveActionEvent/oid/01s-000036-089

Try serotyping only after your complete biochemical characterization. 

Basics: Serotyping is a definitive typing method used for epidemiological characterisation of bacteria. Serotyping of Salmonella strains is carried out by identification of surface antigens (LPS, O-antigens) and flagella antigens (proteins, H-antigens). Most commonly, strains of Salmonella express two phases of H- antigens but aphasic, monophasic and triphasic variants are known. The definition of the serotypes is based on the antigen combination present and is given in the “Kauffmann-White scheme”, Popoff and Le Minor, WHO Centre for Reference and Research on

Salmonella, Institut Pasteur, France, 1997.

Basics: Serotypes are groups within a single species of microorganisms, such as bacteria or viruses, which share distinctive surface structures. For instance, Salmonella bacteria look alike under the microscope but can be separated into many serotypes based on two structures on their surface:  The outermost portion of the bacteria’s surface covering, called the O antigen; and A slender threadlike structure, called the H antigen, that is part of the flagella. The O antigens are distinguished by their different chemical make-up. The H antigens are distinguished by the protein content of the flagella. Each O and H antigen has a unique code number. Scientists determine the serotype based on the distinct combination of O and H antigens.

Dr. Isabelle its a very useful link.

You could consider using Multilocus Sequence Typing (MLST) if you have only a few isolates as this identifies the bast majority of stereotypes well. The typing sera is expensive, it is laborious and often subjective in judging agglutination. Getting both flagella phases is also difficult. Even experienced labs can struggle with atypical isolates such a monophasic Typhimurium like strains.

The agglutination tests according to the Kauffmann-White scheme:

P. Grimont, F.-X. Weill, Antigenic Formulae of the Salmonella Servovars, WHO Collaborating Center Reference Research on Salmonella, 2007, pp. 1-167.

I certainly do agree with the observations/comments made above, that serotyping can be expensive, the procedure tends to be laborious and often time subjective. From my personal experience: Getting around this challenge, I referred the isolates under investigation to a reference Laboratory/Institution reknown for such works. However, you are welcome to read my article entilted ""Characterization of Salmonellae isolated from domestic animals and wildlife in selected areas of Zambia". Therein, the Institution is mentioned without advertising any. I suppose, there could be many other ways or institutions that can assist you to achieve your desired goals i.e. on and above what has been offered to you by other Scientists. Best wishes in your endevour.